Merge remote-tracking branch 'hf-test/main' into claude/custom-mcp-spotter-demo-125da8

.gitignore

@@ -228,3 +228,16 @@ scratch/
 legitdata_project/venv/
 legitdata_project/*.egg-info/
 legitdata_project/.DS_Store
+
+# Portal builder (demoprep_app/portal) — vendored template + specs must be tracked.
+# Overrides the broad *.json / lib/ / dist rules above for this directory only.
+!demoprep_app/portal/**/*.json
+!demoprep_app/portal/template-tse/src/lib/
+!demoprep_app/portal/template-tse/src/lib/**
+demoprep_app/portal/template-tse/node_modules/
+demoprep_app/portal/template-tse/dist/
+# Generated portal apps (codemod output) stay out of the repo
+/*-tse/
+
+# Portal build working dir (scripts/build_portal.py output)
+/build/

CLAUDE.md

@@ -85,6 +85,37 @@ For overall project status, see `PROJECT_STATUS.md` in root.
 | `main` | Production — always stable and shippable | `hf` (prod) |
 | `develop` | Sprint integration — current sprint work | `hf-test` |
 | `feature/*` | Individual features — branch off `develop` | local / `hf-test` |
+| `deploy/mcp` | MCP Space integration — shares history via `develop` merges | `hf-mcp-test`, then `hf-mcp` (prod) |
+
+### Remotes
+
+| Remote | Space | Tier |
+|--------|-------|------|
+| `hf` | `thoughtspot-dp/demoprep` | app — **PROD** |
+| `hf-test` | `thoughtspot-dp/test-demoprep` | app — test |
+| `hf-mcp` | `thoughtspot-demoprep/mcp` | MCP — **PROD** |
+| `hf-mcp-test` | `thoughtspot-demoprep/test-mcp` | MCP — test |
+
+### ⚠️ Reality check (verified 2026-08-26) — the model below is ASPIRATIONAL
+
+The documented flow and the deployed truth have drifted. Verify before trusting:
+
+| Claim in this doc | Actual state on 2026-08-26 |
+|---|---|
+| `develop` → `hf-test` | **False.** `hf-test/main` == `spike/portal-template`. A spike branch is the de facto integration branch. |
+| `develop` is sprint integration | **Stale.** `develop` (tip `8c598cc`) is missing the current CLAUDE.md and the whole "Data Quality Push" sprint section. |
+| Space remotes only receive branch pushes | **False.** `hf`/`hf-mcp` accept direct pushes and routinely hold commits no local branch has (that is how `4b87c36` came to exist only on the Space). |
+
+**Always confirm what is actually deployed before reasoning about it:**
+```bash
+git fetch --all
+git log --oneline -1 hf-mcp/main        # what PROD MCP is really running
+git branch --all --contains <sha>       # who has a given fix
+```
+Deployed sha is also visible per-Space via `HfApi().space_info(repo_id).sha`.
+
+**Do not "fix" this drift by rewriting the model here** — decide the intended
+model with boone first, then make doc and reality match in one deliberate pass.
 
 ### Rules
 
@@ -117,8 +148,38 @@ git push hf-test develop:main
 git push origin main develop
 ```
 
+**MCP Space deploys (`deploy/mcp`) — test tier is MANDATORY:**
+```bash
+git checkout deploy/mcp
+git merge develop                        # pick up sprint work
+# ... MCP-specific work ...
+git push hf-mcp-test deploy/mcp:main     # 1. deploy to TEST first — ALWAYS
+#     verify (see smoke check below), THEN:
+git push hf-mcp deploy/mcp:main          # 2. deploy to PROD
+```
+
+**MCP post-deploy smoke check — a green build log is NOT proof:**
+```bash
+curl -s https://thoughtspot-demoprep-test-mcp.hf.space/
+```
+Expect `{"status":"ok"}`. The health endpoint is unauthenticated, but it only
+proves the container booted — it does **not** exercise a build. Also call the
+`ping` tool, and run one real `build_demo_from_brief` against test before prod.
+
+> **The 2026-08-25 outage happened because logging code that only the MCP path
+> constructs (`PromptLogger`) raised `PermissionError: '/app/logs'` under the
+> Space's non-root user — every MCP build returned `failed`, while `GET /` still
+> answered `ok`. HF Spaces run the container as a non-root user inside a
+> root-owned WORKDIR, and `logs/` is gitignored so it is absent from the image.
+> Anything that writes to disk at runtime must resolve a writable dir or degrade.**
+
 > **Never push sprint/feature work directly to `main`.  
-> Never push to `hf` (prod) from anything other than `main`.**
+> Never push to `hf` (prod) from anything other than `main`.  
+> Never push to `hf-mcp` (prod) without deploying to `hf-mcp-test` first.  
+> Never leave a hotfix living only on a Space remote** — a fix pushed straight to
+> `hf-mcp`/`hf` that is not merged back into `deploy/mcp`, `develop`, AND `main`
+> will be silently reverted by the next routine deploy. After ANY Space hotfix,
+> run `git branch --all --contains <sha>` and confirm every deploy branch lists it.
 
 ### Sprint Doc Workflow - IMPORTANT
 
@@ -209,7 +270,7 @@ Example:
 ## Key Project Facts
 
 ### Environment Setup
-- **Python**: Using virtual environment at `./demoprep/bin/activate`
+- **Python**: Using virtual environment at `./.venv/bin/activate`
 - **NOT using conda** - if you don't see packages, activate the venv first
 - **Supabase**: IS installed and configured - credentials in .env work fine
 - **ThoughtSpot**: Trusted auth via per-environment key (selected from TS Environment dropdown). No global admin user — objects are created under the logged-in user's email.
@@ -218,7 +279,7 @@ Example:
 
 ```bash
 # Run the app (chat_interface.py IS the full app — App tab, Chat tab, Admin, all tabs)
-source ./demoprep/bin/activate && python chat_interface.py
+source ./.venv/bin/activate && python chat_interface.py
 
 # Check git changes
 git diff --stat
@@ -232,7 +293,7 @@ lsof -i :7860
 2. **DO NOT** try to install supabase or other packages - they're already in the venv
 3. **DO NOT** change default values to "thoughtspot.com" - that's for customer URLs, not ThoughtSpot's
 4. **DO NOT** assume worksheets are needed - they're deprecated, we use models now
-5. **ALWAYS** use the venv when running Python: `source ./demoprep/bin/activate && python`
+5. **ALWAYS** use the venv when running Python: `source ./.venv/bin/activate && python`
 
 ### Working Patterns
 - Settings save/load through Supabase - this WORKS when using venv

Dockerfile

@@ -12,13 +12,25 @@ FROM python:3.11-slim
 RUN useradd -m -u 1000 user
 WORKDIR /app
 
-RUN apt-get update && apt-get install -y git curl && rm -rf /var/lib/apt/lists/*
+# node + npm are required by the build_portal tool: the portal codemod is a Node
+# script and the site is built with vite. Debian bookworm ships Node 18, which
+# satisfies vite 5 (^18 || >=20).
+RUN apt-get update && apt-get install -y git curl nodejs npm \
+    && rm -rf /var/lib/apt/lists/*
 
 COPY --chown=user requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
 COPY --chown=user . .
 
+# Pre-install the portal template's node_modules AS ROOT, at image build time.
+# The codemod installs them on first use if absent — but the runtime user cannot
+# write into /app, so that would fail exactly like the 2026-08-25 PromptLogger
+# outage. Installing here means every portal build finds them ready.
+RUN cd demoprep_app/portal/template-tse \
+    && npm install --no-audit --no-fund --loglevel=error \
+    && chown -R user:user node_modules
+
 USER user
 ENV MCP_TRANSPORT=http \
     MCP_HTTP_PORT=7860 \

chat_interface.py

@@ -395,6 +395,7 @@ class ChatDemoInterface:
         self._session_logger = None
         self._prompt_logger = None
         self._demo_bundle = None
+        self._demo_blueprint = None
 
     def clear_deployment_completion(self) -> None:
         self.deployment_completion = None
@@ -3165,6 +3166,7 @@ To change settings, use:
             schema_name = f"DM{yymmdd}_{hhmmss}_{company_clean}_{usecase_clean}"
 
             self._demo_bundle = None
+            self._demo_blueprint = None
             from demoprep_app.pipeline.build_demo import build_demo
 
             row_count_guidance = int(self.settings.get("fact_table_size", 5000) or 5000)
@@ -3194,6 +3196,10 @@ To change settings, use:
                     "Check the warnings above and retry."
                 )
             self._demo_bundle = build.dataset
+            # The blueprint is what the post-load quality gate needs: the
+            # formulas it can recompute a bad column from, and the measures the
+            # demo's story actually depends on.
+            self._demo_blueprint = build.blueprint
             self.demo_builder.schema_generation_results = build.ddl
             self.ddl_code = build.ddl
             for warning in build.warnings:
@@ -3959,22 +3965,34 @@ LegitData will generate realistic, AI-powered data.
             log_progress(f"[OK] {pop_message}")
 
             # Data-quality gate: after every load, profile the MEASURE columns
-            # that were actually written to Snowflake and fail loudly if any is
-            # entirely zero/null (recurring zero-measure defect — derived
-            # measures like TOTAL_*_USD loading as all zeros → blank tiles).
+            # that were actually written to Snowflake, then REPAIR what it can
+            # (recompute a derived column from its formula, clamp a part to its
+            # whole, binarize a flag) and OMIT what it cannot. It aborts the run
+            # only when there is nothing demo-able left — a board with one
+            # missing tile beats an empty schema.
             from demoprep_app.integrations.snowflake import (
-                ZeroMeasureError,
+                MeasureQualityGateError,
+                apply_measure_omissions,
+                measure_formulas,
                 run_measure_quality_gate,
             )
             progress += f"\n\n**Verifying data quality (measure columns)...**"
             yield progress
             log_progress("")
             log_progress("Running data-quality gate on measure columns...")
+            _dq_formulas = measure_formulas(self._demo_bundle) if self._demo_bundle else {}
+            _dq_required = sorted({
+                i.measure for i in getattr(self._demo_blueprint, "insights", []) or [] if i.measure
+            })
             try:
                 dq_profile = run_measure_quality_gate(
-                    deployer.connection, schema_name, progress_callback=log_progress
+                    deployer.connection,
+                    schema_name,
+                    progress_callback=log_progress,
+                    formulas=_dq_formulas,
+                    required_measures=_dq_required,
                 )
-            except ZeroMeasureError as e:
+            except MeasureQualityGateError as e:
                 self._last_population_error = str(e)
                 self._last_schema_name = schema_name
                 _deploy_meta = {
@@ -3991,20 +4009,46 @@ LegitData will generate realistic, AI-powered data.
                 raise Exception(str(e))
 
             dq_warnings = dq_profile.get("warnings", [])
+            dq_repairs = dq_profile.get("repairs", [])
+            dq_omissions = dq_profile.get("omissions", [])
             self._dq_warnings = dq_warnings
             progress += (
-                f"\n[OK] Data quality gate passed "
-                f"({dq_profile['measure_columns_checked']} measure columns "
-                f"across {dq_profile['tables_checked']} tables)"
+                f"\n[OK] Data quality gate: "
+                f"{dq_profile['measure_columns_checked']} measure columns "
+                f"across {dq_profile['tables_checked']} tables"
             )
+            for _dq_r in dq_repairs:
+                progress += f"\n[REPAIR] {_dq_r}"
+            if dq_omissions:
+                # A column dropped in Snowflake has to disappear from the DDL
+                # too: the ThoughtSpot model and liveboard are built from it
+                # later in the run and would otherwise declare a column that no
+                # longer exists.
+                _removed = apply_measure_omissions(self._demo_bundle, dq_omissions)
+                if _removed:
+                    from demoprep_app.ddl import DatasetDdlCompiler
+                    _new_ddl = DatasetDdlCompiler().compile(self._demo_bundle)
+                    self.ddl_code = _new_ddl
+                    self.demo_builder.schema_generation_results = _new_ddl
+                for _dq_o in dq_omissions:
+                    progress += (
+                        f"\n[OMITTED] {_dq_o['table']}.{_dq_o['column']} "
+                        f"({_dq_o['defect']}: {_dq_o['detail']})"
+                    )
             for _dq_w in dq_warnings:
                 progress += f"\n[WARN] {_dq_w}"
             if _slog:
                 _slog.log_verbose(
-                    "deploy", "data quality gate passed",
+                    "deploy", "data quality gate complete",
                     measure_columns_checked=dq_profile["measure_columns_checked"],
                     tables_checked=dq_profile["tables_checked"],
                     dq_warnings=dq_warnings,
+                    dq_repairs=dq_repairs,
+                    dq_omissions=[f"{o['table']}.{o['column']}" for o in dq_omissions],
+                    dq_defects=[
+                        f"{d['table']}.{d['column']}:{d['defect']}"
+                        for d in dq_profile.get("defects", [])
+                    ],
                 )
 
             self._deployed_schema_name = schema_name
@@ -4017,6 +4061,12 @@ LegitData will generate realistic, AI-powered data.
                     "measure_columns_checked": dq_profile["measure_columns_checked"],
                     "tables_checked": dq_profile["tables_checked"],
                     "warnings": dq_warnings,
+                    "repairs": dq_repairs,
+                    "omissions": [f"{o['table']}.{o['column']}" for o in dq_omissions],
+                    "defects": [
+                        f"{d['table']}.{d['column']}:{d['defect']}"
+                        for d in dq_profile.get("defects", [])
+                    ],
                 },
             }
             

demoprep_app/dataset/engine.py

@@ -15,12 +15,15 @@ from __future__ import annotations
 
 import ast
 import math
-import operator
 import random
 from datetime import date, timedelta
 from typing import Any
 
 from demoprep_app.dataset.contracts import DatasetBundle, DatasetColumn, DatasetTable
+from demoprep_app.dataset.formula import (
+    evaluate as eval_formula,
+    identifiers as formula_identifiers,
+)
 from demoprep_app.scenario.blueprint import (
     BlueprintDimension,
     BlueprintMeasure,
@@ -46,7 +49,22 @@ def _safe_date(value):
 class BlueprintEngine:
     """Deterministic executor for DemoBlueprint -> DatasetBundle."""
 
+    def __init__(self) -> None:
+        # Warnings raised while evaluating formulas (currently: division by
+        # zero, which is legitimate but worth surfacing). Everything that is
+        # NOT legitimate raises FormulaEvaluationError instead of silently
+        # writing 0.0 to every row.
+        self._formula_warnings: list[str] = []
+
+    def formula_warnings(self) -> list[str]:
+        """One deduplicated line per distinct formula warning, with row counts."""
+        from collections import Counter
+
+        counts = Counter(self._formula_warnings)
+        return [f"{message} ({count} row(s))" for message, count in sorted(counts.items())]
+
     def generate(self, blueprint: DemoBlueprint, row_count: int | None = None) -> DatasetBundle:
+        self._formula_warnings = []
         rng = random.Random(blueprint.seed or 1)
         total_target = max(500, int(row_count or 5000))
         per_fact = max(500, total_target // max(1, len(blueprint.facts)))
@@ -198,7 +216,27 @@ class BlueprintEngine:
                     # Never clamp a derived value: a negative margin / ending
                     # inventory is legitimate and must still satisfy its formula
                     # (the validator re-checks it against the stored values).
-                    values[m.name] = _format_value(_eval_formula(m.formula, values), m.fmt, clamp=False)
+                    # eval_formula RAISES on an uncomputable formula — it used to
+                    # return 0.0, which shipped all-zero and fabricated columns.
+                    values[m.name] = _format_value(
+                        eval_formula(
+                            m.formula, values, measure=m.name, warnings=self._formula_warnings
+                        ),
+                        m.fmt,
+                        clamp=False,
+                    )
+                    continue
+                if _is_flag_measure(m):
+                    # A *_FLAG column is a boolean event marker: 0 or 1, never
+                    # 5 or 158 (Triumph shipped both). Sample the EVENT, not a
+                    # magnitude, so any rate built on it is meaningful. Insights
+                    # and dimension performance move the probability, so a
+                    # planted story on a flag still shows up.
+                    flag_p = _flag_probability(m)
+                    flag_p *= _insight_multiplier(insights_by_measure.get(m.name, ()), picked, d)
+                    if m.seasonal:
+                        flag_p *= perf_multiplier
+                    values[m.name] = 1 if rng.random() < min(max(flag_p, 0.01), 0.99) else 0
                     continue
                 if m.share_of:
                     # "part of a whole": a fraction of the parent's value. The parent
@@ -209,8 +247,36 @@ class BlueprintEngine:
                     low = m.params.get("low", 0.05)
                     high = m.params.get("high", 0.85)
                     imult = _insight_multiplier(insights_by_measure.get(m.name, ()), picked, d)
-                    frac = min(max(rng.uniform(low, high) * imult, 0.0), 1.0)
-                    values[m.name] = _format_value(values.get(m.share_of, 0.0) * frac, m.fmt)
+                    # The FRACTION must carry dimension performance, or every
+                    # rate built as part/whole is identical across the board.
+                    # perf_multiplier scales the parent AND (without this) the
+                    # part equally, so it cancels: rate = (whole*p*f)/(whole*p)
+                    # = f. A collection rate then reads the same for the best and
+                    # worst provider — AdvancedMD shipped 0.7% spread. Damped,
+                    # because a rate should move less than the money it derives
+                    # from: a 40% better provider collects somewhat more of what
+                    # it bills, not 40% more.
+                    # Performance moves WHERE IN THE BAND this row lands, it does
+                    # not push past it. Multiplying the sampled fraction instead
+                    # drove 16% of rows through the ceiling and pinned them at a
+                    # 100% collection rate — a clamp artifact that reads as a bug
+                    # on a chart. [low, high] is the blueprint's statement of what
+                    # is plausible; a strong performer sits high in that range,
+                    # not outside it.
+                    # WARP the sampling curve rather than scaling and clamping.
+                    # u**(1/p**2) is a smooth monotone map of uniform(0,1) onto
+                    # itself: a strong performer's draws bunch toward the top of
+                    # the band, a weak one's toward the bottom, with no hard edge
+                    # and so no pile-up. Scaling the fraction directly instead
+                    # pinned 16% of rows at a 100% collection rate; positioning
+                    # linearly inside the band removed the artifact but flattened
+                    # the spread back to 5.9%, under the 10-12% the reviewer set.
+                    frac_perf = max(1.0 + (perf_multiplier - 1.0) * RATE_PERF_SENSITIVITY, 0.05)
+                    position = min(max(rng.random() ** (1.0 / frac_perf ** 2) * imult, 0.0), 1.0)
+                    frac = min(max(low + (high - low) * position, 0.0), 1.0)
+                    values[m.name] = _format_value(
+                        values.get(m.share_of, 0.0) * frac, m.fmt, pct_max=_pct_bound(m)
+                    )
                     continue
                 if m.per and f"{m.per}_KEY" in row:
                     # Stable characteristic of an entity: one level per member,
@@ -227,10 +293,20 @@ class BlueprintEngine:
                 mult *= _insight_multiplier(insights_by_measure.get(m.name, ()), picked, d)
                 # Format base values NOW so derived formulas — and the validator,
                 # which re-checks formulas against stored values — see identical numbers.
-                values[m.name] = _format_value(base * mult, m.fmt)
+                values[m.name] = _format_value(
+                    base * mult, m.fmt, pct_max=_pct_bound(m),
+                    signed=getattr(m, "signed", False),
+                )
 
             for m in fact.measures:
-                row[m.name] = values.get(m.name, 0.0)
+                if m.name not in values:
+                    # Dependency ordering covers every measure, so this can only
+                    # mean an engine bug. Say so instead of writing a silent 0.
+                    raise RuntimeError(
+                        f"Fact {fact.name}: measure {m.name} was never evaluated "
+                        f"(kind={m.kind}, formula='{m.formula}') — refusing to store 0."
+                    )
+                row[m.name] = values[m.name]
             rows.append(row)
 
         return DatasetTable(fact.name, fact.grain, columns, rows, is_fact=True)
@@ -266,6 +342,12 @@ def _trend_multiplier(pct_per_year: float, start: date, current: date) -> float:
     return (1.0 + pct_per_year / 100.0) ** years_elapsed
 
 
+# How strongly a part/whole FRACTION responds to dimension performance. Below
+# 1.0 so rates vary less than the money they come from, and so the bound
+# part <= whole is not constantly pushed against.
+RATE_PERF_SENSITIVITY = 1.0
+
+
 def _sample(measure: BlueprintMeasure, rng: random.Random) -> float:
     p = measure.params
     if measure.distribution == "uniform":
@@ -273,6 +355,12 @@ def _sample(measure: BlueprintMeasure, rng: random.Random) -> float:
     mean = p.get("mean", 1.0)
     sigma = p.get("sigma", 0.4)
     if measure.distribution == "normal":
+        if getattr(measure, "signed", False):
+            # sigma is an ABSOLUTE spread here. A signed measure is centred at
+            # or near zero, and the relative branch below (mean * sigma) would
+            # collapse to a stddev of 0 — an all-zero column, the exact defect
+            # this release exists to remove.
+            return rng.gauss(mean, abs(sigma) or 1.0)
         return max(0.0, rng.gauss(mean, abs(sigma) if sigma > 1 else mean * abs(sigma)))
     # lognormal: mean is the desired arithmetic mean, sigma the log-space spread
     sigma = min(max(abs(sigma), 0.05), 1.5)
@@ -298,55 +386,79 @@ def _insight_multiplier(entries, picked: dict[str, str], d: date) -> float:
     return mult
 
 
-def _format_value(value: float, fmt: str, clamp: bool = True) -> Any:
+def _format_value(value: float, fmt: str, clamp: bool = True, pct_max: float = 100.0,
+                  signed: bool = False) -> Any:
     """Round a value for storage. clamp bounds sampled base measures to a sane
-    range (no negative units, pct in 0-100); derived measures pass clamp=False
-    so their stored value stays faithful to the formula the validator checks."""
+    range (no negative units, pct within its natural bound); derived measures
+    pass clamp=False so their stored value stays faithful to the formula the
+    validator checks.
+
+    pct_max is the measure's natural upper bound: 1.0 for a 0-1 rate, 100.0 for
+    a 0-100 percentage. Guessing 100 for a 0-1 rate is how a "utilization rate"
+    of 42.0 (meaning 4200%) reaches a KPI tile.
+    """
     if fmt == "int":
         v = int(round(value))
-        return max(0, v) if clamp else v
+        return max(0, v) if (clamp and not signed) else v
     if fmt in ("currency", "float"):
         return round(value, 2)
     if fmt == "pct":
         if clamp:
-            value = min(max(value, 0.0), 100.0)
-        return round(value, 2)
+            value = min(max(value, -pct_max if signed else 0.0), pct_max)
+        return round(value, 4 if pct_max <= 1.0 else 2)
     return round(value, 2)
 
 
-def _sql_type(measure: BlueprintMeasure) -> str:
-    return "NUMBER" if measure.fmt == "int" else "NUMBER(16,2)"
-
-
-# --- Safe arithmetic formula evaluation (no eval) --------------------------
-
-_OPS = {
-    ast.Add: operator.add,
-    ast.Sub: operator.sub,
-    ast.Mult: operator.mul,
-    ast.Div: lambda a, b: a / b if b else 0.0,
-    ast.USub: operator.neg,
-}
+def _pct_bound(measure: BlueprintMeasure) -> float:
+    """Infer whether a pct measure is on a 0-1 or 0-100 scale from its params.
+
+    The blueprint does not declare the scale, so read it off the numbers the
+    LLM asked for: a mean/high above 1.5 can only mean a 0-100 percentage.
+    """
+    if measure.fmt != "pct":
+        return 100.0
+    declared = [
+        measure.params.get(key)
+        for key in ("mean", "high", "low")
+        if isinstance(measure.params.get(key), (int, float))
+    ]
+    return 100.0 if any(value > 1.5 for value in declared) else 1.0
+
+
+def _is_flag_measure(measure: BlueprintMeasure) -> bool:
+    """A *_FLAG measure is a 0/1 event marker regardless of what the LLM declared."""
+    return measure.name.upper().endswith("_FLAG")
+
+
+def _flag_probability(measure: BlueprintMeasure) -> float:
+    """Probability that a flag is 1, read from whatever the blueprint declared.
+
+    Accepts an explicit params.p, else a mean/high already expressed as a
+    fraction; anything else (a lognormal mean of 40 for a flag) falls back to
+    an even split rather than inventing a magnitude.
+    """
+    for key in ("p", "probability"):
+        value = measure.params.get(key)
+        if isinstance(value, (int, float)) and 0.0 < float(value) <= 1.0:
+            return float(value)
+    for key in ("mean", "high"):
+        value = measure.params.get(key)
+        # Strictly < 1: a declared mean of exactly 1 (or 9.6) is a magnitude,
+        # not a probability, and reading it as one produces an always-1 flag —
+        # which is just the all-constant defect wearing a different hat.
+        if isinstance(value, (int, float)) and 0.0 < float(value) < 1.0:
+            return float(value)
+    return 0.5
 
 
-def _eval_formula(formula: str, values: dict[str, float]) -> float:
-    try:
-        tree = ast.parse(formula, mode="eval")
-        return float(_eval_node(tree.body, values))
-    except Exception:
-        return 0.0
+def _sql_type(measure: BlueprintMeasure) -> str:
+    return "NUMBER" if measure.fmt == "int" else "NUMBER(16,2)"
 
 
-def _eval_node(node: ast.AST, values: dict[str, float]) -> float:
-    if isinstance(node, ast.BinOp) and type(node.op) in _OPS:
-        return _OPS[type(node.op)](_eval_node(node.left, values), _eval_node(node.right, values))
-    if isinstance(node, ast.UnaryOp) and type(node.op) in _OPS:
-        return _OPS[type(node.op)](_eval_node(node.operand, values))
-    if isinstance(node, ast.Constant) and isinstance(node.value, (int, float)):
-        return float(node.value)
-    if isinstance(node, ast.Name):
-        return float(values.get(node.id, values.get(node.id.upper(), 0.0)))
-    raise ValueError(f"Unsupported expression node: {ast.dump(node)}")
+# --- Formula evaluation ----------------------------------------------------
+# The language (parse, validate, evaluate, compile-to-SQL) lives in
+# demoprep_app/dataset/formula.py so the blueprint validator, this engine, the
+# dataset validator and the Snowflake quality gate all agree on it.
 
 
 # --- Measure dependency analysis -------------------------------------------
@@ -387,10 +499,7 @@ def _order_measures(measures: list) -> list:
 
 
 def _formula_names(formula: str) -> set[str]:
-    try:
-        return {n.id for n in ast.walk(ast.parse(formula, mode="eval")) if isinstance(n, ast.Name)}
-    except Exception:
-        return set()
+    return formula_identifiers(formula)
 
 
 def _signed_bases(formula: str, measures_by_name: dict, base_names: set[str]) -> tuple[set[str], set[str]]:
@@ -413,6 +522,11 @@ def _signed_bases(formula: str, measures_by_name: dict, base_names: set[str]) ->
                 walk(node.right, -sign, seen)
         elif isinstance(node, ast.UnaryOp) and isinstance(node.op, ast.USub):
             walk(node.operand, -sign, seen)
+        elif isinstance(node, ast.Call):
+            # ABS/ROUND/MIN/MAX/COALESCE preserve the direction of their inputs
+            # closely enough to pick a base to scale for an insight.
+            for arg in node.args:
+                walk(arg, sign, seen)
         elif isinstance(node, ast.Name):
             if node.id in base_names:
                 (pos if sign > 0 else neg).add(node.id)

demoprep_app/dataset/formula.py

@@ -0,0 +1,456 @@
+"""The measure-formula language: parse, validate, evaluate, compile to SQL.
+
+ONE definition of what a derived-measure formula may contain, shared by
+everything that touches formulas:
+
+  * blueprint.py   — validates formulas BEFORE any generation, Snowflake, or
+                     ThoughtSpot work happens, and repairs the fixable ones.
+  * engine.py      — evaluates them per row while generating the dataset.
+  * validator.py   — re-checks stored values against their formulas.
+  * quality_gate.py— recomputes a bad column IN Snowflake via to_sql().
+
+Why this module exists (2026-08-26): the evaluator used to swallow every
+failure and return 0.0, so an unsupported function (`ABS(...)`), a typo, or a
+reference to a measure in a DIFFERENT fact table silently produced a column of
+zeros — or, worse, a plausible-looking wrong number (`0 - FORECAST_PRICE`).
+Nothing validated formulas upstream. See
+dev_notes/BLOCKING_ISSUES_2026_08_26.md ISSUE A.
+
+Rules of the language:
+  * arithmetic: + - * / and unary minus
+  * numeric literals
+  * identifiers: measures in the SAME fact table, nothing else
+  * functions: only the whitelist below (the ones LLMs actually write)
+  * division by zero yields 0.0 and records a warning — that one is legitimate
+    (a rate with no denominator rows), everything else raises.
+"""
+
+from __future__ import annotations
+
+import ast
+import math
+import operator
+from typing import Any, Callable, Iterable, Mapping, MutableSequence
+
+
+class FormulaEvaluationError(ValueError):
+    """A measure formula could not be evaluated or validated.
+
+    Carries the measure, the formula, and the offending token so the message
+    is actionable instead of a bare "unsupported node".
+    """
+
+    def __init__(self, reason: str, *, measure: str = "", formula: str = "", token: str = ""):
+        self.reason = reason
+        self.measure = measure
+        self.formula = formula
+        self.token = token
+        where = f"measure {measure}" if measure else "formula"
+        detail = f" (offending token: {token})" if token else ""
+        super().__init__(f"{where}: formula '{formula}' cannot be computed — {reason}{detail}")
+
+
+# --- the whitelist -----------------------------------------------------------
+# (min_args, max_args or None for variadic, python impl, sql template kind)
+
+def _round(*args: float) -> float:
+    value = args[0]
+    digits = int(args[1]) if len(args) > 1 else 0
+    return round(value, digits)
+
+
+def _coalesce(*args: float) -> float:
+    for value in args:
+        if value is not None:
+            return value
+    return 0.0
+
+
+ALLOWED_FUNCTIONS: dict[str, tuple[int, int | None, Callable[..., float], str]] = {
+    # name        min max   python                                    sql
+    "ABS":       (1, 1,     lambda x: abs(x),                         "ABS"),
+    "ROUND":     (1, 2,     _round,                                   "ROUND"),
+    "MIN":       (2, None,  lambda *a: min(a),                        "LEAST"),
+    "MAX":       (2, None,  lambda *a: max(a),                        "GREATEST"),
+    "LEAST":     (2, None,  lambda *a: min(a),                        "LEAST"),
+    "GREATEST":  (2, None,  lambda *a: max(a),                        "GREATEST"),
+    "COALESCE":  (1, None,  _coalesce,                                "COALESCE"),
+    "IFNULL":    (2, 2,     _coalesce,                                "COALESCE"),
+}
+
+_BIN_OPS: dict[type, Callable[[float, float], float]] = {
+    ast.Add: operator.add,
+    ast.Sub: operator.sub,
+    ast.Mult: operator.mul,
+}
+
+_SQL_BIN_OPS: dict[type, str] = {
+    ast.Add: "+",
+    ast.Sub: "-",
+    ast.Mult: "*",
+    ast.Div: "/",
+}
+
+ALLOWED_SYNTAX_SUMMARY = (
+    "use + - * / and the functions "
+    + ", ".join(sorted(ALLOWED_FUNCTIONS))
+    + " over other measures in the same fact table"
+)
+
+
+# --- parsing ----------------------------------------------------------------
+
+
+def parse(formula: str, *, measure: str = "") -> ast.expr:
+    """Parse a formula into an AST expression, raising a named error."""
+    text = (formula or "").strip()
+    if not text:
+        raise FormulaEvaluationError("formula is empty", measure=measure, formula=formula or "")
+    try:
+        return ast.parse(text, mode="eval").body
+    except SyntaxError as exc:
+        raise FormulaEvaluationError(
+            f"not valid arithmetic ({exc.msg}); {ALLOWED_SYNTAX_SUMMARY}",
+            measure=measure,
+            formula=text,
+        ) from exc
+
+
+def identifiers(formula: str) -> set[str]:
+    """Every identifier a formula references. Empty set if it does not parse."""
+    try:
+        tree = ast.parse((formula or "").strip() or "0", mode="eval")
+    except SyntaxError:
+        return set()
+    return {
+        node.id
+        for node in ast.walk(tree)
+        if isinstance(node, ast.Name) and not _is_function_slot(node, tree)
+    }
+
+
+def _is_function_slot(name_node: ast.Name, tree: ast.AST) -> bool:
+    """True when this Name is the callee of a call (a function, not a measure)."""
+    for node in ast.walk(tree):
+        if isinstance(node, ast.Call) and node.func is name_node:
+            return True
+    return False
+
+
+def ratio_pairs(formula: str) -> list[tuple[str, str]]:
+    """Every `A / B` where both sides are bare identifiers.
+
+    Used two ways: to infer `share_of` (the part must be a fraction of the
+    whole) and, in the quality gate, to check part <= whole in Snowflake.
+    Finds the pair inside any wrapper — `A / B`, `(A / B) * 100`,
+    `ROUND(A / B, 2)` all yield [(A, B)].
+    """
+    try:
+        tree = ast.parse((formula or "").strip() or "0", mode="eval")
+    except SyntaxError:
+        return []
+    pairs: list[tuple[str, str]] = []
+    for node in ast.walk(tree):
+        if (
+            isinstance(node, ast.BinOp)
+            and isinstance(node.op, ast.Div)
+            and isinstance(node.left, ast.Name)
+            and isinstance(node.right, ast.Name)
+        ):
+            pairs.append((node.left.id, node.right.id))
+    return pairs
+
+
+def validate(formula: str, allowed_identifiers: Iterable[str], *, measure: str = "") -> list[str]:
+    """Report every reason this formula cannot be computed. Empty list == valid.
+
+    Messages are written for two audiences: the LLM repair round (they go
+    straight into the repair prompt) and a human reading the run log.
+    """
+    problems: list[str] = []
+    try:
+        tree = parse(formula, measure=measure)
+    except FormulaEvaluationError as exc:
+        return [f"Measure {measure}: {exc.reason}." if measure else f"{exc.reason}."]
+
+    allowed = {str(name) for name in allowed_identifiers}
+    prefix = f"Measure {measure}: " if measure else ""
+
+    for node in ast.walk(tree):
+        if isinstance(node, ast.Call):
+            func = node.func.id.upper() if isinstance(node.func, ast.Name) else ""
+            if func not in ALLOWED_FUNCTIONS:
+                shown = func or "<expression>"
+                problems.append(
+                    f"{prefix}formula '{formula}' calls unsupported function {shown}() — "
+                    f"{ALLOWED_SYNTAX_SUMMARY}."
+                )
+                continue
+            low, high, _, _ = ALLOWED_FUNCTIONS[func]
+            argc = len(node.args)
+            if node.keywords or argc < low or (high is not None and argc > high):
+                problems.append(
+                    f"{prefix}formula '{formula}' calls {func}() with {argc} argument(s); "
+                    f"it takes {low}{'+' if high is None else f'..{high}'}."
+                )
+        elif isinstance(node, ast.Name):
+            if _is_function_slot(node, tree):
+                continue
+            if node.id == measure:
+                problems.append(f"{prefix}formula '{formula}' references itself.")
+            elif node.id not in allowed:
+                problems.append(
+                    f"{prefix}formula '{formula}' references '{node.id}', which is not a measure "
+                    f"in the same fact table. Every identifier in a formula must be another "
+                    f"measure of that fact."
+                )
+        elif isinstance(node, ast.BinOp):
+            if not isinstance(node.op, (ast.Add, ast.Sub, ast.Mult, ast.Div)):
+                problems.append(
+                    f"{prefix}formula '{formula}' uses an unsupported operator — "
+                    f"{ALLOWED_SYNTAX_SUMMARY}."
+                )
+        elif isinstance(node, ast.UnaryOp):
+            if not isinstance(node.op, ast.USub):
+                problems.append(
+                    f"{prefix}formula '{formula}' uses an unsupported unary operator — "
+                    f"{ALLOWED_SYNTAX_SUMMARY}."
+                )
+        elif isinstance(node, ast.Constant):
+            if not isinstance(node.value, (int, float)) or isinstance(node.value, bool):
+                problems.append(
+                    f"{prefix}formula '{formula}' contains the non-numeric literal "
+                    f"{node.value!r} — {ALLOWED_SYNTAX_SUMMARY}."
+                )
+        elif isinstance(node, (ast.Expression, ast.Load, ast.operator, ast.unaryop, ast.expr_context)):
+            # ast.walk() yields operator nodes (Add, Sub, Div...) as well as the
+            # BinOp/UnaryOp that owns them; the branches above already checked
+            # whether the operator is allowed.
+            continue
+        else:
+            problems.append(
+                f"{prefix}formula '{formula}' contains unsupported syntax "
+                f"({type(node).__name__}) — {ALLOWED_SYNTAX_SUMMARY}."
+            )
+    # dedupe, preserve order
+    seen: set[str] = set()
+    return [p for p in problems if not (p in seen or seen.add(p))]
+
+
+# --- evaluation -------------------------------------------------------------
+
+
+def evaluate(
+    formula: str,
+    values: Mapping[str, Any],
+    *,
+    measure: str = "",
+    warnings: MutableSequence[str] | None = None,
+) -> float:
+    """Evaluate a formula against measure values for ONE row.
+
+    Raises FormulaEvaluationError naming the measure, formula and offending
+    token for anything it cannot compute. It NEVER returns a made-up 0.0 —
+    that behavior shipped columns of zeros and fabricated metrics for months.
+
+    Division by zero returns 0.0 and appends a warning to `warnings` (a rate
+    with an empty denominator is legitimate; a silently wrong metric is not).
+    """
+    node = parse(formula, measure=measure)
+    return float(_eval_node(node, values, measure=measure, formula=formula, warnings=warnings))
+
+
+def _eval_node(
+    node: ast.AST,
+    values: Mapping[str, Any],
+    *,
+    measure: str,
+    formula: str,
+    warnings: MutableSequence[str] | None,
+) -> float:
+    def sub(child: ast.AST) -> float:
+        return _eval_node(child, values, measure=measure, formula=formula, warnings=warnings)
+
+    if isinstance(node, ast.BinOp):
+        op_type = type(node.op)
+        if op_type in _BIN_OPS:
+            return _BIN_OPS[op_type](sub(node.left), sub(node.right))
+        if op_type is ast.Div:
+            numerator = sub(node.left)
+            denominator = sub(node.right)
+            if not denominator:
+                if warnings is not None:
+                    warnings.append(
+                        f"{measure or 'formula'}: division by zero in '{formula}' — stored 0"
+                    )
+                return 0.0
+            return numerator / denominator
+        raise FormulaEvaluationError(
+            f"unsupported operator {op_type.__name__}; {ALLOWED_SYNTAX_SUMMARY}",
+            measure=measure, formula=formula, token=op_type.__name__,
+        )
+
+    if isinstance(node, ast.UnaryOp):
+        if isinstance(node.op, ast.USub):
+            return -sub(node.operand)
+        raise FormulaEvaluationError(
+            f"unsupported unary operator {type(node.op).__name__}; {ALLOWED_SYNTAX_SUMMARY}",
+            measure=measure, formula=formula, token=type(node.op).__name__,
+        )
+
+    if isinstance(node, ast.Call):
+        name = node.func.id.upper() if isinstance(node.func, ast.Name) else ""
+        spec = ALLOWED_FUNCTIONS.get(name)
+        if spec is None or node.keywords:
+            raise FormulaEvaluationError(
+                f"unsupported function {name or type(node.func).__name__}(); "
+                f"{ALLOWED_SYNTAX_SUMMARY}",
+                measure=measure, formula=formula, token=name or "call",
+            )
+        low, high, impl, _ = spec
+        args = [sub(arg) for arg in node.args]
+        if len(args) < low or (high is not None and len(args) > high):
+            raise FormulaEvaluationError(
+                f"{name}() takes {low}{'+' if high is None else f'..{high}'} arguments, got {len(args)}",
+                measure=measure, formula=formula, token=name,
+            )
+        result = float(impl(*args))
+        if math.isnan(result) or math.isinf(result):
+            raise FormulaEvaluationError(
+                f"{name}() produced a non-finite value", measure=measure, formula=formula, token=name
+            )
+        return result
+
+    if isinstance(node, ast.Constant):
+        if isinstance(node.value, bool) or not isinstance(node.value, (int, float)):
+            raise FormulaEvaluationError(
+                f"non-numeric literal {node.value!r}", measure=measure, formula=formula,
+                token=repr(node.value),
+            )
+        return float(node.value)
+
+    if isinstance(node, ast.Name):
+        if node.id in values:
+            return float(values[node.id])
+        upper = node.id.upper()
+        if upper in values:
+            return float(values[upper])
+        raise FormulaEvaluationError(
+            f"'{node.id}' is not a measure of this fact table, so it has no value for this row",
+            measure=measure, formula=formula, token=node.id,
+        )
+
+    raise FormulaEvaluationError(
+        f"unsupported syntax ({type(node).__name__}); {ALLOWED_SYNTAX_SUMMARY}",
+        measure=measure, formula=formula, token=type(node).__name__,
+    )
+
+
+# --- SQL compilation (used by the quality gate to recompute in place) -------
+
+
+def to_sql(formula: str, columns: Mapping[str, str], *, measure: str = "") -> str:
+    """Compile a formula to a Snowflake expression over `columns`.
+
+    columns maps UPPERCASE identifier -> real column name. Raises
+    FormulaEvaluationError if any identifier is not a column of the table, so
+    the caller can fall back to omitting the measure instead of writing SQL
+    that references something that does not exist.
+    """
+    node = parse(formula, measure=measure)
+    return _to_sql_node(node, columns, measure=measure, formula=formula)
+
+
+def _quote(identifier: str) -> str:
+    return '"' + identifier.replace('"', '""') + '"'
+
+
+def _to_sql_node(node: ast.AST, columns: Mapping[str, str], *, measure: str, formula: str) -> str:
+    def sub(child: ast.AST) -> str:
+        return _to_sql_node(child, columns, measure=measure, formula=formula)
+
+    if isinstance(node, ast.BinOp):
+        op = _SQL_BIN_OPS.get(type(node.op))
+        if op is None:
+            raise FormulaEvaluationError(
+                f"unsupported operator {type(node.op).__name__}", measure=measure, formula=formula,
+                token=type(node.op).__name__,
+            )
+        left, right = sub(node.left), sub(node.right)
+        if op == "/":
+            # Same semantics as the Python evaluator: /0 -> 0, never NULL.
+            return f"IFF(({right}) = 0, 0, ({left}) / NULLIF(({right}), 0))"
+        return f"({left} {op} {right})"
+
+    if isinstance(node, ast.UnaryOp) and isinstance(node.op, ast.USub):
+        return f"(-{sub(node.operand)})"
+
+    if isinstance(node, ast.Call):
+        name = node.func.id.upper() if isinstance(node.func, ast.Name) else ""
+        spec = ALLOWED_FUNCTIONS.get(name)
+        if spec is None or node.keywords:
+            raise FormulaEvaluationError(
+                f"unsupported function {name or 'call'}()", measure=measure, formula=formula,
+                token=name or "call",
+            )
+        low, high, _, sql_name = spec
+        args = [sub(a) for a in node.args]
+        if len(args) < low or (high is not None and len(args) > high):
+            raise FormulaEvaluationError(
+                f"{name}() takes {low}{'+' if high is None else f'..{high}'} arguments, got {len(args)}",
+                measure=measure, formula=formula, token=name,
+            )
+        return f"{sql_name}({', '.join(args)})"
+
+    if isinstance(node, ast.Constant) and isinstance(node.value, (int, float)) and not isinstance(node.value, bool):
+        return repr(float(node.value))
+
+    if isinstance(node, ast.Name):
+        actual = columns.get(node.id) or columns.get(node.id.upper())
+        if actual is None:
+            raise FormulaEvaluationError(
+                f"'{node.id}' is not a column of this table",
+                measure=measure, formula=formula, token=node.id,
+            )
+        return _quote(actual)
+
+    raise FormulaEvaluationError(
+        f"unsupported syntax ({type(node).__name__})", measure=measure, formula=formula,
+        token=type(node).__name__,
+    )
+
+
+# --- sign analysis -----------------------------------------------------------
+# Only ever consulted when a formula has to be DEMOTED to a sampled measure
+# (blueprint._repair_one_formula). The sample must keep the sign range the
+# formula would have had: `ACTUAL - FORECAST` is a signed error and sampling it
+# from a strictly-positive lognormal produces a "forecast error" that is never
+# negative — plausible-looking and wrong, which is the failure mode this whole
+# module exists to stop. `ABS(...)` is genuinely one-sided; a difference is not.
+
+_SIGN_PRESERVING_FUNCTIONS = frozenset(
+    {"ROUND", "MIN", "MAX", "LEAST", "GREATEST", "COALESCE", "IFNULL"}
+)
+
+
+def is_signed(formula: str) -> bool:
+    """True when `formula` could legitimately evaluate to a negative number."""
+    try:
+        return _node_is_signed(parse(formula))
+    except FormulaEvaluationError:
+        return False
+
+
+def _node_is_signed(node: ast.AST) -> bool:
+    if isinstance(node, ast.Call) and isinstance(node.func, ast.Name):
+        name = node.func.id.upper()
+        if name in _SIGN_PRESERVING_FUNCTIONS:
+            return any(_node_is_signed(arg) for arg in node.args)
+        return False        # ABS, and anything unrecognised, is non-negative
+    if isinstance(node, ast.BinOp):
+        if isinstance(node.op, ast.Sub):
+            return True
+        return _node_is_signed(node.left) or _node_is_signed(node.right)
+    if isinstance(node, ast.UnaryOp) and isinstance(node.op, ast.USub):
+        return True
+    return False

demoprep_app/dataset/quality.py

@@ -204,6 +204,7 @@ def validate_dataset_quality(dataset: DatasetBundle) -> DatasetQualityReport:
                 if family:
                     dimension_value_sets.setdefault((table.name, column, family), set()).add(text.lower())
 
+    issues.extend(_flat_measure_issues(dataset))
     issues.extend(_duplicate_dimension_set_issues(dimension_value_sets))
     issues.extend(_scenario_contract_issues(dataset, table_map))
     issues.extend(_metric_consistency_issues(dataset))
@@ -359,3 +360,68 @@ def _issue(severity: str, table: str, column: str, value: Any, message: str) ->
         value=str(value)[:255],
         message=message,
     )
+
+
+# A measure whose grouped averages differ by less than this across a dimension
+# cannot answer "which one is underperforming" — the question every one of these
+# boards exists to answer. The reviewer set 10-12% from graded boards; 0.10 is
+# the low end so the gate flags only genuinely dead measures.
+MIN_DIMENSION_SPREAD = 0.10
+
+
+def _flat_measure_issues(dataset: "DatasetBundle") -> list[DatasetQualityIssue]:
+    """Flag measures that are IDENTICAL across the dimensions a board slices by.
+
+    The existing checks catch an all-ZERO column. All-IDENTICAL is a different
+    failure and nearly as demo-fatal: mani.uppala's AdvancedMD build reported
+    success with zero warnings while NET_REVENUE varied 2.4% and COLLECTION_RATE
+    0.7% across providers, so "which providers are underperforming" — the entire
+    point of practice-management analytics — could not be told.
+
+    WARNING, not blocker, deliberately. A flat board is a weak demo; a blocked
+    build is no demo, and no demo is the worse outcome (the governing principle
+    behind the repair-not-veto gate). This tells the SE what is wrong while
+    still handing them something to show.
+    """
+    issues: list[DatasetQualityIssue] = []
+    for table in dataset.tables:
+        if not table.is_fact or not table.rows:
+            continue
+        dim_keys = [c.name for c in table.columns if c.semantic_role == "dimension_key"]
+        measures = [c.name for c in table.columns
+                    if c.semantic_role not in {"dimension_key", "date_key", "fact_key", "date"}
+                    and c.name not in dim_keys]
+        for measure in measures:
+            # Best spread this measure achieves across ANY dimension the board
+            # can group by. Flat on one dimension is fine — a measure priced per
+            # payer is legitimately flat by provider — but flat on ALL of them
+            # means it cannot carry a chart anywhere.
+            best = 0.0
+            best_dim = ""
+            for dim_key in dim_keys:
+                grouped: dict[Any, list[float]] = {}
+                for row in table.rows:
+                    value = _number(row.get(measure))
+                    if value is not None:
+                        grouped.setdefault(row.get(dim_key), []).append(value)
+                if len(grouped) < 2:
+                    continue
+                averages = [sum(v) / len(v) for v in grouped.values() if v]
+                if len(averages) < 2:
+                    continue
+                top = max(averages)
+                if not top:
+                    continue                      # all-zero: the other gate owns it
+                spread = (top - min(averages)) / abs(top)
+                if spread > best:
+                    best, best_dim = spread, dim_key
+            if best_dim and best < MIN_DIMENSION_SPREAD:
+                issues.append(_issue(
+                    "warning", table.name, measure, f"{best * 100:.1f}%",
+                    f"Measure is effectively identical across every dimension "
+                    f"(best spread {best * 100:.1f}% on {best_dim}, need "
+                    f"{MIN_DIMENSION_SPREAD * 100:.0f}%). No chart grouped by a "
+                    f"dimension can show a leader or a laggard, so this measure "
+                    f"cannot carry a story.",
+                ))
+    return issues

demoprep_app/dataset/validator.py

@@ -155,20 +155,36 @@ def _check_integrity(blueprint: DemoBlueprint, fact, tables: dict, fact_rows: li
 
 
 def _check_formulas(fact, fact_rows: list[dict[str, Any]], report: ValidationReport) -> None:
-    from demoprep_app.dataset.engine import _eval_formula  # single source of truth
+    # demoprep_app/dataset/formula.py is the single source of truth for the
+    # formula language — the same evaluator the engine used to produce the rows.
+    from demoprep_app.dataset.formula import FormulaEvaluationError, evaluate
 
     derived = [m for m in fact.measures if m.kind == "derived" and m.formula]
     for m in derived:
         sample = fact_rows[:: max(1, len(fact_rows) // 200)]  # ~200 spot checks
         misses = 0
         for row in sample:
-            expected = _eval_formula(m.formula, {k: float(v) for k, v in row.items() if isinstance(v, (int, float))})
+            try:
+                expected = evaluate(
+                    m.formula,
+                    {k: float(v) for k, v in row.items() if isinstance(v, (int, float))},
+                    measure=m.name,
+                )
+            except FormulaEvaluationError as exc:
+                # The engine could not have produced these rows from this
+                # formula either — name it once and move to the next measure
+                # instead of counting 200 identical failures.
+                report.formula_errors.append(str(exc))
+                break
             actual = float(row.get(m.name, 0.0))
             tolerance = max(1.0, abs(expected) * 0.02)
             if abs(actual - expected) > tolerance:
                 misses += 1
-        if misses > len(sample) * 0.05:
-            report.formula_errors.append(f"{m.name}: {misses}/{len(sample)} sampled rows deviate from formula '{m.formula}'.")
+        else:
+            if misses > len(sample) * 0.05:
+                report.formula_errors.append(
+                    f"{m.name}: {misses}/{len(sample)} sampled rows deviate from formula '{m.formula}'."
+                )
 
 
 def _key_lookup(blueprint: DemoBlueprint, tables: dict) -> dict[str, dict[int, str]]:

demoprep_app/integrations/snowflake/__init__.py

@@ -2,9 +2,18 @@
 
 from demoprep_app.integrations.snowflake.dataset_writer import populate_dataset_bundle
 from demoprep_app.integrations.snowflake.quality_gate import (
+    MeasureQualityGateError,
     ZeroMeasureError,
+    apply_measure_omissions,
+    measure_formulas,
     run_measure_quality_gate,
 )
 
-__all__ = ["populate_dataset_bundle", "run_measure_quality_gate", "ZeroMeasureError"]
-
+__all__ = [
+    "populate_dataset_bundle",
+    "run_measure_quality_gate",
+    "measure_formulas",
+    "apply_measure_omissions",
+    "MeasureQualityGateError",
+    "ZeroMeasureError",
+]

demoprep_app/integrations/snowflake/quality_gate.py

@@ -1,20 +1,54 @@
-"""Post-load data-quality gate for generated demo schemas.
-
-Guards against the recurring zero-measure defect: fact tables where derived
-measure columns (e.g. TOTAL_LOAD_REVENUE_USD, BROKER_MARGIN_USD) load as 100%
-zero/null while their input columns are populated, which renders liveboard
-tiles blank. Runs immediately after data population, against what was actually
-loaded into Snowflake, so it is independent of how the data was generated.
-
-The gate profiles every MEASURE column (numeric columns that are not join
-keys, mirroring ThoughtSpotDeployer._determine_column_type) and raises
-ZeroMeasureError when any measure is entirely zero/null. No silent fallback:
-a failed gate fails the run with the offending columns named.
+"""Post-load data-quality REPAIR stage for generated demo schemas.
+
+Runs immediately after data population, against what was actually loaded into
+Snowflake, so it is independent of how the data was generated.
+
+Why this is a repair stage and not a veto (2026-08-26)
+-----------------------------------------------------
+It used to profile measure columns, raise ZeroMeasureError on the first
+all-zero one, and abort the deployment. On 2026-08-26 that left an SE with a
+schema containing six tables and nothing else: no model, no liveboard, nothing
+to show. An empty schema is indistinguishable from the pipeline never having
+run.
+
+Governing principle (boone): producing NOTHING is the worst possible outcome.
+A visible error beats silence; a liveboard with one omitted or obviously broken
+tile beats no liveboard, because it is demo-able and the defect is fixable.
+Fail-open with a reported gap always beats fail-closed with nothing. This gate
+must never be the reason an SE has no demo.
+
+So the order of preference is:
+
+  1. REPAIR in place — recompute a derived column from its formula, clamp a
+     part to its whole, binarize a flag, clamp a rate to its scale.
+  2. OMIT — drop the measure from the table (and therefore from the model and
+     liveboard that are built from it afterwards) and report the omission.
+  3. HARD FAIL — only when there is no demo left at all: every measure the
+     brief's core claim depends on is unusable.
+
+Detection covers more than the all-zero symptom it used to (see
+dev_notes/BLOCKING_ISSUES_2026_08_26.md ISSUE A: the gate blocked on the benign
+symptom, all-zero, and waved through the dangerous one, a plausible-looking
+column that was exactly minus another column):
+
+  * all zero / null
+  * constant (MIN == MAX)
+  * exactly equal to +/- another column in the same table
+  * *_RATE / *_PCT outside its scale bound
+  * *_FLAG not in {0, 1}
+  * part > whole, for any part/whole pair implied by a ratio formula
 """
 
 from __future__ import annotations
 
-from typing import Any, Callable, Optional
+from typing import Any, Callable, Iterable, Mapping, Optional
+
+from demoprep_app.dataset.formula import (
+    FormulaEvaluationError,
+    identifiers as formula_identifiers,
+    ratio_pairs as formula_ratio_pairs,
+    to_sql as formula_to_sql,
+)
 
 # Snowflake normalizes numeric types, but accept the common aliases too.
 _NUMERIC_DATA_TYPES = {
@@ -27,22 +61,31 @@ _NUMERIC_DATA_TYPES = {
 # (a partial variant of the all-zero defect) but not an outright failure.
 MOSTLY_ZERO_WARN_THRESHOLD = 0.01
 
+# Suffixes that promise a bounded quantity.
+_RATE_SUFFIXES = ("_RATE", "_PCT", "_PERCENT", "_RATIO")
+_FLAG_SUFFIXES = ("_FLAG",)
 
-class ZeroMeasureError(Exception):
-    """One or more measure columns loaded entirely zero/null."""
+# Tolerance when comparing MIN/MAX signatures to shortlist mirrored columns.
+_MIRROR_EPS = 1e-9
 
-    def __init__(self, failures: list[dict], profile: dict):
+
+class MeasureQualityGateError(Exception):
+    """Nothing demo-able could be produced — the only hard failure left.
+
+    Raised when every measure the brief's core claim depends on is unusable, or
+    when the schema has no usable measure columns at all. Anything narrower is
+    repaired or omitted so the build continues.
+    """
+
+    def __init__(self, message: str, failures: list[dict], profile: dict):
         self.failures = failures
         self.profile = profile
-        details = "; ".join(
-            f"{f['table']}.{f['column']} ({f['row_count']} rows, all zero/null)"
-            for f in failures
-        )
-        super().__init__(
-            f"Data quality gate failed — {len(failures)} measure column(s) are "
-            f"entirely zero/null: {details}. The data generator did not compute "
-            f"these derived measures; the run must be regenerated."
-        )
+        super().__init__(message)
+
+
+# Legacy name: chat_interface and older tests catch this. Keeping it as an alias
+# means an existing `except ZeroMeasureError` still catches the hard failure.
+ZeroMeasureError = MeasureQualityGateError
 
 
 def is_measure_column(column_name: str, data_type: str) -> bool:
@@ -57,35 +100,434 @@ def is_measure_column(column_name: str, data_type: str) -> bool:
     return True
 
 
+def is_flag_column(column_name: str) -> bool:
+    return column_name.upper().endswith(_FLAG_SUFFIXES)
+
+
+def is_rate_column(column_name: str) -> bool:
+    return column_name.upper().endswith(_RATE_SUFFIXES)
+
+
+def rate_scale_bound(avg: float | None) -> float:
+    """Whether a rate column is on the 0-1 or the 0-100 scale.
+
+    The blueprint does not record the scale, so read it off the data: a genuine
+    0-100 percentage column averages well above 1. Known failure mode: a 0-100
+    column whose values are nearly all below 1% would be read as a 0-1 rate.
+    That is why an out-of-bounds base rate is clamped only when it is clearly
+    broken (see _repair_plan) rather than on this signal alone.
+    """
+    return 100.0 if (avg is not None and float(avg) > 1.5) else 1.0
+
+
+def measure_formulas(bundle: Any) -> dict[str, str]:
+    """{"TABLE.COLUMN": "FORMULA"} for every derived measure in a DatasetBundle.
+
+    The formulas are what make repair possible: without them a bad column can
+    only be omitted.
+    """
+    formulas: dict[str, str] = {}
+    for table in getattr(bundle, "tables", []) or []:
+        for column in table.columns:
+            if getattr(column, "formula", None):
+                formulas[f"{table.name}.{column.name}"] = column.formula
+    return formulas
+
+
+def apply_measure_omissions(bundle: Any, omissions: Iterable[Mapping[str, str]]) -> list[str]:
+    """Remove omitted measure columns from an in-memory DatasetBundle.
+
+    The ThoughtSpot side of the deploy is driven by DDL compiled from this
+    bundle, so a column dropped in Snowflake has to disappear here too or the
+    model would declare a column that no longer exists.
+    """
+    removed: list[str] = []
+    by_name = {table.name: table for table in getattr(bundle, "tables", []) or []}
+    for omission in omissions:
+        table = by_name.get(omission.get("table", ""))
+        column = omission.get("column", "")
+        if table is None or not column:
+            continue
+        before = len(table.columns)
+        table.columns = [c for c in table.columns if c.name != column]
+        if len(table.columns) != before:
+            removed.append(f"{table.name}.{column}")
+        for row in table.rows:
+            row.pop(column, None)
+    return removed
+
+
 def _quote_ident(identifier: str) -> str:
     escaped = identifier.replace('"', '""')
     return f'"{escaped}"'
 
 
+def _table_ref(schema_name: str, table_name: str) -> str:
+    return f"{_quote_ident(schema_name)}.{_quote_ident(table_name)}"
+
+
+# ---------------------------------------------------------------------------
+# Profiling
+# ---------------------------------------------------------------------------
+
+
+def _profile_table(cursor: Any, schema_name: str, table_name: str, columns: list[str]) -> dict:
+    """One aggregate query per table: row count plus stats for every measure."""
+    select_parts = ["COUNT(*)"]
+    index_of: dict[str, dict[str, int]] = {}
+    for col in columns:
+        q = _quote_ident(col)
+        slots = {}
+        slots["non_zero"] = len(select_parts)
+        select_parts.append(f"SUM(IFF({q} IS NOT NULL AND {q} <> 0, 1, 0))")
+        slots["min"] = len(select_parts)
+        select_parts.append(f"MIN({q})")
+        slots["max"] = len(select_parts)
+        select_parts.append(f"MAX({q})")
+        slots["avg"] = len(select_parts)
+        select_parts.append(f"AVG({q})")
+        if is_flag_column(col):
+            slots["non_binary"] = len(select_parts)
+            select_parts.append(f"SUM(IFF({q} IS NOT NULL AND {q} NOT IN (0, 1), 1, 0))")
+        index_of[col] = slots
+
+    cursor.execute(f"SELECT {', '.join(select_parts)} FROM {_table_ref(schema_name, table_name)}")
+    result = cursor.fetchone()
+    row_count = int(result[0] or 0)
+
+    stats: dict[str, dict] = {}
+    for col, slots in index_of.items():
+        def value(key: str):
+            idx = slots.get(key)
+            return None if idx is None else result[idx]
+
+        stats[col] = {
+            "row_count": row_count,
+            "non_zero": int(value("non_zero") or 0),
+            "min": None if value("min") is None else float(value("min")),
+            "max": None if value("max") is None else float(value("max")),
+            "avg": None if value("avg") is None else float(value("avg")),
+            "non_binary": None if value("non_binary") is None else int(value("non_binary") or 0),
+        }
+    return {"row_count": row_count, "columns": stats}
+
+
+def _mirror_defects(
+    cursor: Any,
+    schema_name: str,
+    table_name: str,
+    stats: dict[str, dict],
+    formulas: Mapping[str, str],
+) -> list[dict]:
+    """Columns that are exactly +/- another column of the same table.
+
+    This is the defect that used to ship silently: a formula referencing a
+    measure that does not exist in the fact resolved that identifier to 0, so
+    `ACTUAL - FORECAST` stored `-FORECAST` in all 5000 rows and looked
+    plausible. Shortlist on the MIN/MAX signature (cheap, from the profile we
+    already have), then verify row by row in ONE query.
+    """
+    # An all-zero column is already reported as all_zero, and two all-zero
+    # columns trivially "mirror" each other — skip them to keep the signal clean.
+    names = [
+        c for c, s in stats.items()
+        if s["row_count"] and s["min"] is not None and s["non_zero"]
+    ]
+    candidates: list[tuple[str, str, int]] = []   # (col, other, sign)
+    for col in names:
+        formula = formulas.get(f"{table_name}.{col}")
+        if not formula:
+            continue    # only a derived column can be "wrong" in this specific way
+        col_min, col_max = stats[col]["min"], stats[col]["max"]
+        for other in names:
+            if other == col:
+                continue
+            # A formula that IS just the other column (or its negation) is
+            # legitimately identical — don't call that a defect.
+            idents = formula_identifiers(formula)
+            if idents == {other} and formula.replace(" ", "").lstrip("-") == other:
+                continue
+            other_min, other_max = stats[other]["min"], stats[other]["max"]
+            if abs(col_min - other_min) <= _MIRROR_EPS and abs(col_max - other_max) <= _MIRROR_EPS:
+                candidates.append((col, other, 1))
+            elif abs(col_min + other_max) <= _MIRROR_EPS and abs(col_max + other_min) <= _MIRROR_EPS:
+                candidates.append((col, other, -1))
+
+    if not candidates:
+        return []
+
+    parts = []
+    for col, other, sign in candidates:
+        left, right = _quote_ident(col), _quote_ident(other)
+        expr = right if sign > 0 else f"-{right}"
+        parts.append(f"SUM(IFF({left} IS DISTINCT FROM {expr}, 1, 0))")
+    cursor.execute(f"SELECT {', '.join(parts)} FROM {_table_ref(schema_name, table_name)}")
+    row = cursor.fetchone()
+
+    defects: list[dict] = []
+    for idx, (col, other, sign) in enumerate(candidates):
+        differing = int(row[idx] or 0)
+        if differing == 0:
+            defects.append({
+                "table": table_name,
+                "column": col,
+                "defect": "mirrors_other_column",
+                "detail": (
+                    f"every row equals {'-' if sign < 0 else ''}{other} — the formula's other "
+                    f"input resolved to nothing"
+                ),
+            })
+    return defects
+
+
+def _part_whole_defects(
+    cursor: Any,
+    schema_name: str,
+    table_name: str,
+    stats: dict[str, dict],
+    formulas: Mapping[str, str],
+) -> list[dict]:
+    """part > whole for every part/whole pair implied by a ratio formula."""
+    pairs: list[tuple[str, str, str]] = []   # (part, whole, ratio_column)
+    for key, formula in formulas.items():
+        tbl, _, ratio_col = key.partition(".")
+        if tbl != table_name:
+            continue
+        for part, whole in formula_ratio_pairs(formula):
+            if part in stats and whole in stats and (part, whole, ratio_col) not in pairs:
+                pairs.append((part, whole, ratio_col))
+    if not pairs:
+        return []
+
+    parts_sql = [
+        f"SUM(IFF({_quote_ident(part)} > {_quote_ident(whole)}, 1, 0))"
+        for part, whole, _ in pairs
+    ]
+    cursor.execute(f"SELECT {', '.join(parts_sql)} FROM {_table_ref(schema_name, table_name)}")
+    row = cursor.fetchone()
+
+    defects: list[dict] = []
+    for idx, (part, whole, ratio_col) in enumerate(pairs):
+        offending = int(row[idx] or 0)
+        if offending:
+            defects.append({
+                "table": table_name,
+                "column": part,
+                "defect": "part_exceeds_whole",
+                "detail": f"{offending} row(s) where {part} > {whole} (used by {ratio_col})",
+                "whole": whole,
+                "ratio_column": ratio_col,
+            })
+    return defects
+
+
+def _column_defects(table_name: str, stats: dict[str, dict]) -> tuple[list[dict], list[str]]:
+    """Per-column defects and warnings derivable from the profile alone."""
+    defects: list[dict] = []
+    warnings: list[str] = []
+    for col, s in stats.items():
+        row_count, non_zero = s["row_count"], s["non_zero"]
+        if not row_count:
+            continue
+        if non_zero == 0:
+            defects.append({
+                "table": table_name, "column": col, "defect": "all_zero",
+                "detail": f"{row_count} rows, all zero/null",
+            })
+            continue
+        if s["min"] is not None and s["max"] is not None and s["min"] == s["max"] and row_count > 1:
+            defects.append({
+                "table": table_name, "column": col, "defect": "constant",
+                "detail": f"every row is {s['min']}",
+            })
+        if is_flag_column(col) and s.get("non_binary"):
+            defects.append({
+                "table": table_name, "column": col, "defect": "flag_not_binary",
+                "detail": f"{s['non_binary']}/{row_count} rows outside {{0,1}} (max {s['max']})",
+            })
+        if is_rate_column(col) and s["max"] is not None:
+            bound = rate_scale_bound(s["avg"])
+            if s["max"] > bound or (s["min"] is not None and s["min"] < -bound):
+                defects.append({
+                    "table": table_name, "column": col, "defect": "rate_out_of_bounds",
+                    "detail": (
+                        f"min={s['min']} max={s['max']} avg={s['avg']} outside +/-{bound:g} "
+                        f"for a rate on that scale"
+                    ),
+                    "bound": bound,
+                })
+        if non_zero / row_count < MOSTLY_ZERO_WARN_THRESHOLD:
+            pct_zero = 100.0 * (1 - non_zero / row_count)
+            warnings.append(
+                f"{table_name}.{col} is {pct_zero:.1f}% zero/null ({non_zero}/{row_count} non-zero)"
+            )
+    return defects, warnings
+
+
+# ---------------------------------------------------------------------------
+# Repair
+# ---------------------------------------------------------------------------
+
+
+def _repair_statements(
+    schema_name: str,
+    table_name: str,
+    defect: dict,
+    columns: list[str],
+    formulas: Mapping[str, str],
+    *,
+    allow_clamp: bool = True,
+) -> list[tuple[str, str]]:
+    """(description, SQL) repairs for one defect, most faithful first.
+
+    Every statement here targets a table this run created seconds earlier in a
+    throwaway demo schema. Nothing outside the generated schema is touched.
+    """
+    column = defect["column"]
+    ref = _table_ref(schema_name, table_name)
+    q = _quote_ident(column)
+    column_map = {c.upper(): c for c in columns}
+    formula = formulas.get(f"{table_name}.{column}")
+
+    def recompute() -> list[tuple[str, str]]:
+        if not formula:
+            return []
+        try:
+            expr = formula_to_sql(formula, column_map, measure=column)
+        except FormulaEvaluationError:
+            return []   # inputs are not in this table — recompute is impossible
+        return [(f"recomputed {table_name}.{column} from '{formula}'", f"UPDATE {ref} SET {q} = {expr}")]
+
+    if defect["defect"] == "part_exceeds_whole":
+        whole = _quote_ident(defect["whole"])
+        repairs = [(
+            f"clamped {table_name}.{column} to {defect['whole']} so the part cannot exceed the whole",
+            f"UPDATE {ref} SET {q} = LEAST({q}, {whole}) WHERE {q} > {whole}",
+        )]
+        ratio_formula = formulas.get(f"{table_name}.{defect.get('ratio_column', '')}")
+        if ratio_formula:
+            ratio_col = defect["ratio_column"]
+            try:
+                expr = formula_to_sql(ratio_formula, column_map, measure=ratio_col)
+            except FormulaEvaluationError:
+                return repairs
+            repairs.append((
+                f"recomputed {table_name}.{ratio_col} from '{ratio_formula}' after clamping",
+                f"UPDATE {ref} SET {_quote_ident(ratio_col)} = {expr}",
+            ))
+        return repairs
+
+    if defect["defect"] == "flag_not_binary":
+        return [(
+            f"binarized {table_name}.{column} to 0/1",
+            f"UPDATE {ref} SET {q} = IFF({q} >= 0.5, 1, 0)",
+        )]
+
+    if defect["defect"] == "rate_out_of_bounds":
+        # Recomputing from the formula is always safe and always preferred.
+        recomputed = recompute()
+        if recomputed:
+            return recomputed
+        if not allow_clamp:
+            # Clamping is destructive if the scale was misread, so the caller
+            # only allows it when the column is broken beyond any reading.
+            return []
+        bound = float(defect.get("bound", 1.0))
+        return [(
+            f"clamped {table_name}.{column} into [0, {bound:g}]",
+            f"UPDATE {ref} SET {q} = LEAST(GREATEST({q}, 0), {bound})",
+        )]
+
+    # all_zero / constant / mirrors_other_column: only a formula can fix these.
+    return recompute()
+
+
+# Repairs run in this order: fix the INPUTS (clamp a part to its whole,
+# binarize a flag) before recomputing anything derived from them, otherwise a
+# recomputed ratio is computed from values that are about to change.
+_REPAIR_ORDER = {"part_exceeds_whole": 0, "flag_not_binary": 1}
+
+
+def _repair_rank(defect: dict) -> int:
+    return _REPAIR_ORDER.get(defect["defect"], 2)
+
+
+# Which unrepaired defects justify removing the measure from the demo.
+# A column that is blank or actively lying is worse than a missing tile; a
+# column that is merely odd (an out-of-scale rate, a part slightly over its
+# whole) is better shown, and reported, than deleted.
+_OMIT_WORTHY = {"all_zero", "mirrors_other_column"}
+
+
+def _should_omit(defect: dict, formulas: Mapping[str, str]) -> bool:
+    kind = defect["defect"]
+    if kind in _OMIT_WORTHY:
+        return True
+    if kind == "constant":
+        # A derived constant means its formula is broken. A base constant may
+        # simply be a flat attribute — keep it and say so.
+        return bool(formulas.get(f"{defect['table']}.{defect['column']}"))
+    return False
+
+
+def _is_clamp_worth_it(defect: dict, stats: dict) -> bool:
+    """Guard for the one destructive repair whose trigger is a heuristic."""
+    if defect["defect"] != "rate_out_of_bounds":
+        return True
+    bound = float(defect.get("bound", 1.0))
+    maximum = stats.get("max")
+    return maximum is not None and (maximum > 100.0 or maximum > bound * 2)
+
+
+# ---------------------------------------------------------------------------
+# Entry point
+# ---------------------------------------------------------------------------
+
+
 def run_measure_quality_gate(
     connection: Any,
     schema_name: str,
     progress_callback: Optional[Callable[[str], None]] = None,
+    *,
+    formulas: Optional[Mapping[str, str]] = None,
+    required_measures: Optional[Iterable[str]] = None,
+    repair: bool = True,
+    drop_unrepairable: bool = True,
 ) -> dict:
-    """Profile every measure column in the schema; fail loudly on all-zero ones.
+    """Profile every measure column in the schema; repair, omit, or (rarely) fail.
 
-    Returns a profile dict:
+    formulas: {"TABLE.COLUMN": "FORMULA"} for derived measures — see
+        measure_formulas(bundle). Without them nothing can be recomputed and a
+        bad column can only be omitted.
+    required_measures: measure names (bare or "TABLE.COLUMN") the brief's core
+        claim depends on. The gate hard-fails only if EVERY one of them ends up
+        unusable — one missing tile is not worth denying an SE a demo.
+    repair / drop_unrepairable: switches for tests and dry runs.
+
+    Returns:
         {
-            "tables_checked": int,
-            "measure_columns_checked": int,
-            "failures": [],  # empty on success (non-empty raises instead)
-            "warnings": ["FACT_X.COL is 99.8% zero/null (1/500 non-zero)", ...],
-            "columns": {"TABLE.COLUMN": {"row_count": n, "non_zero": m}, ...},
+          "tables_checked": int,
+          "measure_columns_checked": int,
+          "failures": [ ... unrepaired defects that were NOT dropped ... ],
+          "defects": [ ... everything detected, with defect type + detail ... ],
+          "repairs": ["recomputed FACT.X from '...'", ...],
+          "omissions": [{"table":..., "column":..., "defect":..., "detail":...}],
+          "warnings": [...],
+          "columns": {"TABLE.COLUMN": {"row_count":n, "non_zero":m, "min":..,
+                                       "max":.., "avg":..}},
         }
 
-    Raises ZeroMeasureError when any measure column in a populated table is
-    entirely zero/null. Profiling errors (bad schema name, connection loss)
-    propagate as-is — the gate never passes silently.
+    Raises MeasureQualityGateError only when there is nothing demo-able left.
+    Profiling errors (bad schema name, connection loss) propagate as-is — the
+    gate never passes silently.
     """
     def log(msg: str) -> None:
         if progress_callback:
             progress_callback(msg)
 
+    formulas = dict(formulas or {})
+    required = {str(m).upper() for m in (required_measures or ())}
+
     cursor = connection.cursor()
     try:
         cursor.execute(
@@ -113,57 +555,171 @@ def run_measure_quality_gate(
             "tables_checked": 0,
             "measure_columns_checked": 0,
             "failures": [],
+            "defects": [],
+            "repairs": [],
+            "omissions": [],
             "warnings": [],
             "columns": {},
         }
 
+        stats_by_table: dict[str, dict] = {}
         for table_name, columns in sorted(measures_by_table.items()):
-            select_parts = ["COUNT(*)"]
-            for col in columns:
-                q = _quote_ident(col)
-                select_parts.append(f"SUM(IFF({q} IS NOT NULL AND {q} <> 0, 1, 0))")
-            cursor.execute(
-                f"SELECT {', '.join(select_parts)} "
-                f"FROM {_quote_ident(schema_name)}.{_quote_ident(table_name)}"
-            )
-            result = cursor.fetchone()
-            row_count = int(result[0] or 0)
-
+            table_profile = _profile_table(cursor, schema_name, table_name, columns)
+            stats_by_table[table_name] = table_profile
             profile["tables_checked"] += 1
-            if row_count == 0:
+            if not table_profile["row_count"]:
                 profile["warnings"].append(
                     f"{table_name} has 0 rows — measure columns could not be verified"
                 )
                 continue
-
-            for idx, col in enumerate(columns):
-                non_zero = int(result[idx + 1] or 0)
+            for col, stats in table_profile["columns"].items():
                 profile["measure_columns_checked"] += 1
                 profile["columns"][f"{table_name}.{col}"] = {
-                    "row_count": row_count,
-                    "non_zero": non_zero,
+                    "row_count": stats["row_count"],
+                    "non_zero": stats["non_zero"],
+                    "min": stats["min"],
+                    "max": stats["max"],
+                    "avg": stats["avg"],
                 }
-                if non_zero == 0:
-                    profile["failures"].append(
-                        {"table": table_name, "column": col, "row_count": row_count}
-                    )
-                elif non_zero / row_count < MOSTLY_ZERO_WARN_THRESHOLD:
-                    pct_zero = 100.0 * (1 - non_zero / row_count)
-                    profile["warnings"].append(
-                        f"{table_name}.{col} is {pct_zero:.1f}% zero/null "
-                        f"({non_zero}/{row_count} non-zero)"
-                    )
+            column_defects, column_warnings = _column_defects(table_name, table_profile["columns"])
+            profile["defects"].extend(column_defects)
+            profile["warnings"].extend(column_warnings)
+            profile["defects"].extend(
+                _part_whole_defects(cursor, schema_name, table_name, table_profile["columns"], formulas)
+            )
+            profile["defects"].extend(
+                _mirror_defects(cursor, schema_name, table_name, table_profile["columns"], formulas)
+            )
 
         log(
             f"Data quality gate: profiled {profile['measure_columns_checked']} measure "
             f"column(s) across {profile['tables_checked']} table(s)"
         )
+        for defect in profile["defects"]:
+            log(f"[DEFECT] {defect['table']}.{defect['column']}: {defect['defect']} — {defect['detail']}")
         for warning in profile["warnings"]:
             log(f"[WARN] Data quality: {warning}")
 
-        if profile["failures"]:
-            raise ZeroMeasureError(profile["failures"], profile)
+        if not profile["defects"]:
+            return profile
+
+        # --- repair pass -------------------------------------------------
+        repaired_tables: set[str] = set()
+        executed_repairs: set[str] = set()
+        if repair:
+            for defect in sorted(profile["defects"], key=_repair_rank):
+                table_name = defect["table"]
+                stats = stats_by_table[table_name]["columns"].get(defect["column"], {})
+                allow_clamp = _is_clamp_worth_it(defect, stats)
+                statements = _repair_statements(
+                    schema_name, table_name, defect,
+                    measures_by_table.get(table_name, []), formulas,
+                    allow_clamp=allow_clamp,
+                )
+                if not statements and not allow_clamp:
+                    profile["warnings"].append(
+                        f"{table_name}.{defect['column']}: {defect['detail']} — left as-is; "
+                        "clamping could be wrong if this column is on the 0-100 scale"
+                    )
+                    continue
+                for description, sql in statements:
+                    if sql in executed_repairs:
+                        continue    # a column can carry two defects with one fix
+                    executed_repairs.add(sql)
+                    cursor.execute(sql)
+                    profile["repairs"].append(description)
+                    repaired_tables.add(table_name)
+                    log(f"[REPAIR] {description}")
+            if repaired_tables:
+                connection.commit()
+
+        # --- re-profile only what we touched -----------------------------
+        remaining: list[dict] = []
+        for defect in profile["defects"]:
+            if defect["table"] not in repaired_tables:
+                remaining.append(defect)
+        for table_name in sorted(repaired_tables):
+            columns = measures_by_table.get(table_name, [])
+            table_profile = _profile_table(cursor, schema_name, table_name, columns)
+            stats_by_table[table_name] = table_profile
+            for col, stats in table_profile["columns"].items():
+                profile["columns"][f"{table_name}.{col}"] = {
+                    "row_count": stats["row_count"],
+                    "non_zero": stats["non_zero"],
+                    "min": stats["min"],
+                    "max": stats["max"],
+                    "avg": stats["avg"],
+                }
+            column_defects, _ = _column_defects(table_name, table_profile["columns"])
+            remaining.extend(column_defects)
+            remaining.extend(
+                _part_whole_defects(cursor, schema_name, table_name, table_profile["columns"], formulas)
+            )
+            remaining.extend(
+                _mirror_defects(cursor, schema_name, table_name, table_profile["columns"], formulas)
+            )
 
+        if not remaining:
+            log("Data quality gate: all detected defects repaired in place")
+            return profile
+
+        # --- decide the hard-fail BEFORE dropping anything ---------------
+        to_omit = [d for d in remaining if _should_omit(d, formulas)]
+        for defect in remaining:
+            if defect in to_omit:
+                continue
+            profile["failures"].append(defect)
+            profile["warnings"].append(
+                f"{defect['table']}.{defect['column']}: {defect['defect']} — {defect['detail']}. "
+                "Kept on the board (visible and fixable beats missing)"
+            )
+        unusable = {f"{d['table']}.{d['column']}".upper() for d in to_omit}
+        unusable |= {d["column"].upper() for d in to_omit}
+        if required and required <= unusable:
+            raise MeasureQualityGateError(
+                "Data quality gate: every measure this demo's story depends on is unusable "
+                f"({', '.join(sorted(required))}). Nothing demo-able could be produced. "
+                f"Schema {schema_name} is left in place for inspection.",
+                to_omit, profile,
+            )
+
+        # --- omit what could not be repaired -----------------------------
+        for defect in to_omit:
+            table_name, column = defect["table"], defect["column"]
+            surviving = [
+                c for c in measures_by_table.get(table_name, [])
+                if c != column and f"{table_name}.{c}".upper() not in
+                {f"{o['table']}.{o['column']}".upper() for o in profile["omissions"]}
+            ]
+            if not drop_unrepairable or not surviving:
+                # Never strip a table's last measure: a fact with no measures
+                # breaks the model outright, which is worse than a bad column.
+                reason = (
+                    "dropping it would leave the table with no measures"
+                    if surviving == [] else "column removal is disabled for this run"
+                )
+                profile["failures"].append(defect)
+                profile["warnings"].append(
+                    f"{table_name}.{column}: {defect['defect']} could not be repaired and was "
+                    f"KEPT ({reason})"
+                )
+                continue
+            cursor.execute(
+                f"ALTER TABLE {_table_ref(schema_name, table_name)} DROP COLUMN {_quote_ident(column)}"
+            )
+            profile["omissions"].append(defect)
+            log(
+                f"[OMIT] dropped {table_name}.{column} ({defect['defect']}: {defect['detail']}) — "
+                "the demo ships without this measure"
+            )
+        if profile["omissions"]:
+            connection.commit()
+
+        if not profile["columns"]:
+            raise MeasureQualityGateError(
+                f"Data quality gate: schema {schema_name} has no usable measure columns at all.",
+                remaining, profile,
+            )
         return profile
     finally:
         cursor.close()

demoprep_app/liveboards/blueprint.py

@@ -21,8 +21,17 @@ def build_liveboard_blueprint(
     use_case: str,
     model_columns: List[Dict[str, Any]] | None = None,
     target_filter_count: int = 3,
+    include_demo_notes_tile: bool = True,
 ) -> Dict[str, Any]:
-    """Build a small liveboard blueprint from model metadata and story context."""
+    """Build a small liveboard blueprint from model metadata and story context.
+
+    include_demo_notes_tile gates the "Demo Guide" tile (ISS-5). Golden-demo
+    style organises a board with GROUPS, and the tile renders as an awkward
+    loose block — worse now that it is appended AFTER the layout pass, so it
+    lands ungrouped on whichever tab happens to be last. Callers on the golden
+    path pass False. Kept rather than deleted so the standalone/portal callers
+    that want it can still ask.
+    """
     model_columns = model_columns or []
     filters = choose_global_filters(model_columns, target_filter_count=target_filter_count)
     return {
@@ -30,7 +39,10 @@ def build_liveboard_blueprint(
         "use_case": use_case or company_data.get("use_case", "Analytics"),
         "app_name": company_data.get("app_name", DEFAULT_APP_NAME),
         "filters": filters,
-        "demo_notes_tile": build_demo_notes_spec(company_data, use_case, filters),
+        "demo_notes_tile": (
+            build_demo_notes_spec(company_data, use_case, filters)
+            if include_demo_notes_tile else {}
+        ),
         "validation_rules": {
             "filter_count_min": 2,
             "filter_count_max": 6,
@@ -218,12 +230,36 @@ def _resolve_demo_guide_id(visualizations: List[Dict[str, Any]], requested_id: s
 
 
 def _place_demo_tile_at_bottom(liveboard: Dict[str, Any], demo_id: str) -> None:
+    """Pin the Demo Guide tile to the bottom of the SHOWCASE tab.
+
+    boone's ruling 2026-09-02 (closes ISS-5): the tile MOVES to the third page —
+    not removed, not gated. The third page is the ungraded experiment zone, so a
+    text tile there does not violate golden-demo style. If the layout has no
+    showcase tab yet (a board with no tables or custom charts), create it: the
+    tile must never fall back onto the executive or manager tab.
+    """
+    from demoprep_app.liveboards.golden_layout import SHOWCASE_TAB
+
     tabs = liveboard.setdefault("layout", {}).setdefault("tabs", [])
     if not tabs:
-        tabs.append({"name": "Executive Overview", "tiles": []})
-    target_tab = tabs[-1]
+        # A board with no layout yet must still lead with the executive tab; the
+        # showcase tab is appended below. Never let "Random Charts" be page one.
+        tabs.append({"name": "Executive Overview", "tiles": [], "group_layouts": []})
+    # Remove any previous placement wherever it was.
+    for tab in tabs:
+        tiles = tab.get("tiles") or []
+        tiles[:] = [tile for tile in tiles if tile.get("visualization_id") != demo_id]
+
+    target_tab = next((tab for tab in tabs if tab.get("name") == SHOWCASE_TAB), None)
+    if target_tab is None:
+        target_tab = {
+            "name": SHOWCASE_TAB,
+            "description": "Custom visuals, exploratory charts, and the demo guide",
+            "tiles": [],
+            "group_layouts": [],
+        }
+        tabs.append(target_tab)
     tiles = target_tab.setdefault("tiles", [])
-    tiles[:] = [tile for tile in tiles if tile.get("visualization_id") != demo_id]
     y = _next_tile_y(tiles)
     tiles.append({"visualization_id": demo_id, "x": 0, "y": y, "height": 7, "width": 12})
 

demoprep_app/liveboards/golden_layout.py

@@ -14,6 +14,9 @@ from typing import Any, Dict, Iterable, List, Tuple
 
 EXECUTIVE_TAB = "Executive Overview"
 OPERATIONS_TAB = "Manager Drilldown"
+# Third tab: the grab-bag. Custom/Muze charts, the one sanctioned table, and any
+# chart too interesting to cut but too odd for the first two tabs.
+SHOWCASE_TAB = "Random Charts"
 
 
 def build_dashboard_plan(
@@ -31,8 +34,11 @@ def build_dashboard_plan(
     and lead with the KPI row. Set True to retain a single branded overview tile.
     """
     target_viz_count = max(8, min(int(target_viz_count or 16), 24))
-    kpi_count = 4 if target_viz_count >= 14 else 3
+    # Four KPIs across the top is THE standard (both golden reference boards),
+    # independent of how many visualizations the board has.
+    kpi_count = KPI_STANDARD_COUNT
     executive_tab, manager_tab = _tab_names_for_use_case(use_case)
+    showcase_tab = SHOWCASE_TAB
     return {
         "company_name": company_name,
         "use_case": use_case,
@@ -42,6 +48,7 @@ def build_dashboard_plan(
         "model_columns": model_columns,
         "executive_tab": executive_tab,
         "manager_tab": manager_tab,
+        "showcase_tab": showcase_tab,
         "target_viz_count": target_viz_count,
         "audiences": [
             {
@@ -54,6 +61,11 @@ def build_dashboard_plan(
                 "tab": manager_tab,
                 "intent": "Operator-level breakdowns for action, ownership, and follow-up.",
             },
+            {
+                "name": "Analyst",
+                "tab": showcase_tab,
+                "intent": "Exploratory and custom visuals, plus row-level detail.",
+            },
         ],
         "sections": [
             {
@@ -98,6 +110,12 @@ def build_dashboard_plan(
                 "role": "detail",
                 "target_count": 2,
             },
+            {
+                "name": "Chart Showcase",
+                "tab": showcase_tab,
+                "role": "showcase",
+                "target_count": SHOWCASE_TAB_TARGET,
+            },
         ],
         "style": {
             "border": "CURVED",
@@ -160,17 +178,39 @@ def apply_golden_two_tab_layout(
     # Auto-build a geo map when the model has coordinates and the board has none.
     geo_viz = _maybe_build_geo_viz(data_visualizations, plan.get("model_columns"))
     geo_change = ""
+    if geo_viz is None:
+        untagged = untagged_geo_candidates(plan.get("model_columns"))
+        if untagged:
+            geo_change = (
+                "No geo map: " + ", ".join(untagged[:3]) + " look geographic but "
+                "carry no geo_config in the model, and ThoughtSpot renders an "
+                "untagged GEO_* chart as a blank map. Fix the model, not the board."
+            )
     if geo_viz is not None:
         data_visualizations.append(geo_viz)
         visualizations.append(geo_viz)
         liveboard["visualizations"] = visualizations
         geo_change = f"Built geo map: {geo_viz['answer']['name']}"
 
+    # One detail table for the showcase tab, built the same way as the geo map.
+    table_viz = _maybe_build_detail_table(data_visualizations, plan.get("model_columns"))
+    table_change = ""
+    if table_viz is not None:
+        data_visualizations.append(table_viz)
+        visualizations.append(table_viz)
+        liveboard["visualizations"] = visualizations
+        table_change = f"Built detail table: {table_viz['answer']['name']}"
+
     kpi_changes = _convert_periodic_metric_charts(data_visualizations)
+    sparkline_changes = _fix_kpi_sparklines(data_visualizations)
+    muze_changes = _muze_upgrade_surplus_heatmaps(data_visualizations)
     chart_changes = _reduce_heatmaps_and_apply_palettes(data_visualizations)
+    # After every type mutation above, and covering types MCP chose itself.
+    axis_changes = repair_secondary_axis_mismatch(data_visualizations)
     mode_changes = _force_chart_mode(data_visualizations)
     color_changes = _apply_color_by_dimension(data_visualizations)
-    groups = _build_groups(data_visualizations, plan)
+    group_notes: List[str] = []
+    groups = _build_groups(data_visualizations, plan, notes=group_notes)
     # Cosmetic title cleanup AFTER classification so roles use the original names.
     title_changes = _clean_viz_titles(data_visualizations)
     if note_viz:
@@ -184,13 +224,19 @@ def apply_golden_two_tab_layout(
         f"Applied golden two-tab layout ({len(groups)} groups)",
         f"Tabs: {_executive_tab_name(plan)}, {_manager_tab_name(plan)}",
     ]
-    for change in (geo_change, kpi_changes, chart_changes, mode_changes, color_changes, title_changes):
+    for change in (geo_change, table_change, kpi_changes, sparkline_changes, chart_changes,
+                   axis_changes, muze_changes, mode_changes, color_changes,
+                   title_changes, *group_notes):
         if change:
             changes.append(change)
     return updated, changes
 
 
-def _build_groups(visualizations: List[Dict[str, Any]], plan: Dict[str, Any]) -> List[Dict[str, Any]]:
+def _build_groups(
+    visualizations: List[Dict[str, Any]],
+    plan: Dict[str, Any],
+    notes: List[str] | None = None,
+) -> List[Dict[str, Any]]:
     buckets = {
         "kpi": [],
         "geo": [],
@@ -199,6 +245,7 @@ def _build_groups(visualizations: List[Dict[str, Any]], plan: Dict[str, Any]) ->
         "breakdown": [],
         "ranking": [],
         "detail": [],
+        "showcase": [],
     }
 
     for viz in visualizations:
@@ -210,7 +257,8 @@ def _build_groups(visualizations: List[Dict[str, Any]], plan: Dict[str, Any]) ->
 
     overflow = []
     assigned = set()
-    for role in ("kpi", "geo", "trend", "driver", "breakdown", "ranking", "detail"):
+    for role in ("kpi", "geo", "trend", "driver", "breakdown", "ranking",
+                 "detail", "showcase"):
         assigned.update(buckets[role])
     for viz in visualizations:
         viz_id = viz.get("id", "")
@@ -218,7 +266,12 @@ def _build_groups(visualizations: List[Dict[str, Any]], plan: Dict[str, Any]) ->
             overflow.append(viz_id)
     buckets["breakdown"].extend(overflow)
 
-    _promote_executive_analysis(buckets)
+    _fill_showcase(buckets, visualizations)
+    quarantine_note = _quarantine_problem_vizzes(buckets, visualizations)
+    if quarantine_note and notes is not None:
+        notes.append(quarantine_note)
+    _demote_heatmaps(buckets, visualizations)
+    _rebalance_tabs(buckets)
 
     groups = []
     for section in plan.get("sections", []):
@@ -241,33 +294,139 @@ def _build_groups(visualizations: List[Dict[str, Any]], plan: Dict[str, Any]) ->
     return groups
 
 
-EXEC_ANALYSIS_TARGET = 5   # trend + driver charts we want on the executive tab
-MANAGER_MIN = 1            # leave at least this many analysis charts for the manager tab
+MANAGER_TAB_TARGET = 6  # tab 2 is a themed drilldown
+SHOWCASE_TAB_TARGET = 4  # tab 3 is a short grab-bag
+EXEC_MIN_ANALYSIS = 4    # non-KPI charts the exec tab keeps before later tabs fill
+MANAGER_MIN_TILES = 3    # handoff 2026-09-02: a one-viz manager tab is a failure; floor is 3
+_MANAGER_ROLES = ("breakdown", "ranking", "detail")
+_EXEC_ANALYSIS_ROLES = ("trend", "driver")
 
+# Chart types that belong on the showcase tab: hand-authored/custom visuals and
+# the exotic types we would not put in front of an executive.
+SHOWCASE_CHART_TYPES = frozenset({
+    "MUZE_STUDIO", "CUSTOM_CHART", "SANKEY", "SPIDER_WEB", "PIE",
+})
 
-def _promote_executive_analysis(buckets: Dict[str, List[str]]) -> None:
-    """Keep the first tab substantial (~EXEC_ANALYSIS_TARGET analysis charts) by
-    promoting breakdown/ranking charts to drivers — without dropping the manager
-    tab below MANAGER_MIN charts."""
-    executive_analysis_count = len(buckets["trend"]) + len(buckets["driver"])
-    if executive_analysis_count >= EXEC_ANALYSIS_TARGET:
-        return
 
-    available = len(buckets["breakdown"]) + len(buckets["ranking"])
-    if available < 4:
-        # Small board: leave breakdowns/rankings on the manager tab as-is.
+def _is_table_viz(viz: Dict[str, Any]) -> bool:
+    return "TABLE" in _chart_type(viz)
+
+
+def _demote_heatmaps(
+    buckets: Dict[str, List[str]],
+    visualizations: List[Dict[str, Any]],
+) -> None:
+    """Heatmaps go to the END of the board — the showcase tab, appended last.
+
+    Run this AFTER _fill_showcase so heatmaps sit behind the custom charts and
+    the table rather than leading the tab.
+    """
+    heatmap_ids = [
+        viz.get("id", "")
+        for viz in visualizations
+        if _chart_type(viz) == "HEATMAP" and viz.get("id")
+    ]
+    if not heatmap_ids:
         return
-    promotable = max(0, available - MANAGER_MIN)
-    if promotable <= 0:
+    heatmap_set = set(heatmap_ids)
+    for role in ("geo", "trend", "driver", "breakdown", "ranking", "detail", "showcase"):
+        buckets[role] = [i for i in buckets[role] if i not in heatmap_set]
+    buckets["showcase"].extend(heatmap_ids)
+
+
+def _fill_showcase(
+    buckets: Dict[str, List[str]],
+    visualizations: List[Dict[str, Any]],
+) -> None:
+    """Populate the showcase tab: custom/exotic charts first, then exactly ONE
+    table, then whatever is left over from the manager tab.
+
+    Tables are banned everywhere else on the board; the showcase tab is the
+    single sanctioned exception, and it gets one.
+    """
+    by_id = {viz.get("id"): viz for viz in visualizations if viz.get("id")}
+    other_roles = ("breakdown", "ranking", "detail", "driver", "trend")
+
+    def take(predicate, limit) -> None:
+        for role in other_roles:
+            for viz_id in list(buckets[role]):
+                if len(buckets["showcase"]) >= limit:
+                    return
+                viz = by_id.get(viz_id)
+                if viz is not None and predicate(viz):
+                    buckets[role].remove(viz_id)
+                    buckets["showcase"].append(viz_id)
+
+    # 1. Custom and exotic visuals — the reason this tab exists.
+    take(lambda v: _chart_type(v) in SHOWCASE_CHART_TYPES, SHOWCASE_TAB_TARGET)
+
+    # 2. Exactly one table.
+    if not any(_is_table_viz(by_id[i]) for i in buckets["showcase"] if i in by_id):
+        take(_is_table_viz, len(buckets["showcase"]) + 1)
+
+    # 3. Any remaining tables go nowhere near the exec/manager tabs.
+    for role in other_roles:
+        keep = []
+        for viz_id in buckets[role]:
+            viz = by_id.get(viz_id)
+            if viz is not None and _is_table_viz(viz):
+                buckets["showcase"].append(viz_id)
+            else:
+                keep.append(viz_id)
+        buckets[role] = keep
+
+
+def _rebalance_tabs(buckets: Dict[str, List[str]]) -> None:
+    """Split analysis charts across the executive and manager tabs.
+
+    Retuned ONCE for three tabs (handoff 2026-09-02, item 3). The showcase tab
+    has already taken its charts by the time this runs, so `chart_total` is what
+    the first two tabs share.
+
+    Priorities, in order:
+      1. The manager tab gets at least MANAGER_MIN_TILES (3). A one-viz manager
+         tab is a failure the SEs called out; "keep that one and add two".
+      2. The executive tab keeps at least one analysis chart under its KPIs.
+      3. Beyond that, the exec tab is front-loaded: it takes everything past
+         MANAGER_TAB_TARGET, and the manager tab does not grow past it.
+
+    Worked examples (analysis charts only; KPIs and showcase excluded):
+      12 charts -> exec 6, manager 6
+      10 charts -> exec 4, manager 6
+       6 charts -> exec 3, manager 3
+       3 charts -> exec 1, manager 2      (floor yields to rule 2)
+       1 chart  -> exec 1, manager tab omitted
+    """
+    chart_total = sum(len(buckets[r]) for r in _EXEC_ANALYSIS_ROLES + _MANAGER_ROLES)
+    if chart_total == 0:
         return
 
-    needed = min(EXEC_ANALYSIS_TARGET - executive_analysis_count, promotable)
-    while needed > 0 and buckets["breakdown"]:
-        buckets["driver"].append(buckets["breakdown"].pop(0))
-        needed -= 1
-    while needed > 0 and buckets["ranking"]:
-        buckets["driver"].append(buckets["ranking"].pop(0))
-        needed -= 1
+    # Rule 3, then rule 1, then rule 2 caps it.
+    manager_target = max(MANAGER_MIN_TILES, chart_total - EXEC_MIN_ANALYSIS)
+    manager_target = min(manager_target, MANAGER_TAB_TARGET)
+    manager_target = min(manager_target, chart_total - 1)  # exec keeps >= 1
+    manager_target = max(manager_target, 0)
+
+    # Too many on the manager tab -> promote to the exec tab's driver section,
+    # taking from the END of each bucket so the tiles that most belong on a
+    # drilldown (multi-dimensional charts, detail tables) stay put — and never
+    # the strongest remaining chart, which sits at the front.
+    surplus = sum(len(buckets[r]) for r in _MANAGER_ROLES) - manager_target
+    for role in _MANAGER_ROLES:
+        while surplus > 0 and buckets[role]:
+            buckets["driver"].append(buckets[role].pop(0 if role == "detail" else -1))
+            surplus -= 1
+
+    # Too few -> borrow from the end of the exec analysis roles, but leave the
+    # exec tab at least one analysis chart in total.
+    deficit = manager_target - sum(len(buckets[r]) for r in _MANAGER_ROLES)
+    for role in ("driver", "trend"):
+        while deficit > 0 and buckets[role]:
+            exec_left = len(buckets["driver"]) + len(buckets["trend"])
+            if exec_left <= 1:
+                break
+            buckets["breakdown"].append(buckets[role].pop())
+            deficit -= 1
 
 
 def _build_tabs(
@@ -278,32 +437,36 @@ def _build_tabs(
     plan = plan or {}
     executive_tab = _executive_tab_name(plan)
     manager_tab = _manager_tab_name(plan)
-    executive_groups = [group for group in groups if group.get("tab") == executive_tab]
-    operations_groups = [group for group in groups if group.get("tab") == manager_tab]
+    showcase_tab = _showcase_tab_name(plan)
+
+    executive_groups = [g for g in groups if g.get("tab") == executive_tab]
+    operations_groups = [g for g in groups if g.get("tab") == manager_tab]
+    showcase_groups = [g for g in groups if g.get("tab") == showcase_tab]
 
     if not executive_groups:
         executive_groups = groups[:1]
-        operations_groups = groups[1:]
+        operations_groups = [g for g in groups[1:] if g.get("tab") != showcase_tab]
 
-    tabs = []
-    tabs.append(
+    tabs = [
         {
             "name": executive_tab,
             "description": "Executive-ready performance summary and key drivers",
             "tiles": _top_level_tiles(executive_groups, note_viz_id=note_viz_id),
             "group_layouts": [_group_layout(group) for group in executive_groups],
         }
-    )
-
-    if operations_groups:
-        tabs.append(
-            {
-                "name": manager_tab,
-                "description": "Manager-level breakdowns, rankings, and detail",
-                "tiles": _top_level_tiles(operations_groups),
-                "group_layouts": [_group_layout(group) for group in operations_groups],
-            }
-        )
+    ]
+    for name, description, tab_groups in (
+        (manager_tab, "Manager-level breakdowns, rankings, and detail", operations_groups),
+        (showcase_tab, "Custom visuals, exploratory charts, and row-level detail",
+         showcase_groups),
+    ):
+        if tab_groups:
+            tabs.append({
+                "name": name,
+                "description": description,
+                "tiles": _top_level_tiles(tab_groups),
+                "group_layouts": [_group_layout(group) for group in tab_groups],
+            })
     return tabs
 
 
@@ -364,8 +527,11 @@ def _group_layout(group: Dict[str, Any]) -> Dict[str, Any]:
     return {"id": group.get("id", ""), "tiles": tiles}
 
 
-KPI_ROW_H = 5
-CHART_ROW_H = 10
+# Golden reference heights: the IoT board's 4-KPI hero row is h=4 (w=3 each) and
+# its chart rows are h=6..8. We were shipping 5/10, which made boards feel loose.
+KPI_ROW_H = 4
+CHART_ROW_H = 8
+KPI_STANDARD_COUNT = 4
 
 
 def _chart_group_widths(n: int) -> List[int]:
@@ -401,11 +567,12 @@ def _apply_style(
         {
             "lb_border_type": style_config.get("border", "CURVED"),
             "lb_brand_color": style_config.get("brand_color", "LBC_A"),
+            # Verbatim from goldendemo/Vizio.com + Manufacturing_IoT (both agree).
             "hide_group_title": "false",
-            "hide_group_description": "true",
+            "hide_group_description": "false",
             "hide_group_tile_description": "false",
-            "hide_tile_description": "false",
-            "kpi_hero_font_size": "L",
+            "hide_tile_description": "true",
+            "kpi_hero_font_size": "M",
         },
     )
 
@@ -504,6 +671,9 @@ def _logo_block(company_name: str, logo_url: str | None) -> str:
     )
 
 
+HEATMAP_LIMIT = 1  # heatmaps read as a wall of green; keep at most this many
+
+
 def _reduce_heatmaps_and_apply_palettes(visualizations: List[Dict[str, Any]]) -> str:
     """Keep heatmaps rare and add varied chart colors to avoid all-blue boards."""
     heatmaps_seen = 0
@@ -517,7 +687,7 @@ def _reduce_heatmaps_and_apply_palettes(visualizations: List[Dict[str, Any]]) ->
 
         if chart_type == "HEATMAP":
             heatmaps_seen += 1
-            if heatmaps_seen > 2:
+            if heatmaps_seen > HEATMAP_LIMIT:
                 chart["type"] = _replacement_chart_type(answer)
                 chart_type = chart["type"]
                 heatmaps_converted += 1
@@ -534,13 +704,175 @@ def _reduce_heatmaps_and_apply_palettes(visualizations: List[Dict[str, Any]]) ->
     return "; ".join(parts)
 
 
+# Chart types we know render correctly from the specs we produce. Anything else
+# is a chart we cannot vouch for, so it goes to the showcase tab rather than in
+# front of an executive.
+KNOWN_GOOD_CHART_TYPES = frozenset({
+    "KPI", "LINE", "AREA", "COLUMN", "BAR", "PIE", "HEATMAP",
+    "STACKED_COLUMN", "STACKED_BAR", "STACKED_AREA", "LINE_COLUMN",
+    "ADVANCED_LINE", "ADVANCED_AREA", "ADVANCED_COLUMN", "ADVANCED_BAR",
+    "ADVANCED_STACKED_COLUMN", "ADVANCED_STACKED_BAR", "ADVANCED_STACKED_AREA",
+    "ADVANCED_LINE_COLUMN", "PIVOT_TABLE", "TABLE", "GRID_TABLE",
+    "GEO_AREA", "GEO_BUBBLE", "MUZE_STUDIO",
+})
+
+MUZE_SHOWCASE_MAX = 2  # how many surplus heatmaps to upgrade rather than downgrade
+
+
+def is_problem_viz(viz: Dict[str, Any]) -> str:
+    """Return a reason string when a viz is one we cannot vouch for, else "".
+
+    Rather than shipping a doubtful tile on the executive tab, move it to the
+    showcase tab where it is clearly experimental. boone's framing: "if we are
+    having trouble with a viz let's move it to this tab".
+    """
+    answer = viz.get("answer") or {}
+    chart = answer.get("chart") or {}
+    chart_type = str(chart.get("type") or "")
+
+    if not chart_type:
+        return "no chart type"
+    if chart_type not in KNOWN_GOOD_CHART_TYPES:
+        return f"unrecognised chart type {chart_type}"
+    # Still contradictory after the repair pass ran.
+    if chart_type in SINGLE_AXIS_CHART_TYPES and _has_secondary_axis(answer):
+        return "chart type cannot render its secondary axis"
+    # Deliberately NOT checking column counts: MCP legitimately leaves
+    # answer_columns/chart_columns empty on exports that render fine, so that
+    # signal quarantines whole boards. Only quarantine what we can prove is bad.
+    return ""
+
+
+def _muze_upgrade_surplus_heatmaps(visualizations: List[Dict[str, Any]]) -> str:
+    """Turn surplus heatmaps into designed Muze cell heatmaps instead of
+    downgrading them all to plain columns.
+
+    The dim x dim -> measure shape is exactly what the Muze cell-heatmap
+    archetype draws, so this upgrades the tile boone objected to rather than
+    flattening it. Capped at MUZE_SHOWCASE_MAX because these are experimental
+    and land on the showcase tab.
+    """
+    from demoprep_app.liveboards.muze_templates import (
+        CELL_HEATMAP, build_custom_visual_props,
+    )
+
+    upgraded = 0
+    seen = 0
+    for viz in visualizations:
+        answer = viz.get("answer") or {}
+        chart = answer.get("chart") or {}
+        if chart.get("type") != "HEATMAP":
+            continue
+        seen += 1
+        if seen <= HEATMAP_LIMIT or upgraded >= MUZE_SHOWCASE_MAX:
+            continue
+
+        columns = _answer_column_names(answer) or _chart_column_names(chart)
+        measures = [c for c in columns if _is_measure_name(c)]
+        dimensions = [c for c in columns
+                      if not _is_measure_name(c) and not _looks_like_date_column(c)]
+        if len(measures) != 1 or len(dimensions) < 2:
+            continue
+        bindings = {"x": dimensions[0], "y": dimensions[1], "measure": measures[0]}
+        try:
+            props = build_custom_visual_props(CELL_HEATMAP, bindings)
+        except ValueError:
+            continue  # e.g. a column name carrying a quote
+
+        chart["type"] = "MUZE_STUDIO"
+        chart["custom_visual_props"] = props
+        # The Muze program draws its own SVG; a stale axis config would only
+        # contradict it.
+        chart.pop("axis_configs", None)
+        upgraded += 1
+
+    if upgraded:
+        return f"Upgraded {upgraded} surplus heatmap(s) to Muze cell heatmaps"
+    return ""
+
+
+def _quarantine_problem_vizzes(
+    buckets: Dict[str, List[str]],
+    visualizations: List[Dict[str, Any]],
+) -> str:
+    """Move doubtful visualizations to the showcase tab, with the reason logged."""
+    by_id = {v.get("id"): v for v in visualizations if v.get("id")}
+    moved = []
+    for role in ("geo", "trend", "driver", "breakdown", "ranking", "detail"):
+        keep = []
+        for viz_id in buckets[role]:
+            reason = is_problem_viz(by_id.get(viz_id, {}))
+            if reason:
+                buckets["showcase"].append(viz_id)
+                moved.append(f"{viz_id} ({reason})")
+            else:
+                keep.append(viz_id)
+        buckets[role] = keep
+    if not moved:
+        return ""
+    return "Quarantined to the showcase tab: " + "; ".join(moved)
+
+
+def repair_secondary_axis_mismatch(visualizations: List[Dict[str, Any]]) -> str:
+    """Give every viz a chart type its OWN axis config allows — not just demoted heatmaps.
+
+    e93e261 fixed this only where WE overwrite chart["type"] (the heatmap cap).
+    But Spotter/MCP hands us two-measure "X vs Y" answers that ALREADY arrive as
+    ADVANCED_COLUMN / ADVANCED_BAR while their client_state_v2 declares a
+    secondary Y axis (isOpposite: true). Those are the majority, and they were
+    never touched — measured on boards built 2026-09-01:
+
+      198e8478 (secloud, mani.uppala)  Viz_7, Viz_9, Viz_10, Viz_11
+      25f8b739 (sebe, boone)           Viz_6, Viz_9, Viz_10
+      2c3fc690 (sebe)                  Viz_5, Viz_6, Viz_8
+
+    Every one is an "X vs Y" / "X and Y" two-measure comparison, and every one
+    rendered "Cannot display the chart".
+
+    Repair chosen: promote the TYPE to the combo chart, not strip the axis.
+    Stripping isOpposite also produces a valid spec, but it silently discards
+    the dual-scale intent and puts a $100M bar next to a 0.5 rate bar. The
+    promotion has direct rendering evidence behind it — on Larry's board
+    (eb6f9238) Viz_6 and Viz_12 are the SAME question with identical
+    axisProperties; the ADVANCED_COLUMN one blanked and the ADVANCED_LINE_COLUMN
+    one rendered.
+
+    Idempotent, and safe to run after every other chart-type mutation.
+
+    NOTE (golden-demo-v3): this is the MCP line's function (913c157), adopted
+    verbatim so the merge carries exactly ONE blank-tile repair. This branch's
+    own equivalent (2e3a32e) was deleted in its favour.
+    """
+    repaired: List[str] = []
+    for viz in visualizations:
+        if "note_tile" in viz:
+            continue
+        answer = viz.get("answer") or {}
+        chart = answer.get("chart")
+        if not isinstance(chart, dict):
+            continue
+        if str(chart.get("type", "")).upper() not in SINGLE_AXIS_CHART_TYPES:
+            continue
+        if not _has_secondary_axis(answer):
+            continue
+        chart["type"] = COMBO_CHART_TYPE
+        repaired.append(viz.get("id") or answer.get("name", "?"))
+
+    if not repaired:
+        return ""
+    return (
+        f"Repaired {len(repaired)} chart(s) whose type could not render their "
+        f"secondary axis ({', '.join(repaired)} -> {COMBO_CHART_TYPE})"
+    )
+
+
 def _force_chart_mode(visualizations: List[Dict[str, Any]]) -> str:
     """Charts rendering as data grids (display_mode=TABLE_MODE) but with a real
     chart type should display as charts. Genuine tables/KPIs are left alone."""
     fixed = 0
     for viz in visualizations:
         chart_type = _chart_type(viz)
-        if chart_type in {"TABLE", "PIVOT_TABLE", "KPI", "NOTE", "UNKNOWN"}:
+        if chart_type in {"TABLE", "GRID_TABLE", "PIVOT_TABLE", "KPI", "NOTE", "UNKNOWN"}:
             continue
         answer = viz.get("answer", {})
         if answer.get("display_mode") != "CHART_MODE":
@@ -600,6 +932,118 @@ _GEO_DIM_HINTS = (
 )
 
 
+_GEO_AREA_PREFERENCE = (
+    # Descending preference. STATE FIRST: demo data is overwhelmingly US and a
+    # state choropleth is the useful one. A country map of a single-country
+    # dataset is one solid blob.
+    (("state", "province"), ("estate", "statement", "status")),
+    (("county",), ()),
+    (("city", "metro"), ()),
+    (("country",), ()),
+)
+
+
+def detect_geo_area_column(model_columns: List[Dict[str, Any]]) -> Dict[str, str] | None:
+    """Find an area-mappable geo attribute (country preferred) plus a measure.
+
+    Unlike detect_geo_columns this needs NO lat/lon — GEO_AREA fills regions by
+    name. The model column must carry geo_config or the map renders blank
+    (see ThoughtSpotDeployer._geo_config_for_column).
+    """
+    names = [(c.get("name", ""), str(c.get("type", "")).upper())
+             for c in (model_columns or []) if c.get("name")]
+    # ThoughtSpot fills regions only for columns tagged with geo_config in the
+    # model. A GEO_AREA on an untagged column renders a BLANK map with no error,
+    # so an untagged candidate is worse than no map at all. When the caller does
+    # not report the flag at all (older callers, tests), fall back to name-based
+    # detection rather than silently disabling geo everywhere.
+    reports_geo_config = any("geo_config" in c for c in (model_columns or []))
+    geo_tagged = {
+        c.get("name", "") for c in (model_columns or []) if c.get("geo_config")
+    }
+    attrs = [
+        n for n, t in names
+        if t == "ATTRIBUTE" and (not reports_geo_config or n in geo_tagged)
+    ]
+    geodim = None
+    for hints, exclusions in _GEO_AREA_PREFERENCE:
+        geodim = next(
+            (n for n in attrs
+             if any(h in n.lower() for h in hints)
+             and not any(x in n.lower() for x in exclusions)),
+            None,
+        )
+        if geodim:
+            break
+    if not geodim:
+        return None
+    measure = _preferred_measure(names)
+    if not measure:
+        return None
+    return {"geodim": geodim, "measure": measure}
+
+
+def _preferred_measure(names: List[Tuple[str, str]]) -> str | None:
+    measures = [n for n, t in names if t == "MEASURE"
+                and not _LAT_RE.search(n) and not _LON_RE.search(n)]
+    return next(
+        (n for n in measures if any(k in n.lower() for k in
+         ("revenue", "sales", "amount", "value", "gmv", "spend"))),
+        measures[0] if measures else None,
+    )
+
+
+def _geo_area_client_state(geodim: str, measure_col: str) -> str:
+    """Golden-demo GEO_AREA state: yellow->red gradient fill with data labels."""
+    gradient = ["#ffffb2", "#fddd87", "#fba35d", "#f75534", "#f9140a", "#d70315", "#b10026"]
+    state = {
+        "version": "V4DOT2",
+        "chartProperties": {
+            "chartSpecific": {"customProps": "{\"tileType\":\"light-v9\"}"},
+        },
+        "columnProperties": [
+            {"columnId": geodim, "columnProperty": {}},
+            {"columnId": measure_col, "columnProperty": {"dataLabels": True}},
+        ],
+        "axisProperties": [
+            {"id": "geo-area-y", "properties": {
+                "axisType": "Y", "linkedColumns": [measure_col], "isOpposite": False}},
+            {"id": "geo-area-x", "properties": {"axisType": "X", "linkedColumns": [geodim]}},
+        ],
+        "systemMultiColorSeriesColors": [
+            {"serieName": measure_col,
+             "colorMap": [{"serieName": geodim, "color": gradient}]}
+        ],
+    }
+    return json_dumps_compact(state)
+
+
+def _build_geo_area_viz(
+    sibling_answer: Dict[str, Any],
+    geo: Dict[str, str],
+    viz_id: str = "Viz_Geo",
+) -> Dict[str, Any]:
+    geodim, measure = geo["geodim"], geo["measure"]
+    t_meas = f"Total {measure}"
+    return {
+        "id": viz_id,
+        "answer": {
+            "name": f"{measure} by {geodim}",
+            "tables": deepcopy(sibling_answer.get("tables", [])),
+            "search_query": f"sum [{measure}] [{geodim}]",
+            "answer_columns": [{"name": t_meas}, {"name": geodim}],
+            "chart": {
+                "type": "GEO_AREA",
+                "chart_columns": [{"column_id": t_meas}, {"column_id": geodim}],
+                "axis_configs": [{"x": [geodim], "y": [t_meas]}],
+                "client_state": "",
+                "client_state_v2": _geo_area_client_state(geodim, t_meas),
+            },
+            "display_mode": "CHART_MODE",
+        },
+    }
+
+
 def detect_geo_columns(model_columns: List[Dict[str, Any]]) -> Dict[str, str] | None:
     """From a model's columns [{name, type}], find lat/lon + a geo attribute + a
     measure. Returns {lat, lon, geodim, measure} or None if not geographic."""
@@ -670,6 +1114,148 @@ def _build_geo_viz(sibling_answer: Dict[str, Any], geo: Dict[str, str], viz_id:
     }
 
 
+def untagged_geo_candidates(model_columns: List[Dict[str, Any]] | None) -> List[str]:
+    """Attribute columns that LOOK geographic but carry no geo_config.
+
+    These are why a board has no map. Returned so the caller can say so out loud
+    rather than silently skipping — the model needs the tag, the liveboard can
+    do nothing about it.
+    """
+    if not model_columns or not any("geo_config" in c for c in model_columns):
+        return []
+    hints = tuple(h for hints, _ in _GEO_AREA_PREFERENCE for h in hints)
+    return [
+        c.get("name", "")
+        for c in model_columns
+        if str(c.get("type", "")).upper() == "ATTRIBUTE"
+        and not c.get("geo_config")
+        and any(h in c.get("name", "").lower() for h in hints)
+        and not any(x in c.get("name", "").lower()
+                    for x in ("estate", "statement", "status"))
+    ]
+
+
+# ThoughtSpot's allowed values for answer.chart.type, verbatim from a TML import
+# rejection (ZZ Lab build6, 2026-09-02). Note: "TABLE" is NOT in it — a table is
+# a valid chart type plus display_mode: TABLE_MODE. Synthesised vizzes must use
+# a type from this set or the WHOLE enhancement import is discarded.
+TS_ALLOWED_CHART_TYPES = frozenset({
+    "ADVANCED_AREA", "ADVANCED_BAR", "ADVANCED_BUBBLE", "ADVANCED_COLUMN",
+    "ADVANCED_DONUT", "ADVANCED_LINE", "ADVANCED_LINE_COLUMN",
+    "ADVANCED_LINE_STACKED_COLUMN", "ADVANCED_PIVOT_TABLE", "ADVANCED_SCATTER",
+    "ADVANCED_STACKED_AREA", "ADVANCED_STACKED_BAR", "ADVANCED_STACKED_COLUMN",
+    "ADVANCED_WATERFALL", "AREA", "BAR", "BUBBLE", "CANDLESTICK", "COLUMN",
+    "CUSTOM_CHART", "FUNNEL", "GEO_AREA", "GEO_BUBBLE", "GEO_EARTH_AREA",
+    "GEO_EARTH_BAR", "GEO_EARTH_BUBBLE", "GEO_EARTH_GRAPH", "GEO_EARTH_HEATMAP",
+    "GEO_HEATMAP", "GRID_TABLE", "HEATMAP", "KPI", "LINE", "LINE_COLUMN",
+    "LINE_STACKED_COLUMN", "MUZE_STUDIO", "PARETO", "PIE", "PIVOT_TABLE",
+    "SANKEY", "SCATTER", "SPIDER_WEB", "STACKED_AREA", "STACKED_BAR",
+    "STACKED_COLUMN", "TREEMAP", "WATERFALL", "WHISKER_SCATTER",
+})
+
+# The chart type we give a synthesised detail table. Golden-demo TABLE_MODE
+# vizzes keep a real chart type; GRID_TABLE is the server's table type and
+# contains "TABLE", which every table-routing check in this module keys on.
+DETAIL_TABLE_CHART_TYPE = "GRID_TABLE"
+
+
+_TABLE_ENTITY_HINTS = (
+    "name", "payer", "customer", "account", "provider", "practice", "store",
+    "product", "rep", "vendor", "supplier", "client", "patient", "employee",
+)
+_TABLE_MEASURE_HINTS = (
+    "revenue", "amount", "sales", "collected", "billed", "cost", "margin",
+    "rate", "count", "total", "spend", "value",
+)
+
+
+def detect_detail_table_columns(model_columns: List[Dict[str, Any]] | None) -> Dict[str, Any] | None:
+    """Pick one entity attribute and up to three measures for a detail table.
+
+    Prefers an entity-like attribute (a "Name" column, payer, provider...) over a
+    low-cardinality category, and business measures over ids. Returns
+    {"entity": str, "measures": [str, ...]} or None when the model cannot support
+    a useful table.
+    """
+    cols = [(c.get("name", ""), str(c.get("type", "")).upper())
+            for c in (model_columns or []) if c.get("name")]
+    attrs = [n for n, t in cols if t == "ATTRIBUTE" and not _looks_like_date_column(n)]
+    measures = [n for n, t in cols if t == "MEASURE"
+                and not _LAT_RE.search(n) and not _LON_RE.search(n)]
+    if not attrs or not measures:
+        return None
+
+    def attr_score(name: str) -> int:
+        low = name.lower()
+        return sum(10 for h in _TABLE_ENTITY_HINTS if h in low) + (5 if "name" in low else 0)
+
+    def measure_score(name: str) -> int:
+        low = name.lower()
+        return sum(10 for h in _TABLE_MEASURE_HINTS if h in low)
+
+    entity = max(attrs, key=attr_score)
+    if attr_score(entity) == 0:
+        return None  # nothing entity-like; a table of categories is not a drill-down
+    ranked = sorted(measures, key=measure_score, reverse=True)
+    return {"entity": entity, "measures": ranked[:3]}
+
+
+def _build_detail_table_viz(
+    sibling_answer: Dict[str, Any],
+    spec: Dict[str, Any],
+    viz_id: str = "Viz_Detail",
+) -> Dict[str, Any]:
+    entity, measures = spec["entity"], spec["measures"]
+    totals = [f"Total {m}" for m in measures]
+    query = f"[{entity}] " + " ".join(f"[{m}]" for m in measures)
+    return {
+        "id": viz_id,
+        "answer": {
+            "name": f"{entity} Detail",
+            "tables": deepcopy(sibling_answer.get("tables", [])),
+            "search_query": query,
+            "answer_columns": [{"name": entity}] + [{"name": t} for t in totals],
+            "table": {
+                "table_columns": [{"column_id": entity}] + [{"column_id": t} for t in totals],
+                "ordered_column_ids": [entity] + totals,
+                "client_state": "",
+                "client_state_v2": '{"tableVizPropVersion": "V1","wrapTableHeader": true}',
+            },
+            "chart": {
+                "type": DETAIL_TABLE_CHART_TYPE,
+                "chart_columns": [{"column_id": entity}] + [{"column_id": t} for t in totals],
+            },
+            "display_mode": "TABLE_MODE",
+        },
+    }
+
+
+def _maybe_build_detail_table(
+    visualizations: List[Dict[str, Any]],
+    model_columns: List[Dict[str, Any]] | None,
+) -> Dict[str, Any] | None:
+    """Synthesise the one sanctioned table when the board has none.
+
+    boone asked for exactly one table chart on the third page. Asking the
+    question generator for it did not work (build4/build5 produced none — a
+    prompt request is advisory), so the table is built deterministically here,
+    the same way the geo map is. _fill_showcase routes it to the showcase tab.
+    """
+    if not model_columns:
+        return None
+    if any("TABLE" in str((v.get("answer") or {}).get("chart", {}).get("type", ""))
+           for v in visualizations if "note_tile" not in v):
+        return None
+    spec = detect_detail_table_columns(model_columns)
+    if not spec:
+        return None
+    sibling = next((v.get("answer", {}) for v in visualizations
+                    if "note_tile" not in v and v.get("answer", {}).get("tables")), None)
+    if not sibling:
+        return None
+    return _build_detail_table_viz(sibling, spec)
+
+
 def _maybe_build_geo_viz(
     visualizations: List[Dict[str, Any]],
     model_columns: List[Dict[str, Any]] | None,
@@ -680,13 +1266,18 @@ def _maybe_build_geo_viz(
     if any(str(v.get("answer", {}).get("chart", {}).get("type", "")).startswith("GEO")
            for v in visualizations if "note_tile" not in v):
         return None
-    geo = detect_geo_columns(model_columns)
-    if not geo:
-        return None
     sibling = next((v.get("answer", {}) for v in visualizations
                     if "note_tile" not in v and v.get("answer", {}).get("tables")), None)
     if not sibling:
         return None
+    # Prefer a filled country/region map (golden-demo style); fall back to
+    # lat/lon bubbles only when no area-mappable attribute exists.
+    area_geo = detect_geo_area_column(model_columns)
+    if area_geo:
+        return _build_geo_area_viz(sibling, area_geo)
+    geo = detect_geo_columns(model_columns)
+    if not geo:
+        return None
     return _build_geo_viz(sibling, geo)
 
 
@@ -746,7 +1337,10 @@ def _convert_periodic_metric_charts(visualizations: List[Dict[str, Any]]) -> str
             {"column_id": date_column},
             {"column_id": measure_column},
         ]
-        chart.pop("client_state_v2", None)
+        # Previously this popped client_state_v2, which silently removed the only
+        # place sparkline/comparison settings can live (root cause of TASK 3:
+        # 0 of 18 graded boards had any sparkline state).
+        _apply_kpi_client_state(chart, date_column, measure_column)
         converted += 1
 
     if converted:
@@ -754,6 +1348,74 @@ def _convert_periodic_metric_charts(visualizations: List[Dict[str, Any]]) -> str
     return ""
 
 
+# Golden-demo KPI customProps. showComparison + showSparkline are what render the
+# trend line and the percent-change range chip; without them a KPI is a bare number.
+# PREV_AVAILABLE compares against the previous populated period, which is resilient
+# to demo data that ends mid-period.
+def _kpi_custom_props() -> Dict[str, Any]:
+    return {
+        "showLabel": True,
+        "showComparison": True,
+        "showSparkline": True,
+        "showAnomalies": True,
+        "showBounds": True,
+        "customCompare": "PREV_AVAILABLE",
+        "showOnlyLatestAnomaly": False,
+    }
+
+
+def _apply_kpi_client_state(
+    chart: Dict[str, Any],
+    date_column: str | None,
+    measure_column: str | None,
+) -> None:
+    """Write sparkline + comparison state onto a KPI chart, preserving whatever
+    else the export already carried."""
+    client_state = _load_chart_client_state(chart)
+    chart_properties = client_state.setdefault("chartProperties", {})
+    chart_properties["responsiveLayoutPreference"] = "AUTO_ON"
+    chart_specific = chart_properties.setdefault("chartSpecific", {})
+    chart_specific["customProps"] = json_dumps_compact(_kpi_custom_props())
+
+    if date_column and measure_column:
+        client_state["columnProperties"] = [
+            {"columnId": measure_column, "columnProperty": {"conditionalFormatting": {}}},
+            {"columnId": date_column, "columnProperty": {}},
+        ]
+        client_state["axisProperties"] = [
+            {"id": "kpi-y", "properties": {
+                "axisType": "Y", "linkedColumns": [measure_column], "isOpposite": False}},
+            {"id": "kpi-x", "properties": {"axisType": "X", "linkedColumns": [date_column]}},
+        ]
+    chart["client_state_v2"] = json_dumps_compact(client_state)
+
+
+def _fix_kpi_sparklines(visualizations: List[Dict[str, Any]]) -> str:
+    """Ensure EVERY KPI tile carries sparkline/comparison state.
+
+    _convert_periodic_metric_charts only touches answers it promotes to KPI;
+    KPIs that MCP already created never passed through it, which is why graded
+    boards showed bare numbers. A KPI with no date column gets the customProps
+    anyway — the comparison chip still renders; only the sparkline needs time.
+    """
+    fixed = 0
+    for viz in visualizations:
+        chart = viz.get("answer", {}).get("chart", {})
+        if chart.get("type") != "KPI":
+            continue
+        if _kpi_state_is_complete(chart):
+            continue
+        date_column, measure_column = _kpi_axis_columns(viz.get("answer", {}))
+        _apply_kpi_client_state(chart, date_column, measure_column)
+        fixed += 1
+    return f"Added sparkline/comparison state to {fixed} KPI(s)" if fixed else ""
+
+
+def _kpi_state_is_complete(chart: Dict[str, Any]) -> bool:
+    raw = chart.get("client_state_v2") or ""
+    return "showSparkline" in raw and "showComparison" in raw
+
+
 def _kpi_axis_columns(answer: Dict[str, Any]) -> Tuple[str | None, str | None]:
     columns = _answer_column_names(answer)
     if len(columns) < 2:
@@ -810,7 +1472,60 @@ def _looks_like_date_column(column: str) -> bool:
     )
 
 
+# Chart types that render a SINGLE Y axis. Assigning one of these to an answer
+# whose saved axis config declares a secondary axis (isOpposite: true) produces a
+# spec ThoughtSpot refuses to render — the tile shows "Cannot display the chart".
+SINGLE_AXIS_CHART_TYPES = frozenset({
+    "ADVANCED_COLUMN", "ADVANCED_BAR",
+    "ADVANCED_STACKED_COLUMN", "ADVANCED_STACKED_BAR",
+})
+
+# The combo type: bars for the money measure, line for the rate. Supports a
+# secondary axis, and is what ThoughtSpot's own Search picks for two measures on
+# different scales.
+COMBO_CHART_TYPE = "ADVANCED_LINE_COLUMN"
+
+
+def _has_secondary_axis(answer: Dict[str, Any]) -> bool:
+    """True when this answer's saved axis config declares a secondary Y axis.
+
+    Two measures on different scales (revenue vs. a rate) get an opposite axis.
+    The flag can sit directly on an axisProperties entry or nested under its
+    "properties" object depending on who wrote the client state, so this walks
+    the decoded structure rather than assuming one shape.
+    """
+    state = _load_chart_client_state(answer.get("chart", {}) or {})
+
+    def walk(node: Any) -> bool:
+        if isinstance(node, dict):
+            if node.get("isOpposite") is True:
+                return True
+            return any(walk(v) for v in node.values())
+        if isinstance(node, list):
+            return any(walk(v) for v in node)
+        return False
+
+    return walk(state)
+
+
 def _replacement_chart_type(answer: Dict[str, Any]) -> str:
+    """Pick a chart type for a demoted heatmap that is VALID for the axis config
+    already attached to the answer.
+
+    The heatmap cap overwrites chart["type"] but leaves client_state_v2 alone.
+    When that state declares a secondary axis, every single-axis type is an
+    invalid spec — this is what blanked ~9 tiles across 4 customer boards
+    (dev_notes/fix_handoff_2026_09_01.md TASK 1). Larry's board proves it:
+    Viz_6 and Viz_12 are the same question with identical axisProperties; Viz_6
+    was overwritten to ADVANCED_COLUMN and went blank, Viz_12 kept
+    ADVANCED_LINE_COLUMN and renders.
+
+    Returning the combo type keeps the dual-scale intent. Stripping the
+    secondary axis instead would "work" but puts a $100M bar next to a 0.5 bar.
+    """
+    if _has_secondary_axis(answer):
+        return COMBO_CHART_TYPE
+
     name = answer.get("name", "").lower()
     query = answer.get("search_query", "").lower()
     text = f"{name} {query}"
@@ -928,6 +1643,10 @@ def _executive_tab_name(plan: Dict[str, Any]) -> str:
     return plan.get("executive_tab") or EXECUTIVE_TAB
 
 
+def _showcase_tab_name(plan: Dict[str, Any]) -> str:
+    return plan.get("showcase_tab") or SHOWCASE_TAB
+
+
 def _manager_tab_name(plan: Dict[str, Any]) -> str:
     return plan.get("manager_tab") or OPERATIONS_TAB
 
@@ -1000,11 +1719,19 @@ def _role_for_viz(viz: Dict[str, Any]) -> str:
         return "geo"
     if "TABLE" in chart_type:
         return "detail"
-    # A single-metric periodic line is really a KPI (also handled upstream).
+    # A single-metric periodic question is a KPI regardless of the chart type MCP
+    # happened to draw it with. Keying off LINE/AREA alone lost the 4th KPI on the
+    # AdvancedMD board: "Net Collection Rate Quarterly" is one measure over one
+    # date column, but MCP rendered it as ADVANCED_COLUMN so it was filed as a
+    # trend and the board shipped 3 KPIs instead of the standard 4.
+    #
+    # The two guards that keep real charts out: no " by " (a breakdown), and at
+    # most one measure (two measures is a comparison chart, not a KPI).
     if (
-        chart_type in {"LINE", "AREA", "ADVANCED_LINE", "ADVANCED_AREA"}
+        chart_type not in {"TABLE", "PIVOT_TABLE", "NOTE"}
         and any(period in text for period in ("weekly", "monthly", "quarterly", "daily", "yearly"))
         and " by " not in text
+        and len([c for c in _answer_column_names(viz.get("answer", {})) if _is_measure_name(c)]) <= 1
     ):
         return "kpi"
     # Rankings before breakdowns: "Top 10 X by Y" is a ranking, not a plain breakdown.

demoprep_app/liveboards/muze_templates.py

@@ -0,0 +1,335 @@
+"""Templated ThoughtSpot custom charts (MUZE_STUDIO / BYOC).
+
+A MUZE_STUDIO visualization carries hand-authored JS + CSS + HTML, base64'd
+inside ``answer.chart.custom_visual_props``. The programs are parameterised by a
+handful of column-name constants, so they can be retargeted at a new model by
+substitution — that is what makes them templatable at all.
+
+**Why these are not copied verbatim from the reference board.** The four
+MUZE_STUDIO charts in ``goldendemo/Manufacturing_IoT`` each violate several
+documented BYOC hard rules, every one of which fails *silently* as a blank tile:
+
+* no ``try``/``catch`` — the BYOC sandbox replaces the real error with
+  "Something went wrong", so there is nothing to debug;
+* ``emitRenderCompletedEvent()`` on the success path only — a tile that throws
+  never reports in, which hangs Liveboard PDF export for the **whole board**;
+* an unbounded CDN ``<script>`` load with no timeout and (mostly) no fallback
+  host — a request that hangs fires neither ``onload`` nor ``onerror``, so the
+  awaited promise never settles and the tile renders blank;
+* three of the four omit ``html, body { height: 100% }``, so a percentage-height
+  stage can compute to 0px and the tile is blank while the console stays clean.
+
+The templates below take the *shape* of the reference charts and none of those
+bugs. The CDN dependency is removed outright — these draw plain SVG rather than
+pulling d3 — which deletes that entire failure class instead of bounding it.
+
+Reference: BibekTS/Thoughtspot-Amuzing-chart-skills (knowledge/hard-rules.md,
+knowledge/byoc-data-modes.md).
+"""
+
+from __future__ import annotations
+
+import base64
+import json
+from copy import deepcopy
+from typing import Any, Dict, List
+
+
+MUZE_CHART_TYPE = "MUZE_STUDIO"
+
+
+# --- Archetype A: cell heatmap (dim1 x dim2 -> measure) ----------------------
+# The shape behind "Gross Churn Rate by Region and Plan Family". Replaces the
+# default green HEATMAP, which is the tile boone objected to.
+
+_CELL_HEATMAP_JS = """\
+/**
+ * Cell heatmap: MEASURE aggregated over X_FIELD x Y_FIELD.
+ * Column constants below must match the ANSWER column display names (which
+ * carry the aggregation prefix, e.g. "Average Foo"), not the model column
+ * names. A name the search does not return is silently absent, not an error.
+ */
+const X_FIELD = '__X_FIELD__';
+const Y_FIELD = '__Y_FIELD__';
+const M_FIELD = '__M_FIELD__';
+
+// Resolve the host object. The documented entry point is the BARE identifier;
+// a host that passes viz as a wrapper argument leaves globalThis.viz undefined.
+// typeof alone is not enough: on a binding in its temporal dead zone it throws.
+function getViz() {
+  try { if (typeof viz !== 'undefined' && viz) return viz; } catch (e) {}
+  try { if (globalThis.viz) return globalThis.viz; } catch (e) {}
+  return null;
+}
+
+const host = getViz();
+
+function done() {
+  // Must fire on BOTH the success and failure paths, or Liveboard PDF export
+  // blocks on this tile forever.
+  try {
+    if (host && host.events && host.events.emitRenderCompletedEvent) {
+      host.events.emitRenderCompletedEvent();
+    }
+  } catch (e) {
+    console.warn('emitRenderCompletedEvent failed', e);
+  }
+}
+
+// Build our own mount point if the host has not laid one down yet - the tab
+// order is not contractual, and a null container throws before anything renders.
+function stage() {
+  let el = document.getElementById('chart');
+  if (!el) {
+    el = document.createElement('div');
+    el.id = 'chart';
+    document.body.appendChild(el);
+  }
+  return el;
+}
+
+function fail(err) {
+  // Never paint into the element that caused the throw.
+  const target = document.getElementById('chart') || document.body;
+  const pre = document.createElement('pre');
+  pre.className = 'chart-error';
+  pre.textContent = String((err && err.stack) || err);
+  target.appendChild(pre);
+  console.error(err);
+  done();
+}
+
+// Cells sometimes arrive wrapped as { value, formatted }; always unwrap.
+function cellVal(c) {
+  if (c && typeof c === 'object' && 'value' in c) return c.value;
+  return c;
+}
+
+function loadRows() {
+  const gd = host && host.getDataFromSearchQuery;
+  if (typeof gd !== 'function') return [];
+  const raw = gd().getData();
+  return raw.data.map(function (arr) {
+    const o = {};
+    raw.schema.forEach(function (c, i) { o[c.name] = cellVal(arr[i]); });
+    return o;
+  });
+}
+
+function lerp(a, b, t) { return a + (b - a) * t; }
+
+function rampColor(t, stops) {
+  const clamped = Math.max(0, Math.min(1, t));
+  const span = 1 / (stops.length - 1);
+  const i = Math.min(stops.length - 2, Math.floor(clamped / span));
+  const local = (clamped - i * span) / span;
+  const from = stops[i];
+  const to = stops[i + 1];
+  return 'rgb(' +
+    Math.round(lerp(from[0], to[0], local)) + ',' +
+    Math.round(lerp(from[1], to[1], local)) + ',' +
+    Math.round(lerp(from[2], to[2], local)) + ')';
+}
+
+const RAMP = [[224, 242, 254], [8, 145, 178], [12, 74, 110], [217, 119, 6], [220, 38, 38]];
+
+function svgEl(name, attrs) {
+  const el = document.createElementNS('http://www.w3.org/2000/svg', name);
+  Object.keys(attrs || {}).forEach(function (k) { el.setAttribute(k, attrs[k]); });
+  return el;
+}
+
+function render() {
+  const el = stage();
+  el.innerHTML = '';
+
+  const rows = loadRows();
+  if (!rows.length) {
+    el.appendChild(document.createTextNode('No data returned for this search.'));
+    return;
+  }
+  for (const field of [X_FIELD, Y_FIELD, M_FIELD]) {
+    if (!(field in rows[0])) {
+      throw new Error('Search did not return the column "' + field +
+        '". Available: ' + Object.keys(rows[0]).join(', '));
+    }
+  }
+
+  const xLabels = Array.from(new Set(rows.map(function (r) { return r[X_FIELD]; }))).sort();
+  const yLabels = Array.from(new Set(rows.map(function (r) { return r[Y_FIELD]; }))).sort();
+  const values = rows.map(function (r) { return Number(r[M_FIELD]); })
+                     .filter(function (v) { return isFinite(v); });
+  if (!values.length) throw new Error('Column "' + M_FIELD + '" has no numeric values.');
+  const min = Math.min.apply(null, values);
+  const max = Math.max.apply(null, values);
+  const span = max - min || 1;
+
+  const margin = { top: 16, right: 16, bottom: 52, left: 132 };
+  const totalW = 720;
+  const cellH = 56;
+  const gap = 6;
+  const innerW = totalW - margin.left - margin.right;
+  const totalH = yLabels.length * (cellH + gap) - gap;
+  const cellW = (innerW - (xLabels.length - 1) * gap) / xLabels.length;
+
+  const svg = svgEl('svg', {
+    viewBox: '0 0 ' + totalW + ' ' + (totalH + margin.top + margin.bottom),
+    preserveAspectRatio: 'xMidYMid meet'
+  });
+  const g = svgEl('g', { transform: 'translate(' + margin.left + ',' + margin.top + ')' });
+  svg.appendChild(g);
+
+  rows.forEach(function (r) {
+    const xi = xLabels.indexOf(r[X_FIELD]);
+    const yi = yLabels.indexOf(r[Y_FIELD]);
+    const v = Number(r[M_FIELD]);
+    if (xi < 0 || yi < 0 || !isFinite(v)) return;
+    const t = (v - min) / span;
+    const cx = xi * (cellW + gap);
+    const cy = yi * (cellH + gap);
+    g.appendChild(svgEl('rect', {
+      x: cx, y: cy, width: cellW, height: cellH, rx: 8, ry: 8, fill: rampColor(t, RAMP)
+    }));
+    const label = svgEl('text', {
+      x: cx + cellW / 2, y: cy + cellH / 2 + 5, 'text-anchor': 'middle',
+      'font-size': '15', 'font-weight': '600', fill: t < 0.15 ? '#0C4A6E' : '#ffffff'
+    });
+    label.textContent = Math.abs(v) >= 1000 ? Math.round(v).toLocaleString() : v.toFixed(2);
+    g.appendChild(label);
+  });
+
+  yLabels.forEach(function (name, i) {
+    const t = svgEl('text', {
+      x: -12, y: i * (cellH + gap) + cellH / 2 + 5, 'text-anchor': 'end',
+      'font-size': '13', fill: '#64748B'
+    });
+    t.textContent = name;
+    g.appendChild(t);
+  });
+  xLabels.forEach(function (name, i) {
+    const t = svgEl('text', {
+      x: i * (cellW + gap) + cellW / 2, y: totalH + 26, 'text-anchor': 'middle',
+      'font-size': '13', fill: '#64748B'
+    });
+    t.textContent = name;
+    g.appendChild(t);
+  });
+
+  el.appendChild(svg);
+}
+
+try {
+  render();
+  done();
+} catch (err) {
+  fail(err);
+}
+"""
+
+# html, body height is load-bearing: a percentage height resolves only against
+# an ancestor chain that is definite the whole way up, and the tile's <body> has
+# no explicit height. Without this the chart renders into a 0px-tall stage and
+# the tile is blank with a clean console.
+_CELL_HEATMAP_CSS = """\
+html, body { height: 100%; margin: 0; padding: 0; }
+body { font-family: 'DM Sans', system-ui, -apple-system, sans-serif; background: #fff; }
+#chart { height: 100%; width: 100%; }
+#chart svg { width: 100%; height: 100%; display: block; }
+.chart-error { color: #b91c1c; font-size: 12px; white-space: pre-wrap; padding: 8px; }
+"""
+
+_CELL_HEATMAP_HTML = """<div id="chart"></div>"""
+
+
+CELL_HEATMAP = "cell_heatmap"
+
+_ARCHETYPES: Dict[str, Dict[str, Any]] = {
+    CELL_HEATMAP: {
+        "js": _CELL_HEATMAP_JS,
+        "css": _CELL_HEATMAP_CSS,
+        "html": _CELL_HEATMAP_HTML,
+        "fields": ("x", "y", "measure"),
+    },
+}
+
+
+def available_archetypes() -> List[str]:
+    return sorted(_ARCHETYPES)
+
+
+def _b64(text: str) -> str:
+    return base64.b64encode(text.encode("utf-8")).decode("ascii")
+
+
+def build_custom_visual_props(archetype: str, bindings: Dict[str, str]) -> str:
+    """Return the ``custom_visual_props`` string for a MUZE_STUDIO chart.
+
+    ``bindings`` maps the archetype's field slots to ANSWER column display
+    names — the aggregated names ("Average Foo"), not model column names.
+    """
+    spec = _ARCHETYPES.get(archetype)
+    if spec is None:
+        raise ValueError(
+            f"unknown Muze archetype {archetype!r}; have {available_archetypes()}"
+        )
+    missing = [f for f in spec["fields"] if not bindings.get(f)]
+    if missing:
+        raise ValueError(f"{archetype} requires bindings for {missing}")
+
+    js = spec["js"]
+    for slot, placeholder in (
+        ("x", "__X_FIELD__"), ("y", "__Y_FIELD__"), ("measure", "__M_FIELD__"),
+    ):
+        if slot in spec["fields"]:
+            value = bindings[slot]
+            if "'" in value or "\\" in value:
+                raise ValueError(f"column name cannot contain a quote or backslash: {value!r}")
+            js = js.replace(placeholder, value)
+
+    if "</script>" in js or "</script>" in spec["html"]:
+        raise ValueError("a literal </script> breaks the host's HTML parser")
+
+    client_state = {
+        "version": 1,
+        "playground": {
+            "version": "v1",
+            "code": {
+                "jsCodeBase64": _b64(js),
+                "cssCodeBase64": _b64(spec["css"]),
+                "htmlCodeBase64": _b64(spec["html"]),
+            },
+        },
+    }
+    return json.dumps({"clientState": json.dumps(client_state)})
+
+
+def build_muze_viz(
+    viz_id: str,
+    name: str,
+    archetype: str,
+    bindings: Dict[str, str],
+    search_query: str,
+    sibling_answer: Dict[str, Any],
+) -> Dict[str, Any]:
+    """Build a MUZE_STUDIO visualization bound to the given columns.
+
+    The caller owns the search query, and it MUST select every bound column —
+    a column the chart names and the search omits is silently absent.
+    """
+    columns = [bindings[f] for f in _ARCHETYPES[archetype]["fields"]]
+    return {
+        "id": viz_id,
+        "answer": {
+            "name": name,
+            "tables": deepcopy(sibling_answer.get("tables", [])),
+            "search_query": search_query,
+            "answer_columns": [{"name": c} for c in columns],
+            "chart": {
+                "type": MUZE_CHART_TYPE,
+                "chart_columns": [{"column_id": c} for c in columns],
+                "client_state": "",
+                "custom_visual_props": build_custom_visual_props(archetype, bindings),
+            },
+            "display_mode": "CHART_MODE",
+        },
+    }

demoprep_app/liveboards/tml_yaml.py

@@ -0,0 +1,41 @@
+"""Safe YAML loading for ThoughtSpot TML.
+
+TML is YAML, but ThoughtSpot emits constructs that `yaml.safe_load` refuses:
+
+* Date filters contain a bare ``oper: =``. In YAML 1.1 a lone ``=`` is the
+  reserved *value* tag (``tag:yaml.org,2002:value``), and SafeLoader has no
+  constructor for it, so the whole document raises ConstructorError. Any
+  liveboard carrying a ``LAST_N_PERIOD`` filter — which both golden reference
+  boards do — is unparseable without this.
+
+Use :func:`load_tml` anywhere TML is read back from ThoughtSpot.
+"""
+
+from __future__ import annotations
+
+from typing import Any, Dict
+
+import yaml
+
+
+class TMLLoader(yaml.SafeLoader):
+    """SafeLoader that tolerates ThoughtSpot's TML quirks."""
+
+
+def _construct_value(loader: yaml.Loader, node: yaml.Node) -> str:
+    """Return the reserved YAML ``=`` tag as the plain string it is in TML."""
+    return loader.construct_scalar(node)
+
+
+TMLLoader.add_constructor("tag:yaml.org,2002:value", _construct_value)
+
+
+def load_tml(text: str | bytes) -> Dict[str, Any]:
+    """Parse a TML document. Returns {} for empty input.
+
+    Raises yaml.YAMLError for genuinely malformed TML — callers should not
+    swallow that, since it means ThoughtSpot returned something unexpected.
+    """
+    if not text:
+        return {}
+    return yaml.load(text, Loader=TMLLoader) or {}

demoprep_app/pipeline/build_demo.py

@@ -30,6 +30,7 @@ from dataclasses import dataclass, field
 
 from demoprep_app.dataset.contracts import DatasetBundle
 from demoprep_app.dataset.engine import BlueprintEngine
+from demoprep_app.dataset.formula import FormulaEvaluationError
 from demoprep_app.dataset.validator import ValidationReport, validate_bundle
 from demoprep_app.ddl import DatasetDdlCompiler
 from demoprep_app.scenario.blueprint import DemoBlueprint
@@ -41,6 +42,10 @@ from demoprep_app.scenario.contract import ScenarioContract
 MAX_TUNING_ROUNDS = 3
 # When an insight is too weak, multiply its distance-from-1.0 by this.
 BOOST_FACTOR = 1.6
+# How many times we repair a formula the engine refused to compute and try
+# again. Formula validation at blueprint time should make this dead code; it
+# exists so an unforeseen formula failure costs one measure, not the demo.
+MAX_FORMULA_REPAIR_ATTEMPTS = 3
 
 
 @dataclass(slots=True)
@@ -133,14 +138,35 @@ def build_demo(
     dataset: DatasetBundle | None = None
     for round_num in range(1, MAX_TUNING_ROUNDS + 1):
         progress(f"Generating dataset (round {round_num})...")
-        dataset = engine.generate(blueprint, row_count_guidance)
+        dataset = _generate_with_formula_repair(
+            engine, blueprint, row_count_guidance, warnings, progress
+        )
         report = validate_bundle(blueprint, dataset)
         progress(report.summary())
         if report.passed:
             break
-        if report.integrity_errors or report.formula_errors:
-            # Engine bugs, not tuning issues — do not loop on these.
+        if report.integrity_errors:
+            # Dangling keys / column mismatches are engine bugs and the
+            # Snowflake load would fail on them anyway — fail here, clearly.
             raise RuntimeError("Generated dataset failed integrity checks:\n" + report.summary())
+        if report.formula_errors:
+            # A stored derived value disagrees with its formula. Repair the
+            # formula and regenerate; if there is nothing left to repair, SHIP
+            # with a loud warning instead of aborting — the post-load quality
+            # gate recomputes or omits the column, and a board with one suspect
+            # tile beats no board at all (BLOCKING_ISSUES_2026_08_26, ISSUE C).
+            repaired = blueprint.repair_formulas()
+            if repaired:
+                for fix in repaired:
+                    progress(f"formula repair: {fix}")
+                warnings.extend(f"formula repair: {fix}" for fix in repaired)
+                continue
+            warnings.append(
+                "Stored derived values deviate from their formulas; shipping anyway so the "
+                "demo exists. The Snowflake quality gate will recompute or omit them:\n"
+                + report.summary()
+            )
+            break
         _boost_weak_insights(blueprint, report)
 
     if report is not None and not report.passed:
@@ -165,6 +191,43 @@ def build_demo(
     )
 
 
+def _generate_with_formula_repair(
+    engine: BlueprintEngine,
+    blueprint: DemoBlueprint,
+    row_count_guidance: int | None,
+    warnings: list[str],
+    progress,
+) -> DatasetBundle:
+    """Generate rows, repairing (never swallowing) a formula the engine rejects.
+
+    The engine now RAISES on an uncomputable formula instead of writing 0.0 to
+    every row. That correctness fix must not become a new way to produce
+    nothing, so each failure costs the offending measure and the run continues.
+    """
+    last_error: FormulaEvaluationError | None = None
+    for _ in range(MAX_FORMULA_REPAIR_ATTEMPTS):
+        try:
+            dataset = engine.generate(blueprint, row_count_guidance)
+        except FormulaEvaluationError as exc:
+            last_error = exc
+            progress(f"Formula rejected during generation: {exc}")
+            fixes = blueprint.repair_formulas()
+            if not fixes and exc.measure:
+                fixes = blueprint.drop_measure(exc.measure, reason=exc.reason)
+            if not fixes:
+                raise
+            for fix in fixes:
+                progress(f"formula repair: {fix}")
+            warnings.extend(f"formula repair: {fix}" for fix in fixes)
+            continue
+        warnings.extend(engine.formula_warnings())
+        return dataset
+    raise RuntimeError(
+        "Dataset generation kept failing on measure formulas after "
+        f"{MAX_FORMULA_REPAIR_ATTEMPTS} repair attempts. Last error: {last_error}"
+    )
+
+
 def _boost_weak_insights(blueprint: DemoBlueprint, report: ValidationReport) -> None:
     """Contract-level cleanup: strengthen only what failed, deterministically."""
     failed = {c.insight_id for c in report.failed_insights()}

demoprep_app/portal/README.md

@@ -0,0 +1,55 @@
+# Portal builder assets (`build_portal` groundwork)
+
+Everything needed to turn a finished DemoPrep run into a branded static portal
+site. The `build_portal(run_id)` MCP tool is **not implemented yet** — this
+directory holds the verified building blocks it will use.
+
+Design doc: `docs/agentspot_portal_handoff_2026_08_24.md`
+Verification results: `dev_notes/portal_template_spike.md`
+
+## Provenance — do not edit `template-tse/` by hand
+
+`template-tse/`, `scripts/apply-spec.mjs`, and `references/` are **vendored** from
+the `rebrand-thoughtspot-portal` skill in ThoughtSpot's public `tse_demos` repo:
+
+    repo:   https://github.com/thoughtspot/tse_demos
+    path:   .claude/skills/rebrand-thoughtspot-portal/
+    commit: 0c68520f51b15dcd12088e638273b914af2c7077   (2026-08-25)
+
+The template is Mani's team's and evolves independently. Upgrade deliberately by
+re-vendoring at a new pinned commit and re-testing — never by patching in place.
+
+`node_modules/` and `dist/` are intentionally excluded (121 MB of deps; the
+codemod runs `npm install` itself on first use).
+
+## How it works
+
+    interview -> spec.json -> deterministic codemod -> npm run build -> static host
+
+DemoPrep skips the interview: it already knows the host, liveboard GUID, model
+GUID, columns, company and palette because it created the run, so it authors
+`spec.json` directly. Spec contract: `references/spec-schema.md`.
+
+## Build a portal from a spec (verified working)
+
+    node demoprep_app/portal/scripts/apply-spec.mjs demoprep_app/portal/specs/triumph.json
+    (cd <slug>-tse && npm run build)     # output in dist/
+
+`specs/triumph.json` and `specs/vantage.json` are real worked examples (live
+GUIDs on sebe). Asset paths in a spec resolve relative to the current directory —
+see `assets/` for the placeholder logos those two specs reference.
+
+## Known gotchas (hit during verification — keep these fixed in any generator)
+
+1. `client` must be a single alphanumeric token (it becomes component/file names).
+   `"Vantage Capital"` produces `AskVantage Capital.tsx` and ~40 tsc errors. Pass
+   CamelCase (`VantageCapital`); the display name belongs in `content.company`.
+2. The codemod's tab pruning leaves dangling imports: with `inline:false` /
+   `action:false` it deletes `Carriers.tsx`/`Capacity.tsx` but not their `import`
+   lines in `src/App.tsx`, nor the now-unused `Truck`/`Activity` icons in
+   `src/components/TopBar.tsx`. `tsc` fails on both. Strip them post-codemod.
+3. `spec.action` and every `spec.columns.*` field are read unconditionally, even
+   when those features are off — always populate them.
+4. Hosting a portal needs the **exact** hostname in the cluster's CORS whitelist;
+   subdomain wildcards (`*.static.hf.space`) are rejected. `build_portal` must
+   return that instruction — the config API is 403 for non-admins.

demoprep_app/portal/builder.py

@@ -0,0 +1,565 @@
+"""Build a branded ThoughtSpot portal from a deployed liveboard.
+
+The importable core behind the `build_portal` MCP tool. No CLI, no argparse:
+callers pass values and get a dict back, or a PortalBuildError with a specific
+message. `scripts/build_portal.py` is a thin CLI over this for local testing.
+
+Chain: liveboard GUID -> model TML -> spec.json -> codemod -> npm run build
+       -> static HF Space -> URL (+ the CORS entry the cluster requires)
+
+Two Hugging Face Space constraints shape this module:
+
+1. **Everything is written to a writable temp dir, never the repo.** HF Spaces
+   run the container as a non-root user inside a root-owned WORKDIR, so writing
+   build output next to the code raises PermissionError at runtime — the exact
+   shape of the 2026-08-25 MCP outage.
+2. **The template's node_modules must already exist in the image.** The codemod
+   installs them itself when missing, which would be a write into that same
+   read-only tree. The Dockerfile pre-installs them as root at build time.
+"""
+from __future__ import annotations
+
+import json
+import os
+import re
+import shutil
+import subprocess
+import sys
+import tempfile
+from pathlib import Path
+
+REPO = Path(__file__).resolve().parent.parent.parent
+PORTAL = REPO / "demoprep_app" / "portal"
+CODEMOD = PORTAL / "scripts" / "apply-spec.mjs"
+
+
+class PortalBuildError(Exception):
+    """Any failure that should be reported to the caller verbatim."""
+
+
+# TS environment label -> (base url, .env key var). Mirrors the app's env table.
+TS_ENVS = {
+    "secloud": ("https://se-thoughtspot-cloud.thoughtspot.cloud", "TS_ENV_1_KEY_VAR"),
+    "sebe": ("https://sebe.thoughtspotstaging.cloud", "TS_ENV_2_KEY_VAR"),
+    "sebe-demo": ("https://sebe.thoughtspotstaging.cloud", "TS_ENV_3_KEY_VAR"),
+}
+
+# Measures whose name matches these read as money in the portal UI.
+CURRENCY_HINTS = ("usd", "revenue", "amount", "cost", "pay", "price", "margin usd",
+                  "balance", "savings", "spend")
+# Attributes matching these make good grouping/segment filters.
+SEGMENT_HINTS = ("tier", "segment", "category", "type", "grade", "status", "group", "class")
+
+
+# ---------------------------------------------------------------- ThoughtSpot
+
+def ts_client(env: str, as_user: str = ""):
+    """Authenticated deployer. Fails loudly if the env key is missing.
+
+    `as_user` selects the ThoughtSpot identity to act as via trusted auth. This
+    matters because TML export is permission-scoped: a model built by another SE
+    is FORBIDDEN to everyone else, so a portal over their liveboard has to be
+    built as them.
+    """
+    from thoughtspot_deployer import ThoughtSpotDeployer
+
+    if env not in TS_ENVS:
+        raise PortalBuildError(f"unknown ts_env '{env}'. choose from: {', '.join(TS_ENVS)}")
+    base, key_var = TS_ENVS[env]
+    secret = os.getenv(key_var)
+    if not secret:
+        raise PortalBuildError(f"{key_var} is not set in .env — cannot authenticate to {base}")
+    user = as_user or os.getenv("TEST_USER") or os.getenv("USER_EMAIL")
+    if not user:
+        raise PortalBuildError("no authenticated user available (set TEST_USER in .env)")
+    d = ThoughtSpotDeployer(base_url=base, username=user, secret_key=secret)
+    d.authenticate_trusted()
+    return d, base
+
+
+def export_tml(client, base: str, guid: str, associated: bool = False) -> list[dict]:
+    r = client.session.post(
+        f"{base}/api/rest/2.0/metadata/tml/export",
+        json={"metadata": [{"identifier": guid}], "export_associated": associated,
+              "export_fqn": True},
+    )
+    if not r.ok:
+        raise PortalBuildError(f"TML export failed for {guid}: HTTP {r.status} {r.text[:200]}"
+                 if hasattr(r, "status") else
+                 f"TML export failed for {guid}: HTTP {r.status_code} {r.text[:200]}")
+    return r.json()
+
+
+def resolve_model(client, base: str, liveboard_guid: str) -> str:
+    """Pull the model GUID out of a liveboard's first visualization."""
+    import yaml
+
+    docs = export_tml(client, base, liveboard_guid)
+    info = docs[0].get("info", {})
+    if info.get("type") != "liveboard":
+        raise PortalBuildError(f"{liveboard_guid} is a '{info.get('type')}', not a liveboard — "
+                 "a viz GUID in a liveboard slot is the #1 cause of a blank embed")
+    lb = yaml.safe_load(docs[0]["edoc"])["liveboard"]
+    for viz in lb.get("visualizations", []):
+        for tbl in (viz.get("answer", {}) or {}).get("tables", []) or []:
+            if tbl.get("fqn"):
+                return tbl["fqn"]
+    raise PortalBuildError("could not find a model GUID in that liveboard's visualizations")
+
+
+def model_columns(client, base: str, model_guid: str):
+    """Return (attributes, measures, date_columns, col_table).
+
+    `col_table` maps display name -> source table, which is what lets the date
+    column be matched to the fact its metrics actually live on.
+
+    Date columns need the associated TABLE TMLs: a model TML export strips the
+    `calendar` property, so dates are only identifiable via the underlying
+    column data types.
+    """
+    import yaml
+
+    docs = export_tml(client, base, model_guid, associated=True)
+    readable = [d for d in docs if d.get("edoc")]
+    if not readable:
+        detail = ""
+        for d in docs:
+            msg = ((d.get("info") or {}).get("status") or {}).get("error_message", "")
+            if msg:
+                detail = msg.splitlines()[0]
+                break
+        owner_hint = (
+            " The model is owned by another user and TML export is permission-scoped: "
+            "build as its owner (as_user=...) or have them share it."
+            if "FORBIDDEN" in detail.upper() else ""
+        )
+        raise PortalBuildError(
+            f"cannot read model {model_guid}: {detail or 'export returned no TML'}.{owner_hint}"
+        )
+    model_doc = next((d for d in readable if "model" in yaml.safe_load(d["edoc"])), None)
+    if not model_doc:
+        raise PortalBuildError(f"{model_guid} did not export as a model")
+    model = yaml.safe_load(model_doc["edoc"])["model"]
+
+    attributes, measures, by_id, col_table = [], [], {}, {}
+    for col in model.get("columns", []) or []:
+        props = col.get("properties", {}) or {}
+        if props.get("is_hidden"):
+            continue
+        name = col["name"]
+        cid = col.get("column_id", "")
+        by_id[cid] = name
+        col_table[name] = cid.split("::")[0] if "::" in cid else ""
+        (measures if props.get("column_type") == "MEASURE" else attributes).append(name)
+
+    # physical data types -> which model columns are dates
+    date_cols: list[str] = []
+    for d in readable:
+        doc = yaml.safe_load(d["edoc"])
+        tbl = doc.get("table")
+        if not tbl:
+            continue
+        tname = tbl.get("name", "")
+        for col in tbl.get("columns", []) or []:
+            dtype = str((col.get("db_column_properties") or {}).get("data_type", "")).upper()
+            if dtype in ("DATE", "DATETIME", "TIMESTAMP"):
+                display = by_id.get(f"{tname}::{col.get('db_column_name', col.get('name'))}")
+                if display and display not in date_cols:
+                    date_cols.append(display)
+    # fall back to name-shaped detection only if types gave us nothing
+    if not date_cols:
+        date_cols = [a for a in attributes
+                     if re.search(r"\b(date|month|day|week|quarter|year)\b", a, re.I)]
+    return attributes, measures, date_cols, col_table
+
+
+# ---------------------------------------------------------------- spec build
+
+def camel_token(name: str) -> str:
+    """`client` becomes component and FILE names, so it must be one bare token.
+
+    "Vantage Capital" -> "VantageCapital". A space here produces
+    `AskVantage Capital.tsx` and ~40 tsc errors.
+    """
+    tok = re.sub(r"[^A-Za-z0-9]+", " ", name).title().replace(" ", "")
+    if not tok:
+        raise PortalBuildError(f"cannot derive a component token from company name {name!r}")
+    return tok
+
+
+def pick_filters(attributes: list[str]) -> tuple[list[str], list[str]]:
+    """Two candidate lists for the portal's filter chips: first that returns data wins."""
+    names = [a for a in attributes if a.lower().endswith("name")]
+    segments = [a for a in attributes if any(h in a.lower() for h in SEGMENT_HINTS)]
+    primary = (names[:1] + segments[:1]) or attributes[:2]
+    remaining = [a for a in names[1:] + segments if a not in primary]
+    secondary = remaining[:2] or [a for a in attributes if a not in primary][:2]
+    if not primary or not secondary:
+        raise PortalBuildError("model has too few attribute columns to build filters from")
+    return primary, secondary
+
+
+def metric_spec(measures: list[str]) -> list[dict]:
+    out = []
+    for m in measures[:2]:
+        fmt = "currency" if any(h in m.lower() for h in CURRENCY_HINTS) else "number"
+        label = re.sub(r"\s+Usd$", "", m, flags=re.I)
+        label = re.sub(r"\s+Pct$", " %", label, flags=re.I)
+        out.append({"label": label, "format": fmt, "candidates": [m]})
+    return out
+
+
+def theme(primary: str, accent: str) -> dict:
+    """Full :root token dictionaries — the codemod replaces the whole block, so
+    every token the template reads must be present in both light and dark."""
+    def rgb(hex_):
+        h = hex_.lstrip("#")
+        return ", ".join(str(int(h[i:i + 2], 16)) for i in (0, 2, 4))
+
+    font = "'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif"
+    light = {
+        "--sl-cream": "#f7f8fa", "--sl-cream-2": "#eceef2", "--sl-white": "#ffffff",
+        "--sl-evergreen": primary, "--sl-evergreen-2": primary,
+        "--sl-green": primary, "--sl-green-dark": accent,
+        "--sl-primary-rgb": rgb(primary),
+        "--intime-blue": accent, "--sl-gold": "#e8b53c", "--sl-gold-ink": "#5f4708",
+        "--intime-grad": f"linear-gradient(90deg, {primary} 0%, {accent} 100%)",
+        "--sl-mint": "#e8f1fb", "--sl-mint-2": "#d5e6f7", "--sl-coral": "#a2191f",
+        "--sl-warn": "#fffbe6", "--sl-warn-ink": "#5f4708",
+        "--sl-ink": "#111827", "--sl-muted": "#5a6a7e",
+        "--sl-border": "#e2e6ec", "--sl-border-soft": "#eef1f5",
+        "--font-serif": font, "--font-sans": font, "--topbar-h": "66px",
+        "--shadow-card": "0 1px 2px rgba(15, 23, 42, 0.05), 0 8px 24px rgba(15, 23, 42, 0.07)",
+        "--shadow-lift": "0 6px 14px rgba(15, 23, 42, 0.08), 0 18px 40px rgba(15, 23, 42, 0.12)",
+        "--brand-navy": primary, "--brand-accent": accent,
+    }
+    dark = dict(light)
+    dark.update({
+        "--sl-cream": "#0b1622", "--sl-cream-2": "#101f2f", "--sl-white": "#14293d",
+        "--sl-evergreen": "#eef4fa", "--sl-evergreen-2": "#dbe6f0",
+        "--sl-green": accent, "--sl-green-dark": accent,
+        "--sl-primary-rgb": rgb(accent),
+        "--sl-mint": "#102a40", "--sl-mint-2": "#16354e", "--sl-coral": "#f0616a",
+        "--sl-warn": "#2a2413", "--sl-warn-ink": "#e8b53c",
+        "--sl-ink": "#eaf1f7", "--sl-muted": "#9fb2c6",
+        "--sl-border": "#23415a", "--sl-border-soft": "#172f44",
+        "--shadow-card": "0 1px 2px rgba(0, 0, 0, 0.5), 0 8px 24px rgba(0, 0, 0, 0.55)",
+        "--shadow-lift": "0 6px 14px rgba(0, 0, 0, 0.6), 0 18px 40px rgba(0, 0, 0, 0.65)",
+    })
+    return {"default": "light", "light": light, "dark": dark}
+
+
+def swaps(primary: str, accent: str) -> dict:
+    """Fixed left-hand template constants; only the right-hand side is brand."""
+    return {
+        "embedSwaps": [["#4F5BD5", accent], ["#3F49B8", accent], ["#333C99", primary],
+                       ["#22B8CF", accent], ["#0b0d10", "#0b1622"], ["#14181f", "#14293d"],
+                       ["#1a212b", "#16354e"], ["#232a34", "#23415a"], ["#0b1f19", "#101f2f"],
+                       ["rgba(45,164,191,0.14)", "rgba(122,180,232,0.16)"]],
+        "greenSwaps": [["#0f9a63", primary], ["#0c8554", primary], ["#d9ecea", "#d5e6f7"],
+                       ["#cfe3e2", "#c2ddf4"], ["#1bb978", accent], ["#15ae6e", primary]],
+        "cssSwaps": [["#0e1116", "#101f2f"], ["#10222a", "#101f2f"], ["#123047", "#14293d"]],
+    }
+
+
+def content(company: str, use_case: str, ai_name: str, website: str,
+            primary_label: str, secondary_label: str, samples: list[str]) -> dict:
+    return {
+        "company": company,
+        "aiName": ai_name,
+        "website": website,
+        "loginTitleHtml": f"{use_case},<br />in one place.",
+        "loginSubtitle": (f"Your live view of {use_case.lower()} — with AI-powered answers "
+                          f"on real data. Built on {company}, powered by ThoughtSpot."),
+        "loginStats": [{"value": "Live", "label": "Operational Data"},
+                       {"value": "AI", "label": "Natural-Language Answers"},
+                       {"value": "Real-time", "label": "Decision Support"}],
+        "loginCardSubtitle": "Continue with your ThoughtSpot session",
+        "loginDemoNoteHtml": ("<strong>Continue</strong> using your existing ThoughtSpot "
+                              "session on <code>{host}</code>. This origin must be "
+                              "CORS-allowlisted on that cluster for data to load."),
+        "platformNav": ["Overview", "Analytics", "Insights"],
+        "tabs": {"myReports": "My Reports", "analytics": "Analytics", "inline": "Detail",
+                 "action": "Actions", "ask": f"Ask {ai_name}", "spotter": "Spotter"},
+        "analyticsSubtitle": f"Your live {use_case.lower()} dashboard",
+        "spotterSubtitle": f"Ask questions of your {company} data in natural language.",
+        "inline": {"subtitleReady": "Select a row to view its insights",
+                   "subtitleIdle": "Key metrics across your portfolio",
+                   "loading": "Loading…", "errorTitle": "Couldn't load data",
+                   "errorMsg": "Unable to load.", "moreLabel": "View insights",
+                   "hideLabel": "Hide insights"},
+        "filters": {"primaryLabel": primary_label, "secondaryLabel": secondary_label},
+        "sampleQuestions": samples,
+        "askWelcome": (f"Ask me anything about {company} — or ask about your own data "
+                       "and I will chart it live."),
+        "askEmptySub": ("Ask any analytical question. Answers come straight from your live "
+                        "data with an interactive chart you can drill into."),
+        "action": {"label": "Take Action", "subtitlePrefix": "Right-click a row and choose",
+                   "subtitleSuffix": "to act on it.", "modalTitle": "New Request",
+                   "modalLead": "Pre-filled for the selected row.",
+                   "submitLabel": "Submit", "successTitle": "Request submitted"},
+        "chatbot": {"welcome": f"Hi! I'm {ai_name}. Ask me about your data.",
+                    "overview": f"{company}: {use_case}.",
+                    "greetingExample": samples[0] if samples else "show me the trend",
+                    "examples": samples[1:3]},
+        "trialQuestionTrigger": 3,
+        "upgradeReason": f"Upgrade to unlock drill-down, Ask AI and downloads.",
+        "analyticsKeywords": sorted({w.lower() for s in samples for w in re.findall(r"[a-z]{4,}", s.lower())}),
+    }
+
+
+def pick_date(dates: list[str], measures: list[str], col_table: dict[str, str]) -> str:
+    """Choose a date column on the SAME fact table as the metrics.
+
+    A date from a different fact is the cross-fact chasm bug: the filter appears
+    to work but slices a table the measures do not live on, so tiles come back
+    empty or silently unfiltered. Never fall back to "any date column".
+    """
+    if not dates:
+        raise PortalBuildError("no date column found on the model — the portal's date filter needs one")
+    fact = col_table.get(measures[0], "")
+    same_fact = [d for d in dates if col_table.get(d, "") == fact]
+    if not same_fact:
+        raise PortalBuildError(
+            f"no date column on '{fact}', the fact holding the chosen metrics "
+            f"({measures[0]}). Dates found: "
+            + ", ".join(f"{d} on {col_table.get(d, '?')}" for d in dates)
+            + ". Pick metrics from a dated fact, or add a date to that fact — "
+              "using another fact's date silently filters the wrong table."
+        )
+    return same_fact[0]
+
+
+def build_spec(*, company, use_case, host, liveboard, model, attributes, measures,
+               dates, primary, accent, logo_rel, website, col_table) -> dict:
+    fp, fs = pick_filters(attributes)
+    if len(measures) < 2:
+        raise PortalBuildError(f"model has only {len(measures)} measure(s); need at least 2 for the metrics")
+    date_col = pick_date(dates, measures, col_table)
+    segments = [a for a in attributes if any(h in a.lower() for h in SEGMENT_HINTS)]
+    ai_name = f"{company.split()[0]} AI"
+    samples = [f"{measures[0]} by {fp[0]}", f"{measures[1]} by {fs[0]}",
+               f"{measures[0]} over time", f"{measures[1]} by {segments[0] if segments else fp[0]}"]
+    return {
+        "client": camel_token(company),
+        "slug": re.sub(r"[^a-z0-9]+", "-", company.lower()).strip("-"),
+        "host": host,
+        "auth": "none",
+        "ids": {"analyticsLiveboard": liveboard, "model": model,
+                "inlineLiveboard": liveboard, "actionLiveboard": liveboard,
+                "actionViz": liveboard},
+        "action": {"id": "take-action", "name": "Take Action"},
+        "columns": {
+            "inlineName": fp,
+            "inlineMetrics": metric_spec(measures),
+            "filterPrimary": fp,
+            "filterSecondary": fs,
+            "date": date_col,
+            "hierarchy": {"segment": segments[0] if segments else fp[0],
+                          "rep": fp[0], "cadence": fs[0]},
+        },
+        "logo": {"svgPath": logo_rel},
+        "favicon": ('<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 32 32" width="32" '
+                    f'height="32"><rect width="32" height="32" rx="7" fill="{primary}"/>'
+                    f'<circle cx="16" cy="16" r="7" fill="{accent}"/></svg>'),
+        "theme": theme(primary, accent),
+        **swaps(primary, accent),
+        "features": {"askMode": "both", "inline": False, "action": False,
+                     "monetize": False, "tiers": False, "pinning": False},
+        "content": content(company, use_case, ai_name, website, fp[0], fs[0], samples),
+    }
+
+
+# ---------------------------------------------------------------- build steps
+
+def wordmark(company: str, primary: str, path: Path) -> None:
+    label = company.upper()[:26]
+    width = 40 + len(label) * 11
+    path.write_text(
+        f'<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 {width} 40" '
+        f'width="{width}" height="40"><text x="4" y="27" font-family="Inter, '
+        f'system-ui, sans-serif" font-size="18" font-weight="700" letter-spacing="0.5" '
+        f'style="fill:{primary}">{label}</text></svg>\n'
+    )
+
+
+def _tool(name: str) -> str:
+    """Absolute path to node/npm. Containers often lack them on the default PATH."""
+    found = shutil.which(name) or shutil.which(name, path="/usr/bin:/usr/local/bin:"
+                                                          f"{os.path.expanduser('~/.node/bin')}")
+    if not found:
+        raise PortalBuildError(
+            f"'{name}' is not available. The portal build needs Node to run the codemod "
+            "and vite. In the MCP Space this comes from the Dockerfile."
+        )
+    return found
+
+
+def run(cmd: list[str], cwd: Path) -> None:
+    p = subprocess.run(cmd, cwd=cwd, capture_output=True, text=True)
+    if p.returncode != 0:
+        raise PortalBuildError(
+            f"command failed: {' '.join(cmd)}\n{p.stdout[-1500:]}\n{p.stderr[-1500:]}")
+
+
+def patch_dangling_imports(app_dir: Path) -> None:
+    """The codemod's tab pruning deletes tab FILES but leaves their imports.
+
+    With inline/action off it removes Carriers.tsx / Capacity.tsx plus their JSX
+    and nav entries, but not the `import` lines in App.tsx nor the now-unused
+    lucide icons in TopBar.tsx — and `tsc --noUnusedLocals` fails on both.
+    """
+    app = app_dir / "src" / "App.tsx"
+    if app.exists():
+        kept = [ln for ln in app.read_text().splitlines(keepends=True)
+                if not re.match(r"^import (Carriers|Capacity) from", ln)]
+        app.write_text("".join(kept))
+    top = app_dir / "src" / "components" / "TopBar.tsx"
+    if top.exists():
+        kept = [ln for ln in top.read_text().splitlines(keepends=True)
+                if not re.match(r"^\s*(Truck|Activity),\s*$", ln)]
+        top.write_text("".join(kept))
+
+
+def deploy_space(dist: Path, owner: str, space: str) -> str:
+    from huggingface_hub import HfApi
+
+    token = os.getenv("HF_TOKEN")
+    if not token:
+        raise PortalBuildError("HF_TOKEN is not set in .env — cannot deploy the Space")
+    api = HfApi(token=token)
+    repo_id = f"{owner}/{space}"
+    api.create_repo(repo_id=repo_id, repo_type="space", space_sdk="static",
+                    private=False, exist_ok=True)
+    api.upload_folder(folder_path=str(dist), repo_id=repo_id, repo_type="space",
+                      commit_message="portal build")
+    return f"https://{owner.replace('.', '-')}-{space}.static.hf.space".replace("_", "-")
+
+
+
+# ---------------------------------------------------------------- public API
+
+def build_portal(
+    *,
+    company_name: str,
+    liveboard_guid: str,
+    model_guid: str = "",
+    use_case: str = "Analytics",
+    ts_env: str = "sebe",
+    primary: str = "#12283f",
+    accent: str = "#2f6fb0",
+    website: str = "",
+    logo_path: str = "",
+    space_name: str = "",
+    hf_owner: str = "thoughtspot-demoprep",
+    work_dir: str = "",
+    as_user: str = "",
+    deploy: bool = True,
+    on_progress=None,
+) -> dict:
+    """Build (and optionally deploy) a portal for one deployed liveboard.
+
+    Returns a dict with url/spec/paths. Raises PortalBuildError with a specific
+    message on any failure — no silent fallbacks.
+
+    work_dir defaults to a fresh temp directory, because the repo tree is not
+    writable in the MCP Space (see module docstring).
+    """
+    def say(msg: str) -> None:
+        if on_progress:
+            on_progress(msg)
+
+    if not (company_name or "").strip():
+        raise PortalBuildError("company_name is required")
+    if not (liveboard_guid or "").strip():
+        raise PortalBuildError("liveboard_guid is required")
+    for label, hexval in (("primary", primary), ("accent", accent)):
+        if not re.fullmatch(r"#[0-9A-Fa-f]{6}", hexval or ""):
+            raise PortalBuildError(f"{label} colour must be a 6-digit hex, got {hexval!r}")
+    if not CODEMOD.exists():
+        raise PortalBuildError(f"codemod missing at {CODEMOD}")
+
+    work = Path(work_dir) if work_dir else Path(tempfile.mkdtemp(prefix="portal-"))
+    work.mkdir(parents=True, exist_ok=True)
+
+    say(f"authenticating to {ts_env}" + (f" as {as_user}" if as_user else ""))
+    client, base = ts_client(ts_env, as_user)
+
+    guid = model_guid or resolve_model(client, base, liveboard_guid)
+    if not model_guid:
+        say(f"model derived from liveboard: {guid}")
+
+    say("reading model columns")
+    attributes, measures, dates, col_table = model_columns(client, base, guid)
+    say(f"{len(attributes)} attributes, {len(measures)} measures, {len(dates)} date column(s)")
+
+    logo_name = "brand-logo.svg"
+    if logo_path:
+        shutil.copy(logo_path, work / logo_name)
+    else:
+        wordmark(company_name, primary, work / logo_name)
+
+    host = base.replace("https://", "").replace("http://", "")
+    spec = build_spec(
+        company=company_name, use_case=use_case, host=host, liveboard=liveboard_guid,
+        model=guid, attributes=attributes, measures=measures, dates=dates,
+        primary=primary, accent=accent, logo_rel=logo_name, col_table=col_table,
+        website=website or f"https://{host}",
+    )
+    (work / "spec.json").write_text(json.dumps(spec, indent=2))
+    say(f"spec: filters {spec['columns']['filterPrimary'][0]} / "
+        f"{spec['columns']['filterSecondary'][0]}, date {spec['columns']['date']}")
+
+    say("running codemod")
+    run([_tool("node"), str(CODEMOD), "spec.json"], cwd=work)
+    app_dir = work / f"{spec['slug']}-tse"
+    if not app_dir.is_dir():
+        raise PortalBuildError(f"codemod did not produce {app_dir}")
+
+    patch_dangling_imports(app_dir)
+
+    say("building (vite)")
+    run([_tool("npm"), "run", "build"], cwd=app_dir)
+    dist = app_dir / "dist"
+    if not (dist / "index.html").exists():
+        raise PortalBuildError(f"build produced no dist/index.html in {app_dir}")
+
+    out = {
+        "status": "built",
+        "spec": spec,
+        "work_dir": str(work),
+        "app_dir": str(app_dir),
+        "dist": str(dist),
+        "liveboard_guid": liveboard_guid,
+        "model_guid": guid,
+        "resolved_columns": {
+            "filter_primary": spec["columns"]["filterPrimary"][0],
+            "filter_secondary": spec["columns"]["filterSecondary"][0],
+            "date": spec["columns"]["date"],
+            "metrics": [m["candidates"][0] for m in spec["columns"]["inlineMetrics"]],
+        },
+    }
+
+    if not deploy or not space_name:
+        out["message"] = "Built locally; not deployed (no space_name)."
+        return out
+
+    say(f"deploying to {hf_owner}/{space_name}")
+    url = deploy_space(dist, hf_owner, space_name)
+    hostname = url.replace("https://", "")
+    out.update({
+        "status": "success",
+        "url": url,
+        "space": f"{hf_owner}/{space_name}",
+        "cors_setup_required": {
+            "cluster": host,
+            "add_exact_host": hostname,
+            "where": "Develop -> Customizations -> Security Settings -> "
+                     "CORS whitelisted domains",
+            "note": "Wildcards such as *.static.hf.space are rejected by validation. "
+                    "Without this entry the portal renders but its filters and "
+                    "metrics cannot load data.",
+        },
+        "auth_note": "auth=none — viewers need an active ThoughtSpot session on that cluster.",
+    })
+    return out

demoprep_app/portal/references/SKILL.md

@@ -0,0 +1,186 @@
+---
+name: rebrand-thoughtspot-portal
+description: Guided builder that turns the reusable Vite + React + @thoughtspot/visual-embed-sdk portal template into a fully-skinned client app — brand, theme, data, tabs, custom actions, monetization and user tiers — from a plain-language interview. Use when someone wants to build / skin / white-label / clone a ThoughtSpot embedded analytics portal for a company (e.g. "build the portal for Acme", "skin the app for <client>", "make a ThoughtSpot demo app for X"). Walks a branching question tree (no coding required from the user), then clones the template and wires everything up, building and verifying at each step. Not for building a brand-new app architecture from scratch, or non-ThoughtSpot apps.
+---
+
+# Build a client ThoughtSpot portal (guided)
+
+You are running a **non-technical intake**: the user answers plain-language
+questions, you do all the code. Flow: **interview → build → verify → (ask before) deploy.**
+Never make the user touch code or GUIDs beyond pasting the values you ask for.
+
+## The base template
+
+The base app is **bundled inside this skill** at `template-tse/` (next to this
+`SKILL.md`) — a brand-neutral ("Northwind") build that already contains every
+feature toggleable: Analytics liveboard, inline-insights list tab, custom-action
+viz tab, standalone Spotter tab, "fancy" Ask-AI chat, monetization paywall,
+Premium/Basic tiers, host filters, light+dark theme. You keep what the answers ask
+for and strip the rest.
+
+**First-time setup: none needed.** The bundled template's deps are installed
+**automatically** by the codemod on the first build (it detects a missing
+`node_modules` and runs `npm install` once, looking in `~/.node/bin` if npm isn't on
+PATH). The codemod resolves the template relative to itself, so the skill works from
+any project directory. Each generated `<slug>-tse/` app symlinks its `node_modules` to
+the bundled template's, so nothing to install per app.
+
+The codemod copies the template to `<client>-tse/` (kebab-case) in the current project,
+excluding `node_modules dist .vercel .env.local`.
+
+## Step 0 — Show the "gather-first" card (before any question)
+
+People stall mid-interview when a question asks for a GUID they don't have open. So
+**before** asking anything, post this checklist in one plain-language message and let
+them collect it all at once. Keep it short — don't explain fields yet, just list what
+to have ready:
+
+> **Before we start, grab these — you'll paste them as we go (takes ~2 min):**
+> 1. **Your ThoughtSpot host URL** (e.g. `https://your-co.thoughtspot.cloud`)
+> 2. **The main dashboard's Liveboard ID** — open the liveboard; it's the long ID in
+>    the URL after `/pinboard/` or `/liveboard/`.
+> 3. **Your data model / worksheet ID** — the source the dashboard is built on.
+> 4. **A brand screenshot** — your website, product UI, or logo. I'll pull the colors
+>    from it. (A hex color works too if you have no image.)
+> 5. *(optional)* Any extra Liveboard/viz IDs if you want more than one tab.
+> 6. Your **logo** file (SVG or PNG), or say the word and I'll generate a wordmark.
+>
+> Building a demo with no real cluster? Just say so — I'll make up realistic values.
+
+Wait for a "ready" (or their first pastes), then start the interview. Don't block on
+it — if they'd rather answer as they go, proceed; the card just prevents the ambush.
+
+## Step 1 — Run the interview (the question tree)
+
+Ask with `AskUserQuestion`, **one section per round**, offering sensible defaults
+and only expanding a branch when its gate is "yes". Collect answers into the
+**spec** (schema at the bottom). Free-text values (GUIDs, column names, a pasted
+screenshot) you request in a normal message.
+
+**Sequencing rule (important):** never fire `AskUserQuestion` in the *same turn* as
+a free-text request — the popup blocks the text box, so the user can't paste. When a
+section has both, **ask the free-text first, wait for the reply, then** pop the
+choice(s). Batch multiple choices of the same section into one popup. If a section is
+purely choices, popup directly; if purely free-text, just ask in the message.
+
+Walk the tree in this order. **Section 0 comes first — it prunes the tree:**
+
+0. **Basic or Advanced demo?** — *Basic* = a quick demo (Analytics dashboards +
+   Spotter + Ask-AI + theme). *Advanced* = the full build. **If Basic, skip
+   sections 3, 4 and 6 entirely** (no inline tab, no custom-action tab, no
+   Add-Report). Record as `mode`.
+1. **Brand** — company name; AI assistant name (default "Ask <Company> AI");
+   one-line "what they do"; website URL; **ask them to attach the logo**
+   (SVG preferred, PNG fine) — or offer to generate a wordmark if they have none.
+2. **Data & filters** — ThoughtSpot host URL; main liveboard GUID; model/worksheet
+   GUID; **auth** (None = rides existing session · Basic = typed creds · Trusted =
+   advanced); and **which columns filter the main liveboard** (multiselect) + an
+   optional **date column**.
+3. **Inline-insights tab?** *(advanced only)* → if yes: inline **liveboard** GUID
+   (confirm it's a *liveboard*, not a viz/answer), tab name, the list's **name
+   column**, and up to 3 **metric columns**.
+4. **Custom-action workflow?** *(advanced only; optional — ask first, skip entirely
+   if no.)* Only if yes: **viz** GUID + its **liveboard** GUID, action label, tab name, and
+   **— required —** a description of the workflow: what screen/modal opens on click, which
+   row fields pre-fill which inputs, and what "Submit" does. **You build exactly that screen**
+   (rebuilding the scaffold modal); do NOT fall back to the generic "Request Bid" form.
+   A screenshot of the target screen is ideal — ask for one. If the user can't describe a
+   workflow, ask whether a simple labeled form is acceptable or the action should be dropped.
+5. **Ask-AI** — Normal Spotter (standalone), Fancy chat, or both.
+   - **Floating chatbot in the bottom-right corner?** (yes/no)
+   - **Monetization?** → if yes: which question number triggers the paywall.
+   - **Feature gating (tiers)?** → if yes: which actions Basic loses (drill-down,
+     Ask AI, downloads, …).
+6. **Pinning / "Add Report"?** *(advanced only)* — the Analytics "Add Report"
+   split-panel that lets a user build an answer (Report Builder / Spotter) and **pin
+   it to the dashboard**. If yes, **warn**: the Report Builder (`SearchEmbed`) renders
+   poorly in **dark** theme — suggest defaulting to light, or accepting the caveat. If
+   no, strip the Add-Report menu, the pin flow, and `PinModal`.
+7. **Theme** — **REQUIRED FIRST: explicitly ask the user to UPLOAD a screenshot for
+   color reference — and WAIT for it before anything else in this section.** Say it in
+   plain words, e.g. *"Please upload a screenshot for reference (your website, product UI,
+   a marketing shot, or your logo) and I'll derive the palette from it."* Do NOT jump to a
+   color-picker or "what's your primary color?" — the screenshot is the reference. A pasted
+   **URL** is an acceptable substitute; a **named color** is a last resort only when the
+   user genuinely has no image at all (and even then, ask). Derive the whole palette from
+   the uploaded image. Then: light / dark / both (+ default); font(s) for titles vs body;
+   confirm the primary/accent you read from the image; any specific asks (gradients, hues,
+   "make titles/KPIs pop").
+
+There is **no deploy question** — always **build + run locally** by default. Only
+deploy to Vercel if the user explicitly asks for it afterward.
+
+### Completeness gate — do NOT skip, do NOT fabricate
+
+Before you build, you MUST have an explicit answer for **every section 0–7**, including the
+free-text ones. Never invent or silently default a value the user could provide.
+
+- **Section 2 (Data) is REQUIRED free-text — actually ask for it and wait.** You must
+  collect the host URL, the Analytics liveboard GUID, the model/worksheet GUID, (if inline
+  enabled) the inline liveboard GUID, (if action enabled) the action viz + its liveboard
+  GUID, the filter columns, and the date column. **Do NOT invent GUIDs or a host.** If the
+  user is building a fictitious/demo app, ask explicitly "want me to make these up?" — get
+  a yes per-value; don't assume it.
+- **Section 7 (Theme) — ask for BOTH the screenshot/URL AND the font** (a Google font by
+  name, or an attached file). Never pick a font yourself without asking.
+- Batching choice-questions into one `AskUserQuestion` popup is fine, but it must **not
+  drop the free-text asks** that belong to the same section (Data GUIDs, font, columns).
+
+Then **echo the COMPLETE spec back in plain language** — every field, section by section,
+flagging anything the user let you make up — and get an explicit **"go"** before building.
+If you catch yourself about to fabricate a value or skip a section, stop and ask instead.
+
+## Step 2 — Build (spec-driven: interview → spec.json → codemod → hand-finish)
+
+The interview becomes a **`spec.json`**; a codemod does the mechanical ~90% in under a
+second; you hand-finish the few things a script can't.
+
+1. **Write `spec.json`.** Copy **`references/spec.example.json`** (a complete working
+   example) and edit every value per the interview. Field-by-field mapping + the swap-array
+   rules are in **`references/spec-schema.md`**. Derive the `theme.light`/`theme.dark` token
+   dicts from the brand screenshot/URL; in the `embedSwaps`/`greenSwaps`/`cssSwaps` arrays
+   change only the **right-hand** (brand) values — the left-hand hexes are fixed template
+   constants.
+2. **Run the codemod** from the project root (where you want the app written):
+   ```bash
+   node .claude/skills/rebrand-thoughtspot-portal/scripts/apply-spec.mjs spec.json
+   ```
+   It clones the bundled `template-tse` → `<slug>-tse/`, renames the brand, writes `content.ts`, patches
+   `config.ts` (host/GUIDs/columns/embed theme), switches auth, regenerates the `globals.css`
+   theme tokens + `@font-face`, drops in logo/favicon/font, and prunes the tabs `features`
+   turns off.
+3. **Hand-finish what the codemod can't** (it prints reminders):
+   - **Custom-action modal** — if a workflow was described, **MANDATORY**: rebuild the
+     scaffold modal to match it (build-playbook §5). Never leave the generic Request Bid form.
+   - Any bespoke UI the user flagged (custom tables, live lookups).
+   (Tabs and the monetize/tiers/pinning toggles are handled by the codemod — nothing to do.)
+4. **Build:** `export PATH="$HOME/.node/bin:$PATH" && (cd <slug>-tse && npm run build)`.
+   Fix any `tsc` errors and rebuild until clean.
+
+**`references/build-playbook.md`** is the reference for WHAT each field maps to and the
+non-obvious rules (host protection, solid-color embed buttons, masterpieces flag,
+liveboard-vs-viz, `Action` enum names, candidate-column fallback) — consult it when a
+codemod result looks off or when hand-editing. (The playbook's per-file manual steps still
+work as a fallback if the codemod can't run.)
+
+## Step 3 — Verify, run, (ask to) deploy
+
+- `npm run dev` and give the user the localhost URL.
+- Print a **spec recap** + a **"verify on your cluster"** list: exact column names
+  the candidate-lists resolved, the CSP/CORS **allowlist** step for the domain, and
+  the auth/session requirement.
+- Deploy to Vercel only if the user asked (see the playbook's deploy section).
+
+## Spec schema
+
+The interview's answers are written to a **`spec.json`** consumed by the codemod. The full
+field-by-field schema, the interview→field mapping, and the fixed-LHS swap rules live in
+**`references/spec-schema.md`**; **`references/spec.example.json`** is a complete worked
+example to copy and edit. (Deploy is always LOCAL by default — not interviewed; Vercel only
+on an explicit later ask.)
+
+## Rules of engagement
+- Confirm ambiguous ThoughtSpot IDs by **type** (liveboard / viz / saved answer) — the
+  #1 source of blank embeds.
+- Prefer **candidate-column lists** over a single guessed name; the app falls back gracefully.
+- Keep the user in plain language; you own the terminal, the build, and the fixes.

demoprep_app/portal/references/build-playbook.md

@@ -0,0 +1,154 @@
+# Build playbook — spec → edits
+
+Deterministic steps to turn the interview **spec** into a working `<client>-tse/`.
+Run `npm run build` (with `export PATH="$HOME/.node/bin:$PATH"`) after each stage.
+The template brand token is **`Northwind`** (casings: `Northwind` / `NORTHWIND` /
+`northwind`); the placeholder data tokens are `YOUR-CLUSTER.thoughtspot.cloud` and
+`REPLACE_*_GUID`.
+
+**Mode:** `basic` = strip the inline tab, the custom-action tab, and Add-Report
+(keep My Reports · Analytics · Ask-AI · Spotter · theme). `advanced` = per the
+answers. Deploy is always **local** (`npm run dev`) unless the user later asks for Vercel.
+
+## 0. Clone (manual fallback — the codemod does this for you)
+The template is bundled inside the skill. From the project root:
+```bash
+TPL=.claude/skills/rebrand-thoughtspot-portal/template-tse
+rsync -a --exclude node_modules --exclude dist --exclude .vercel --exclude .env.local \
+  "$TPL"/ <client>-tse/
+```
+Build locally by symlinking deps to the bundled template's:
+`ln -s "$(cd "$TPL" && pwd)/node_modules" <client>-tse/node_modules` (remove before any commit).
+
+## 1. Rebrand (Northwind → <Client>)
+Replace in all casings, then rename the two brand-named files. **zsh does not
+word-split an unquoted `$var`** — always use `find -exec … {} +`, never
+`perl … $(find …)`.
+```bash
+PERL='s/Northwind/<Client>/g; s/NORTHWIND/<CLIENT>/g; s/northwind/<client>/g;'
+find src -type f \( -name '*.ts' -o -name '*.tsx' -o -name '*.css' \) -exec perl -0pi -e "$PERL" {} +
+perl -0pi -e "$PERL" index.html
+perl -0pi -e 's/northwind-thoughtspot-portal/<client>-thoughtspot-portal/g' package.json
+mv src/components/NorthwindLogo.tsx src/components/<Client>Logo.tsx
+mv src/tabs/AskNorthwind.tsx src/tabs/Ask<Client>.tsx
+```
+Display casing: use title-case for the visible brand and "Ask <Client> AI". If the
+client brand is stylized ALL-CAPS, do a targeted follow-up replace on visible
+strings only (not const names). Re-grep for `northwind` — nothing should remain.
+
+**Logo:** if the user attached one, write it to `src/assets/<client>-logo.svg`
+(make the wordmark `fill="currentColor"` so it flips with the theme; keep one
+accent color fixed) and a 32×32 `public/<client>-icon.svg` favicon. Otherwise keep
+the generated wordmark. The import paths were already rewritten by the rename pass.
+
+## 2. Data (`src/config.ts`)
+Replace the placeholders with the spec:
+- `THOUGHTSPOT_HOST` ← host
+- `ANALYTICS_LIVEBOARD_ID` ← main liveboard GUID
+- `CADENCE_WORKSHEET_ID` **and** `WORKSHEET_ID` ← model/worksheet GUID
+- `INLINE_INSIGHTS_LIVEBOARD_ID` ← inline liveboard GUID (only if inline enabled)
+- `CAPACITY_LIVEBOARD_ID` / `CAPACITY_VIZ_ID` ← the action viz's liveboard + viz (only if action enabled)
+
+**Confirm each ID's type.** A whole-liveboard GUID passed as `vizId` renders the
+liveboard-empty "Nothing to see yet" state. Inline tab = `liveboardId` only (no
+vizId), filtered by the name column. Action tab = `liveboardId` + `vizId`.
+
+**Auth** (`src/lib/thoughtspot.ts` `init`): `none` → `AuthType.None` (rides the
+existing browser session; the login screen is a demo gate — keep the note telling
+users to be signed in to the cluster in another tab). `basic` → `AuthType.Basic`
+with the login form's username/password. `trusted` → wire a token endpoint (advanced).
+
+## 3. Columns (candidate lists in `config.ts`)
+Every column the app reads is a **candidate list** (first that returns data wins;
+graceful fallback). Set the user's real names as the first entry, keep sensible fallbacks:
+- `CADENCE_COLUMN_CANDIDATES` ← inline list's **name column** (e.g. carrier/process name)
+- `CADENCE_DETAIL_COLUMNS` ← the inline **metric columns** (label + format `number|text|currency|date` + candidates); keys are arbitrary, up to 3
+- `REGION_COLUMN_CANDIDATES` / `CORRIDOR_COLUMN_CANDIDATES` → rename to the spec's filter columns
+- `DATE_COLUMN` ← the date filter column
+- Org scoping is off by default (`ORG_COLUMN_CANDIDATES = []`); populate only if the client wants a tenant filter.
+
+## 4. Tabs / nav (`src/components/TopBar.tsx` + `src/App.tsx`)
+Template tabs: My Reports · Analytics · **Carriers** (inline) · **Capacity** (action) ·
+**Ask <Client> AI** · **Spotter** (standalone). Keep/rename/remove per spec:
+- Left platform nav (`PLATFORM_NAV`) → the client's context words.
+- Rename the inline tab (`carriers`→spec tabName) and action tab (`capacity`→spec tabName).
+- **Inline disabled** → remove the Carriers tab + route + delete `tabs/Carriers.tsx`.
+- **Action disabled** → remove the Capacity tab + route + delete `tabs/Capacity.tsx` + `components/BidModal.tsx`.
+- **Ask-AI mode**: `fancy` keeps `Ask<Client>` chat; `normal` keeps only the `Spotter` tab; `both` keeps both.
+Update `TabId`, the `TABS` array, and the `{tab === …}` routes together; `tsc` catches mismatches.
+
+## 5. Custom-action workflow (only if enabled) — `tabs/Capacity.tsx` + `components/BidModal.tsx`
+- `REQUEST_BID_ACTION_ID/NAME` ← the action label.
+- Custom action: `position: CustomActionsPosition.CONTEXTMENU`, `target: CustomActionTarget.VIZ`, scoped via `metadataIds`.
+- `onCustomAction` builds context from the clicked row (`buildSignalContext` in `lib/signalContext.ts` reads `contextMenuPoints.clickedPoint`).
+- **MANDATORY: rebuild the modal to match the workflow the user described — never ship the
+  generic "Request Bid" form for a client who described a workflow.** `BidModal.tsx` is a
+  *scaffold*, not the deliverable. Rename it to the workflow (e.g. `LeaveRequestModal.tsx`),
+  rebuild its fields to match the description (what opens on click, which row fields pre-fill
+  which inputs, what "Submit" confirms), and update the action label + tab name. The generic
+  bid form is only acceptable when the user explicitly said "any simple form is fine" or
+  declined to describe one. If a `spec.json`/codemod scaffolds the app, this modal is the
+  one step you STILL implement by hand afterward — the codemod prints a NOTE reminding you.
+- Verify at the end: right-click the viz → the action appears with the user's label → it opens
+  *their* described screen (not the bid form), pre-filled from the clicked row.
+
+## 6. Ask-AI, chatbot, monetization, tiers
+- **Floating chatbot** (`components/ChatBot.tsx`, mounted in `App.tsx`): keep if the user wants the bottom-right bot; otherwise remove the `<ChatBot …/>` mount.
+- **Fancy Ask tab** copy: welcome, subtitle, and `<Client>_SAMPLE_QUESTIONS` → the client's domain.
+- **Monetization**: paywall is `components/TrialModal.tsx`, shown once on the Nth question (trigger in `tabs/Ask<Client>.tsx`, `if (nextCount === N)`). Set N from spec. The template's modal is already the professional version.
+- **Add Report / pinning** (Analytics tab, `tabs/Analytics.tsx`): the "Add Report"
+  split-panel = a Report Builder (`SearchEmbed`) + Spotter answer that pins to the
+  dashboard (`HostEvent.Pin`), via `components/PinModal.tsx`. If **not** wanted,
+  remove the Add-Report menu button, the `panel` split, the Search/Spotter report
+  column, the pin flow, and the `PinModal` import/mount. **Caveat to surface if
+  wanted:** `SearchEmbed` (Report Builder) renders poorly in **dark** theme — advise
+  defaulting to light, or accept it.
+- **Tiers**: `context/TierContext.tsx` (Premium/Basic switch in TopBar) + `lib/tierActions.ts` `BASIC_DISABLED_ACTIONS`. Map requested features to `Action` enum members — **verify names in the installed SDK** (`grep "enum Action" node_modules/@thoughtspot/visual-embed-sdk/dist/*.d.ts`): it's `Action.AskAi` (not `AISearch`), `Action.DrillDown`, `Action.Download` / `DownloadAsCsv` / `DownloadAsPdf` / `DownloadAsXlsx`, `Action.SpotIQAnalyze`. Pass `disabledActions` + `disabledActionReason` on each liveboard embed. If tiers not wanted, remove the switcher + the disabledActions props.
+
+## 7. Theme (`src/styles/globals.css` tokens + `config.ts` embed vars)
+Derive the palette from the **screenshot/URL**. Set:
+- Host tokens on `:root` (light) and `:root[data-theme='dark']`: `--sl-cream`,
+  `--sl-white`, `--sl-evergreen` (logo/ink), `--sl-green` (**primary**),
+  `--sl-green-dark`, `--sl-mint`, `--sl-coral` (alert), `--sl-ink`, `--sl-muted`, borders.
+- `--<client>-grad` gradient for titles; `.page-title` already clips it.
+- Fonts: swap the Google-Fonts `<link>` in `index.html` and `--font-serif`/`--font-sans`
+  + `TS_FONT_URL`. Titles vs body weights per spec.
+- **Theme mode**: default in `context/ThemeContext.tsx` (`'dark'` or `'light'`); if
+  the client is single-theme, still keep both palettes defined.
+- **Dark base is ANY color** (black, navy, red, green…): it's fully token-driven — the
+  template has no hardcoded hues. Set the `:root[data-theme='dark']` block
+  (`--sl-cream` page bg, `--sl-white` cards, `--sl-ink`/`--sl-muted` text,
+  `--sl-green`/`--sl-green-dark` primary, borders) **and `--sl-primary-rgb`** (the
+  primary as `r,g,b`) — every glow/rim uses `rgba(var(--sl-primary-rgb), a)` and every
+  shadow is neutral slate, so changing those tokens re-themes the whole dark mode.
+- **Embed vars**: retint `TS_CSS_VARIABLES` (light) and `TS_VARS_DARK` (dark).
+
+**Embed gotchas (these bit us — do not repeat):**
+- **Buttons must be a SOLID color, never a gradient**, on `--ts-var-*-button-*-background`
+  — the SDK applies them as `background-color`, so a gradient renders as invisible/white.
+  Put gradients only on host CSS and on customCSS **rules** (`Triumph_EMBED_RULES` →
+  rename to `<Client>_EMBED_RULES`), e.g. KPI-tile background and viz-title text.
+- **Dark embed must blend**: set `--ts-var-liveboard-layout-background` / nav / group to
+  the app's page background, tiles one step lighter. Purge any leftover green
+  (`rgba(27,185,120,…)` ring in `--ts-var-viz-box-shadow`, `#143329` table header).
+- **`isLiveboardMasterpiecesEnabled: true`** must be on **every** LiveboardEmbed
+  (it's in `LIVEBOARD_EMBED_FLAGS` — spread it everywhere, incl. the inline tab) or
+  custom/BYOC vizzes throw "features disabled by your admin".
+- CustomCSS selectors for KPI tiles / viz titles / the AI-Highlights button are
+  version-dependent — they're best-effort; tell the user to Inspect and share a class
+  if a surface is missed.
+
+## 8. Copy
+Login hero tagline + subtitle (`pages/Login.tsx`), the demo-access note (host name),
+chatbot FAQ/system prompt (`lib/chatbot.ts`), and sample questions — all to the
+client's domain, derived from "what they do".
+
+## 9. Verify + deploy
+- `npm run build` clean, then `npm run dev` → give the localhost URL.
+- Recap the spec; list the resolved column names, the **CSP/CORS allowlist** step for
+  the deploy domain, and the auth/session requirement.
+- **Deploy only if asked.** Vercel CLI is cached (`npx vercel@latest`, scope
+  `koushik426s-projects` — personal can't be passed as `--scope <name>`):
+  `vercel project add <name>` → `vercel link --yes --project <name> --scope …` →
+  `vercel deploy --prod --yes --scope …`. A project named `<name>` serves
+  `<name>.vercel.app`. Remove the `node_modules` symlink before any git commit.

demoprep_app/portal/references/intake.md

@@ -0,0 +1,52 @@
+# Portal intake (plain English)
+
+Answer what you can — anything skipped uses a sensible default, and the assistant
+will ask follow-ups. No technical knowledge needed; just paste the values.
+
+## 0. How deep?
+- **Basic** (quick demo: dashboards + Spotter + Ask-AI + theme) or **Advanced**
+  (full build)? If Basic, sections 3, 4 and 6 are skipped.
+
+## 1. Brand
+- Company name:
+- Name for the AI assistant (default "Ask <Company> AI"):
+- One line on what the company does:
+- Website URL:
+- **Logo** — attach an SVG (preferred) or PNG. No logo? Say "generate one".
+
+## 2. Your ThoughtSpot data & filters
+- ThoughtSpot URL (e.g. `your-co.thoughtspot.cloud`):
+- Main dashboard (liveboard) — paste its GUID:
+- Data model / worksheet — paste its GUID:
+- Sign-in: (a) rely on being logged into ThoughtSpot in another tab · (b) type real credentials · (c) advanced/trusted
+- Which fields should filter the main dashboard (e.g. Region, Corridor)?
+- A date field to filter by?
+
+## 3. Inline-insights list tab?  (a table of items, each opening its own dashboard)
+- Want it? yes / no
+- If yes → dashboard (liveboard) GUID to open inline · tab name · the column of item
+  names (e.g. Carrier, Process) · up to 3 numbers to show per row.
+
+## 4. Custom-action workflow?  (right-click a chart row → run a Triumph-style action)  — OPTIONAL
+- Want it? yes / no  (if no, we skip this entirely)
+- If yes → visualization GUID + the dashboard it lives on · action label (e.g.
+  "Request Bid") · describe what should happen on click and which row details pre-fill it · tab name.
+
+## 5. Ask-AI
+- Which style: (a) standalone Spotter · (b) fancy chat experience · (c) both
+- Floating chatbot in the bottom-right corner? yes / no
+- Paywall/monetization popup? yes / no — if yes, after how many questions?
+- Premium vs Basic users? yes / no — if yes, what should Basic NOT be able to do
+  (drill-down, Ask AI, downloads, …)?
+
+## 6. Pinning / "Add Report"
+- Want the "Add Report" feature — build a chart in-app and pin it to the dashboard? yes / no
+  (Heads-up: the Report Builder doesn't render well in **dark** mode — light theme recommended if you enable it.)
+
+## 7. Look & feel
+- **Attach a screenshot of the brand's site** and/or its URL (colors are taken from it).
+- Light, dark, or both? Which is the default?
+- Fonts (title vs body), if specific.
+- Primary color, and any specific asks (gradients, hues, "make the titles/KPIs pop").
+
+_The app is built and run **locally** by default — ask afterward if you want it deployed to a public URL._

demoprep_app/portal/references/spec-schema.md

@@ -0,0 +1,79 @@
+# spec.json — schema + interview mapping
+
+The interview produces a **`spec.json`**; the codemod `scripts/apply-spec.mjs` turns it
+into a working `<slug>-tse/` app in well under a second. **Copy
+`references/spec.example.json`** (a complete, working InTime example) and edit every
+value per the interview — do not hand-write from scratch.
+
+Run from the project root (the codemod finds its bundled `template-tse/` relative to
+itself, and resolves your `spec.json` asset paths against the current directory):
+
+```bash
+node .claude/skills/rebrand-thoughtspot-portal/scripts/apply-spec.mjs spec.json
+```
+
+## Top-level fields
+
+| Field | From interview | Notes |
+|---|---|---|
+| `client` | Brand name, Title-case | e.g. `"InTime"`. Renames the `Northwind` token in all casings. |
+| `slug` | derived, kebab/lower | output dir `<slug>-tse/`; also namespaces storage keys. |
+| `host` | Data section | bare host, e.g. `team2.thoughtspot.cloud` (script adds `https://`). |
+| `auth` | Data section | `none` \| `basic` \| `trusted`. `basic` wires `AuthType.Basic` + login creds. |
+| `ids` | Data section | `analyticsLiveboard`, `model`, `inlineLiveboard`, `actionLiveboard`, `actionViz`. **Confirm each by TYPE** (liveboard vs viz) — a viz GUID in a liveboard slot = blank embed. |
+| `action` | Custom-action section | `{ id, name }` — the context-menu action label. |
+| `columns` | Data / inline / filters | `inlineName[]` (candidate list), `inlineMetrics[]` (`{label,format,candidates[]}`, ≤3), `filterPrimary[]`, `filterSecondary[]`, `date`, `hierarchy{segment,rep,cadence}`. Candidate lists = first that returns data wins. |
+| `appFont` | — | **Ignored.** The embed font is now derived automatically from `font.family` (see below). Safe to omit. |
+| `font` | Theme | the brand font. Two forms, both handled: **(1) Self-hosted** `{ family, file, srcPath }` — copied to `public/<file>` and `@font-face`'d in `globals.css` (host UI = real font); embeds get the nearest Google font (below). **(2) Google font by name** `{ family }` (optionally `{ googleUrl }`) — host loads it via `index.html`'s `<link>` (host UI = real font, no `@font-face`); embeds load it too (it's on a CDN → real font in embeds as well). Set `--font-serif`/`--font-sans` in `theme.light/dark` to `family`. **Embeds** always request `family` first, then the **nearest free Google font** as the reachable fallback (`FONT_FALLBACKS` in `apply-spec.mjs`: Larsseit→DM Sans, Gilroy→Poppins, Futura→Jost, Google faces→themselves, unknown→Inter); the real font renders in embeds only if it's a Google font or the cluster admin uploaded it. **Never self-host a font into the embeds** — it fails in the sandboxed iframe and blanks the theme. Omit `font` to keep Inter everywhere. |
+| `logo` | Brand | `{ svgPath }` (repo-relative). Script rewrites `fill:#hex`→`currentColor` so it flips with theme. |
+| `favicon` | Brand | inline SVG string → `public/<slug>-icon.svg`. |
+| `theme` | Theme | see below. |
+| `features` | multiple | see below — controls **tab pruning**. |
+| `content` | all copy | the entire `content.ts` object (every human string). |
+| `embedSwaps` / `greenSwaps` / `cssSwaps` | Theme | hex remaps — **fixed left side, brand right side** (see below). |
+
+## `theme`
+
+- `default`: `"light"` | `"dark"`.
+- `light` / `dark`: the **full** `:root` token dictionaries (the codemod replaces the whole
+  block, so include every token — colors, fonts, `--topbar-h`, shadows). Derive the palette
+  from the brand screenshot/URL. Must include `--brand-navy` and `--brand-accent` (fixed
+  brand colors, **same value in light and dark** — they drive the chatbot, which is a light
+  card with a dark header in both themes).
+
+## `features` (tab pruning the codemod performs)
+
+- `askMode`: `"both"` (Ask + Spotter) · `"fancy"` (removes Spotter tab) · `"normal"` (removes Ask tab).
+- `inline`: `false` removes the inline (Carriers/Shifts) tab + file.
+- `action`: `false` removes the action (Capacity/Leaves) tab + BidModal.
+- `monetize` / `tiers` / `pinning`: **automated** — the codemod writes `src/flags.ts` and the
+  template gates these in place (`{FLAGS.tiers && …}` etc.). `false` hides the paywall, the
+  tier switcher (tier stays Premium → everything unlocked), and the Add-Report panel. No hand-step.
+
+## The swap arrays — fixed LHS, brand RHS
+
+Each is `[from, to]`. The **left-hand values are constants** (the template's neutral
+palette); change only the **right-hand** values to the brand equivalents. This is how the
+embed CSS variables (indigo `config.ts`), the liveboard tab/button accents (green
+`thoughtspot.ts`), and a few leftover dark-override rules (`globals.css`) get rebranded.
+
+- **`embedSwaps`** (config.ts) LHS: `#4F5BD5` (primary), `#3F49B8` (hover), `#333C99` (active),
+  `#22B8CF` (gradient end), `#0b0d10` `#14181f` `#1a212b` `#232a34` (dark embed bg family),
+  `#0b1f19` (dark input), `rgba(45,164,191,0.14)` (ring).
+- **`greenSwaps`** (thoughtspot.ts liveboard accents) LHS: `#0f9a63` `#0c8554` `#d9ecea`
+  `#cfe3e2` `#1bb978` `#15ae6e`.
+- **`cssSwaps`** (globals.css dark-override rules) LHS: `#0e1116` `#10222a` `#123047`.
+
+> The chatbot palette is **token-driven** (`--brand-navy`/`--brand-accent`), so it needs
+> **no** swaps — just set those two theme tokens.
+
+## After the codemod — hand-steps (what a script can't do)
+
+1. **Custom-action modal** (if a workflow was described): **MANDATORY** — rebuild
+   `BidModal.tsx` to match the described screen (rename it, e.g. `LeaveRequestModal.tsx`),
+   per `build-playbook.md` §5. **Never ship the generic Request Bid form** for a described
+   workflow.
+2. Any bespoke UI the user flagged (custom tables, live lookups).
+
+Then: `npm run build` (fix TS errors), `npm run dev`, and print the spec recap + the
+"verify on your cluster" list (resolved column names, CORS/CSP allowlist, auth requirement).

demoprep_app/portal/references/spec.example.json

@@ -0,0 +1,312 @@
+{
+  "client": "InTime",
+  "slug": "yourco",
+  "host": "YOUR-CLUSTER.thoughtspot.cloud",
+  "auth": "basic",
+  "ids": {
+    "analyticsLiveboard": "REPLACE_MAIN_LIVEBOARD_GUID",
+    "model": "REPLACE_MODEL_GUID",
+    "inlineLiveboard": "REPLACE_INLINE_LIVEBOARD_GUID",
+    "actionLiveboard": "REPLACE_MAIN_LIVEBOARD_GUID",
+    "actionViz": "REPLACE_ACTION_VIZ_GUID"
+  },
+  "action": {
+    "id": "request-leave",
+    "name": "Request Leave"
+  },
+  "columns": {
+    "inlineName": [
+      "Rank Name",
+      "Rank",
+      "Rank Type",
+      "Employee"
+    ],
+    "inlineMetrics": [
+      {
+        "label": "Manual Overtime Hours",
+        "format": "number",
+        "candidates": [
+          "Total Manual Overtime Hours",
+          "Manual Overtime Hours",
+          "Overtime Hours"
+        ]
+      },
+      {
+        "label": "Overtime Cost",
+        "format": "currency",
+        "candidates": [
+          "Overtime Cost",
+          "Total Overtime Cost",
+          "OT Cost"
+        ]
+      }
+    ],
+    "filterPrimary": [
+      "Branch Name",
+      "Company Name",
+      "Branch",
+      "Dim Branch"
+    ],
+    "filterSecondary": [
+      "Rank Name",
+      "Rank",
+      "Rank Type"
+    ],
+    "date": "Sched Start Time",
+    "hierarchy": {
+      "segment": "Branch Name",
+      "rep": "Rank Name",
+      "cadence": "Employee"
+    }
+  },
+  "appFont": "'Larsseit', 'Inter', 'Segoe UI', sans-serif",
+  "font": {
+    "family": "Manrope",
+    "googleUrl": "https://fonts.googleapis.com/css2?family=Manrope:wght@400;500;600;700&display=swap"
+  },
+  "logo": {
+    "svgPath": "your-logo.svg"
+  },
+  "favicon": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 32 32\" width=\"32\" height=\"32\"><rect width=\"32\" height=\"32\" rx=\"7\" fill=\"#0E3147\"/><circle cx=\"16\" cy=\"9\" r=\"2.4\" fill=\"#F4D14E\"/><rect x=\"13.7\" y=\"13\" width=\"4.6\" height=\"12.5\" rx=\"2.3\" fill=\"#ffffff\"/></svg>",
+  "theme": {
+    "default": "light",
+    "light": {
+      "--sl-cream": "#f4f7fb",
+      "--sl-cream-2": "#e9f0f7",
+      "--sl-white": "#ffffff",
+      "--sl-evergreen": "#0e3147",
+      "--sl-evergreen-2": "#0a2436",
+      "--sl-green": "#0e3147",
+      "--sl-green-dark": "#16466a",
+      "--sl-primary-rgb": "14, 49, 71",
+      "--intime-blue": "#2f6fb0",
+      "--sl-gold": "#f4d14e",
+      "--sl-gold-ink": "#6b5410",
+      "--intime-grad": "linear-gradient(90deg, #0e3147 0%, #2f6fb0 100%)",
+      "--sl-mint": "#e6f1fb",
+      "--sl-mint-2": "#d5e8f8",
+      "--sl-coral": "#a2191f",
+      "--sl-warn": "#fffbe6",
+      "--sl-warn-ink": "#6b5410",
+      "--sl-ink": "#12283f",
+      "--sl-muted": "#5a6a7e",
+      "--sl-border": "#e1e8f0",
+      "--sl-border-soft": "#eef2f7",
+      "--font-serif": "'Larsseit', 'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--font-sans": "'Larsseit', 'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--topbar-h": "66px",
+      "--shadow-card": "0 1px 2px rgba(15, 23, 42, 0.05), 0 8px 24px rgba(15, 23, 42, 0.07)",
+      "--shadow-lift": "0 6px 14px rgba(15, 23, 42, 0.08), 0 18px 40px rgba(15, 23, 42, 0.12)",
+      "--brand-navy": "#0e3147",
+      "--brand-accent": "#2f6fb0"
+    },
+    "dark": {
+      "--sl-cream": "#0a2233",
+      "--sl-cream-2": "#0e293c",
+      "--sl-white": "#12314a",
+      "--sl-evergreen": "#eef4fa",
+      "--sl-evergreen-2": "#dbe6f0",
+      "--sl-green": "#7fb0e0",
+      "--sl-green-dark": "#5e97cc",
+      "--sl-primary-rgb": "127, 176, 224",
+      "--intime-blue": "#7fb0e0",
+      "--sl-gold": "#f4d14e",
+      "--sl-gold-ink": "#0a2233",
+      "--intime-grad": "linear-gradient(90deg, #7fb0e0 0%, #bfe0f5 100%)",
+      "--sl-mint": "#10293d",
+      "--sl-mint-2": "#16344c",
+      "--sl-coral": "#f0616a",
+      "--sl-ink": "#eaf1f7",
+      "--sl-muted": "#9fb2c6",
+      "--sl-border": "#23415a",
+      "--sl-border-soft": "#172f44",
+      "--shadow-card": "0 1px 2px rgba(0, 0, 0, 0.5), 0 8px 24px rgba(0, 0, 0, 0.55)",
+      "--shadow-lift": "0 6px 14px rgba(0, 0, 0, 0.6), 0 18px 40px rgba(0, 0, 0, 0.65)",
+      "--brand-navy": "#0e3147",
+      "--brand-accent": "#2f6fb0"
+    }
+  },
+  "embedSwaps": [
+    [
+      "#4F5BD5",
+      "#2f6fb0"
+    ],
+    [
+      "#3F49B8",
+      "#285f97"
+    ],
+    [
+      "#333C99",
+      "#22527f"
+    ],
+    [
+      "#22B8CF",
+      "#7fb0e0"
+    ],
+    [
+      "#0b0d10",
+      "#0a2233"
+    ],
+    [
+      "#14181f",
+      "#12314a"
+    ],
+    [
+      "#1a212b",
+      "#16344c"
+    ],
+    [
+      "#232a34",
+      "#23415a"
+    ],
+    [
+      "#0b1f19",
+      "#0e293c"
+    ],
+    [
+      "rgba(45,164,191,0.14)",
+      "rgba(127,176,224,0.16)"
+    ]
+  ],
+  "greenSwaps": [
+    [
+      "#0f9a63",
+      "#16344c"
+    ],
+    [
+      "#0c8554",
+      "#12314a"
+    ],
+    [
+      "#d9ecea",
+      "#d5e8f8"
+    ],
+    [
+      "#cfe3e2",
+      "#c2ddf4"
+    ],
+    [
+      "#1bb978",
+      "#7fb0e0"
+    ],
+    [
+      "#15ae6e",
+      "#0e3147"
+    ]
+  ],
+  "cssSwaps": [
+    [
+      "#0e1116",
+      "#0e293c"
+    ],
+    [
+      "#10222a",
+      "#0e293c"
+    ],
+    [
+      "#123047",
+      "#12314a"
+    ]
+  ],
+  "features": {
+    "askMode": "both",
+    "inline": true,
+    "action": true,
+    "monetize": true,
+    "tiers": true,
+    "pinning": true
+  },
+  "content": {
+    "company": "InTime",
+    "aiName": "InTime AI",
+    "website": "https://www.intime.com",
+    "loginTitleHtml": "Scheduling built for<br />public safety.",
+    "loginSubtitle": "Your command center for police, fire, corrections and dispatch \u2014 schedules, overtime and coverage, with AI-powered answers on live data. Built on InTime, powered by ThoughtSpot.",
+    "loginStats": [
+      {
+        "value": "24/7",
+        "label": "Coverage Managed"
+      },
+      {
+        "value": "98%",
+        "label": "Shifts Filled"
+      },
+      {
+        "value": "Real-time",
+        "label": "Overtime Intel"
+      }
+    ],
+    "loginCardSubtitle": "Sign in with your ThoughtSpot credentials",
+    "loginDemoNoteHtml": "<strong>Sign in</strong> with your ThoughtSpot username &amp; password for <code>{host}</code>. This origin must be CORS-allowlisted on that cluster for the embeds to load.",
+    "platformNav": [
+      "Police",
+      "Fire",
+      "Corrections"
+    ],
+    "tabs": {
+      "myReports": "My Reports",
+      "analytics": "Analytics",
+      "inline": "Shifts",
+      "action": "Leaves",
+      "ask": "Ask InTime AI",
+      "spotter": "Spotter"
+    },
+    "analyticsSubtitle": "Your live scheduling, overtime & coverage dashboard",
+    "spotterSubtitle": "Ask questions of your InTime scheduling & overtime data in natural language.",
+    "inline": {
+      "subtitleReady": "Overtime & hours by rank \u2014 select a rank to view its shift insights",
+      "subtitleIdle": "Overtime and hours across your ranks",
+      "loading": "Loading ranks\u2026",
+      "errorTitle": "Couldn't load shifts",
+      "errorMsg": "Unable to load shift data.",
+      "moreLabel": "View insights",
+      "hideLabel": "Hide insights"
+    },
+    "filters": {
+      "primaryLabel": "Branch",
+      "secondaryLabel": "Rank"
+    },
+    "sampleQuestions": [
+      "Overtime hours by rank",
+      "Overtime cost by branch",
+      "Worked hours by employee",
+      "Leave hours over time"
+    ],
+    "askWelcome": "Ask me anything about InTime \u2014 scheduling, overtime, leave, coverage, best practices \u2014 or ask about your own data and I will chart it live.",
+    "askEmptySub": "Ask any analytical question about your schedules, overtime, leave, and coverage. Answers come straight from your live data with an interactive chart you can drill into.",
+    "action": {
+      "label": "Request Leave",
+      "subtitlePrefix": "Officer schedules and coverage. Right-click a row and choose",
+      "subtitleSuffix": "to open a leave request pre-filled for that officer.",
+      "modalTitle": "New Leave Request",
+      "modalLead": "Pre-filled for the selected officer. Review and submit to request time off through InTime.",
+      "submitLabel": "Submit",
+      "successTitle": "Leave request submitted"
+    },
+    "chatbot": {
+      "welcome": "Hi! I'm InTime AI. Ask me about scheduling, overtime and coverage, or ask a question about your data.",
+      "overview": "InTime is workforce scheduling and time & attendance software built for public safety \u2014 police, sheriff, fire, corrections, and dispatch. It manages shift schedules, overtime, leave, court time, and certifications, with AI-powered answers on live workforce data.",
+      "greetingExample": "overtime hours by rank",
+      "examples": [
+        "worked hours by week",
+        "overtime cost by branch"
+      ]
+    },
+    "trialQuestionTrigger": 3,
+    "upgradeReason": "Upgrade to InTime Premium to unlock drill-down, Ask AI and downloads.",
+    "analyticsKeywords": [
+      "overtime",
+      "worked",
+      "leave",
+      "coverage",
+      "shift",
+      "shifts",
+      "headcount",
+      "staffing",
+      "rank",
+      "branch",
+      "officer",
+      "hours"
+    ]
+  }
+}

demoprep_app/portal/scripts/apply-spec.mjs

@@ -0,0 +1,294 @@
+#!/usr/bin/env node
+// ---------------------------------------------------------------------------
+// apply-spec.mjs — spec.json -> a fully-skinned <slug>-tse/ app.
+//
+//   node scripts/apply-spec.mjs <spec.json>
+//
+// Deterministic codemod for the mechanizable ~90% of a reskin: clone the
+// template, rename the brand token, write content.ts (all copy), patch config.ts
+// (host / GUIDs / columns / action / embed theme), patch auth + fonts, regenerate
+// the globals.css theme tokens, drop in the logo/favicon/font, and prune features
+// the spec turns off. Bespoke workflow modals (e.g. a custom leave form) are NOT
+// generated — they stay a manual step, and the script prints a reminder.
+//
+// Run from your project root. The skill is self-contained: the base template is
+// bundled next to this script (../template-tse), and the generated <slug>-tse/ app
+// is written into the current working directory. Paths in the spec (logo.svgPath,
+// font.srcPath) are resolved relative to the current working directory.
+// ---------------------------------------------------------------------------
+import fs from 'node:fs';
+import path from 'node:path';
+import { fileURLToPath } from 'node:url';
+import { execSync } from 'node:child_process';
+
+const ROOT = process.cwd();                                   // where the app is written
+const SKILL_DIR = path.resolve(path.dirname(fileURLToPath(import.meta.url)), '..');
+const TEMPLATE = path.join(SKILL_DIR, 'template-tse');        // bundled with the skill
+const SKIP = new Set(['node_modules', 'dist', '.vercel', '.git']);
+
+const specPath = process.argv[2];
+if (!specPath) { console.error('usage: node scripts/apply-spec.mjs <spec.json>'); process.exit(1); }
+const spec = JSON.parse(fs.readFileSync(specPath, 'utf8'));
+const t0 = Date.now();
+
+const slug = spec.slug;
+const client = spec.client;                    // Title-case, e.g. "InTime"
+const CLIENT = client.toUpperCase();
+const clientLower = slug;                       // lowercase token
+const OUT = path.join(ROOT, `${slug}-tse`);
+
+// Embed font: the brand font can't be pushed into ThoughtSpot's sandboxed iframe
+// unless it's on a CDN or uploaded to the cluster. So we request the brand font
+// first (renders if the cluster has it) and load the NEAREST Google font as the
+// reachable fallback. Map common brand faces -> closest free Google font.
+const gfontUrl = (fam) =>
+  `https://fonts.googleapis.com/css2?family=${fam.replace(/ /g, '+')}:wght@400;500;600;700&display=swap`;
+const FONT_FALLBACKS = {
+  larsseit: 'DM Sans', circular: 'Mulish', gilroy: 'Poppins', gotham: 'Montserrat',
+  avenir: 'Nunito Sans', 'avenir next': 'Nunito Sans', 'proxima nova': 'Montserrat',
+  metric: 'Inter', graphik: 'Public Sans', futura: 'Jost', 'brandon grotesque': 'Montserrat',
+  helvetica: 'Inter', arial: 'Inter',
+  // Google faces map to themselves (so an explicit Google font just works)
+  inter: 'Inter', roboto: 'Roboto', 'open sans': 'Open Sans', lato: 'Lato',
+  montserrat: 'Montserrat', poppins: 'Poppins', 'dm sans': 'DM Sans', mulish: 'Mulish',
+  nunito: 'Nunito', 'nunito sans': 'Nunito Sans', jost: 'Jost', manrope: 'Manrope',
+};
+const brandFamily = (spec.font && spec.font.family) || 'Inter';
+const fbFamily = FONT_FALLBACKS[brandFamily.toLowerCase()] || 'Inter';
+const embedFamily = `'${brandFamily}', '${fbFamily}', system-ui, -apple-system, 'Segoe UI', sans-serif`;
+const embedFontUrl = (spec.font && spec.font.googleUrl) || gfontUrl(fbFamily);
+// Host-UI font: a self-hosted file is @font-face'd; a Google font (given by name
+// or googleUrl) is loaded via index.html's <link>. isBrandGoogleFont is true when
+// the brand font IS itself a Google font (map returns it unchanged).
+const isBrandGoogleFont = fbFamily.toLowerCase() === brandFamily.toLowerCase() && brandFamily.toLowerCase() !== 'inter';
+const hostGoogleUrl = spec.font
+  ? (spec.font.googleUrl || (isBrandGoogleFont && !spec.font.file ? gfontUrl(brandFamily) : null))
+  : null;
+
+// ---- helpers ---------------------------------------------------------------
+function copyDir(src, dst) {
+  fs.mkdirSync(dst, { recursive: true });
+  for (const ent of fs.readdirSync(src, { withFileTypes: true })) {
+    if (SKIP.has(ent.name)) continue;
+    const s = path.join(src, ent.name), d = path.join(dst, ent.name);
+    if (ent.isDirectory()) copyDir(s, d);
+    else fs.copyFileSync(s, d);
+  }
+}
+function walk(dir, fn) {
+  for (const ent of fs.readdirSync(dir, { withFileTypes: true })) {
+    if (SKIP.has(ent.name)) continue;
+    const p = path.join(dir, ent.name);
+    if (ent.isDirectory()) walk(p, fn);
+    else fn(p);
+  }
+}
+const read = (p) => fs.readFileSync(p, 'utf8');
+const write = (p, s) => fs.writeFileSync(p, s);
+function edit(p, fn) { write(p, fn(read(p))); }
+function renameBrand(s) {
+  return s.split('Northwind').join(client)
+          .split('NORTHWIND').join(CLIENT)
+          .split('northwind').join(clientLower);
+}
+
+// ---- 0. clone --------------------------------------------------------------
+// Zero-setup: the generated apps share the bundled template's node_modules. If it
+// hasn't been installed yet (fresh clone of the skill), install it now — once —
+// instead of failing the first build with a confusing missing-deps error.
+ensureTemplateDeps();
+
+if (fs.existsSync(OUT)) fs.rmSync(OUT, { recursive: true, force: true });
+copyDir(TEMPLATE, OUT);
+// symlink deps to the bundled template's node_modules. Absolute target so it works
+// wherever the generated app is created.
+try { fs.symlinkSync(path.join(TEMPLATE, 'node_modules'), path.join(OUT, 'node_modules')); } catch {}
+
+function ensureTemplateDeps() {
+  // A real install has node_modules AND the SDK dep resolved (a bare/partial dir
+  // would still fail the build). Check for the SDK to catch a half-populated dir.
+  const nm = path.join(TEMPLATE, 'node_modules');
+  const sdk = path.join(nm, '@thoughtspot', 'visual-embed-sdk');
+  if (fs.existsSync(sdk)) return;
+  // node/npm often aren't on the default PATH (they live at ~/.node/bin here) —
+  // prepend it so `npm` resolves whether or not the caller exported it.
+  const home = process.env.HOME || '';
+  const PATH = `${path.join(home, '.node', 'bin')}:${process.env.PATH || ''}`;
+  console.log('[apply-spec] template deps not installed — running `npm install` once (this can take a minute)…');
+  try {
+    execSync('npm install --no-audit --no-fund', {
+      cwd: TEMPLATE,
+      stdio: 'inherit',
+      env: { ...process.env, PATH },
+    });
+  } catch {
+    console.error(
+      `[apply-spec] auto-install failed. Install manually, then re-run:\n` +
+      `  (cd "${TEMPLATE}" && npm install)`,
+    );
+    process.exit(1);
+  }
+}
+
+// ---- 1. brand rename (all casings) + file renames --------------------------
+walk(path.join(OUT, 'src'), (p) => {
+  if (/\.(ts|tsx|css)$/.test(p)) edit(p, renameBrand);
+});
+edit(path.join(OUT, 'index.html'), renameBrand);
+// Host webfont: point index.html's font <link> at the brand Google font when the
+// host font is a Google font (self-hosted fonts are handled by @font-face above).
+if (hostGoogleUrl) {
+  edit(path.join(OUT, 'index.html'), (s) =>
+    s.replace('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap', hostGoogleUrl));
+}
+edit(path.join(OUT, 'package.json'), (s) => s.replace(/northwind-thoughtspot-portal/g, `${slug}-thoughtspot-portal`));
+const mv = (a, b) => { const A = path.join(OUT, a), B = path.join(OUT, b); if (fs.existsSync(A)) fs.renameSync(A, B); };
+mv('src/components/NorthwindLogo.tsx', `src/components/${client}Logo.tsx`);
+mv('src/tabs/AskNorthwind.tsx', `src/tabs/Ask${client}.tsx`);
+
+// ---- 2. content.ts (all copy, from spec) -----------------------------------
+const contentFile =
+`// Generated by apply-spec.mjs — single source of truth for all portal copy.
+export interface LoginStat { value: string; label: string; }
+export const CONTENT = ${JSON.stringify(spec.content, null, 2)} as const;
+export type Content = typeof CONTENT;
+`;
+write(path.join(OUT, 'src/content.ts'), contentFile);
+
+// ---- 2b. flags.ts — feature toggles (monetize / tiers / pinning) -----------
+// Tabs (inline/action/askMode) are PRUNED below; these three are gated in-place
+// via FLAGS so shared components (paywall, tier switcher, Add-Report) render only
+// when on. Default true when the spec omits a flag.
+{
+  const ff = (spec.features || {});
+  const b = (v) => (v === false ? 'false' : 'true');
+  write(path.join(OUT, 'src/flags.ts'),
+    '// Generated by apply-spec.mjs from spec.features.\n' +
+    'export const FLAGS = {\n' +
+    `  monetize: ${b(ff.monetize)},\n` +
+    `  tiers: ${b(ff.tiers)},\n` +
+    `  pinning: ${b(ff.pinning)},\n` +
+    '} as const;\n');
+}
+
+// ---- 3. config.ts — IDs / host / action / columns / embed theme ------------
+edit(path.join(OUT, 'src/config.ts'), (s) => {
+  const i = spec.ids;
+  s = s.replace('https://YOUR-CLUSTER.thoughtspot.cloud', `https://${spec.host}`)
+       .replace('REPLACE_MAIN_LIVEBOARD_GUID', i.analyticsLiveboard)
+       .replace('REPLACE_INLINE_LIVEBOARD_GUID', i.inlineLiveboard)
+       .replace(/REPLACE_MODEL_GUID/g, i.model)
+       .replace('REPLACE_ACTION_LIVEBOARD_GUID', i.actionLiveboard)
+       .replace('REPLACE_ACTION_VIZ_GUID', i.actionViz)
+       .replace("'request-bid'", `'${spec.action.id}'`)
+       .replace("'Request Bid'", `'${spec.action.name}'`);
+  // candidate column arrays
+  const arr = (name, items) =>
+    s = s.replace(new RegExp(`export const ${name} = \\[[\\s\\S]*?\\];`),
+      `export const ${name} = [\n${items.map((x) => `  '${x}',`).join('\n')}\n];`);
+  arr('CADENCE_COLUMN_CANDIDATES', spec.columns.inlineName);
+  arr('REGION_COLUMN_CANDIDATES', spec.columns.filterPrimary);
+  arr('CORRIDOR_COLUMN_CANDIDATES', spec.columns.filterSecondary);
+  // inline detail metrics (CADENCE_DETAIL_COLUMNS)
+  const keys = spec.columns.inlineMetrics.map((_, n) => `metric${n + 1}`);
+  const body = spec.columns.inlineMetrics.map((m, n) =>
+    `  ${keys[n]}: {\n    label: '${m.label}',\n    format: '${m.format}',\n    candidates: [${m.candidates.map((c) => `'${c}'`).join(', ')}],\n  },`).join('\n');
+  s = s.replace(/export const CADENCE_DETAIL_COLUMNS: Record<[\s\S]*?\n> = \{[\s\S]*?\n\};/,
+    `export const CADENCE_DETAIL_COLUMNS: Record<\n  ${keys.map((k) => `'${k}'`).join(' | ')},\n  { label: string; format: 'text' | 'date' | 'number' | 'currency'; candidates: string[] }\n> = {\n${body}\n};`);
+  // hierarchy + date filter columns
+  s = s.replace(/export const SEGMENT_COLUMN = '[^']*';/, `export const SEGMENT_COLUMN = '${spec.columns.hierarchy.segment}';`)
+       .replace(/export const REP_COLUMN = '[^']*';/, `export const REP_COLUMN = '${spec.columns.hierarchy.rep}';`)
+       .replace(/export const CADENCE_NAME_COLUMN = '[^']*';/, `export const CADENCE_NAME_COLUMN = '${spec.columns.hierarchy.cadence}';`)
+       .replace(/export const DATE_COLUMN = '[^']*';/, `export const DATE_COLUMN = '${spec.columns.date}';`);
+  // embed theme hex swaps (indigo -> brand)
+  for (const [a, b] of spec.embedSwaps || []) s = s.split(a).join(b);
+  // Brand-tint the LIGHT embed surfaces (secondary buttons, chips, hovers, tile
+  // borders, action buttons, menus) by blending the primary toward white — so
+  // surfaces follow the brand instead of a flat grey. The neutral hexes below are
+  // the template's surface placeholders acting as swap keys.
+  {
+    const h2 = (h) => { h = h.replace('#', ''); return [parseInt(h.slice(0, 2), 16), parseInt(h.slice(2, 4), 16), parseInt(h.slice(4, 6), 16)]; };
+    const x = (n) => Math.round(n).toString(16).padStart(2, '0');
+    const mixW = (hex, t) => { const [r, g, b] = h2(hex); return '#' + x(r + (255 - r) * t) + x(g + (255 - g) * t) + x(b + (255 - b) * t); };
+    const prim = (spec.theme && spec.theme.light && spec.theme.light['--sl-green']) || '#4F5BD5';
+    const surf = {
+      '#FCFCFD': mixW(prim, 0.975), '#F5F6F8': mixW(prim, 0.96), '#F5F6F9': mixW(prim, 0.96),
+      '#EEF0F4': mixW(prim, 0.92), '#EEF0F5': mixW(prim, 0.92), '#F1F2F6': mixW(prim, 0.945),
+      '#E3E6EC': mixW(prim, 0.86), '#E8EAF1': mixW(prim, 0.90), '#E2E4EC': mixW(prim, 0.88),
+      '#D7DBE3': mixW(prim, 0.80),
+    };
+    for (const [g, v] of Object.entries(surf)) s = s.split(g).join(v);
+  }
+  // embed font: brand font first, nearest reachable Google font as fallback
+  s = s.replace(/export const EMBED_FONT_FAMILY = '[^']*';/,
+    `export const EMBED_FONT_FAMILY = ${JSON.stringify(embedFamily)};`);
+  // the stylesheet actually loaded into the iframe — always a reliable Google URL
+  s = s.replace(/export const TS_FONT_URL =\n  '[^']*';/,
+    `export const TS_FONT_URL =\n  ${JSON.stringify(embedFontUrl)};`);
+  return s;
+});
+
+// ---- 4. thoughtspot.ts — auth + embed font + accent swaps ------------------
+edit(path.join(OUT, 'src/lib/thoughtspot.ts'), (s) => {
+  if (spec.auth === 'basic') {
+    s = s.replace(/void username;\n\s*void password;\n/, '')
+         .replace('authType: AuthType.None,', 'authType: AuthType.Basic,\n    username,\n    password,');
+  }
+  s = s.replace(/const APP_FONT = "[^"]*";/, `const APP_FONT = ${JSON.stringify(embedFamily)};`);
+  for (const [a, b] of spec.embedSwaps || []) s = s.split(a).join(b);
+  for (const [a, b] of spec.greenSwaps || []) s = s.split(a).join(b);
+  return s;
+});
+
+// ---- 5. globals.css — regenerate theme tokens + @font-face -----------------
+edit(path.join(OUT, 'src/styles/globals.css'), (s) => {
+  // Guarantee the fixed brand tokens the chatbot depends on — if the spec's theme
+  // dicts omit them, the chat FAB/header background (var(--brand-accent/navy)) would
+  // be undefined and render transparent. Default from the light primary + ink.
+  for (const key of ['light', 'dark']) {
+    const d = spec.theme[key] || {};
+    if (!d['--brand-accent']) d['--brand-accent'] = spec.theme.light['--sl-green'] || '#4F5BD5';
+    if (!d['--brand-navy']) d['--brand-navy'] = spec.theme.light['--sl-evergreen'] || '#12203a';
+    spec.theme[key] = d;
+  }
+  const block = (dict) => Object.entries(dict).map(([k, v]) => `  ${k}: ${v};`).join('\n');
+  s = s.replace(/:root \{\n[\s\S]*?\n\}/, `:root {\n${block(spec.theme.light)}\n}`);
+  s = s.replace(/:root\[data-theme='dark'\] \{\n[\s\S]*?\n\}/, `:root[data-theme='dark'] {\n${block(spec.theme.dark)}\n}`);
+  // leftover dark-override rules (login-left gradient, branded header fills)
+  for (const [a, b] of spec.cssSwaps || []) s = s.split(a).join(b);
+  if (spec.font && spec.font.file) {
+    const face =
+`@font-face {\n  font-family: '${spec.font.family}';\n  font-style: normal;\n  font-weight: normal;\n  font-display: swap;\n  src: url('/${spec.font.file}') format('woff2');\n}\n\n`;
+    s = face + s;
+  }
+  return s;
+});
+
+// ---- 6. theme default + assets --------------------------------------------
+edit(path.join(OUT, 'src/context/ThemeContext.tsx'), (s) =>
+  s.replace(/: 'dark';/, `: '${spec.theme.default}';`).replace(/: 'light';/, `: '${spec.theme.default}';`));
+if (spec.logo?.svgPath) {
+  const svg = read(path.join(ROOT, spec.logo.svgPath)).replace(/fill:\s*#[0-9A-Fa-f]{3,6}/g, 'fill:currentColor');
+  write(path.join(OUT, `src/assets/${slug}-logo.svg`), svg);
+}
+if (spec.favicon) write(path.join(OUT, `public/${slug}-icon.svg`), spec.favicon);
+if (spec.font?.srcPath) fs.copyFileSync(path.join(ROOT, spec.font.srcPath), path.join(OUT, `public/${spec.font.file}`));
+// remove stale template assets
+for (const f of ['public/northwind-icon.svg', 'src/assets/northwind-logo.svg', 'public/figma-reference.png'])
+  { const p = path.join(OUT, f); if (fs.existsSync(p)) fs.rmSync(p); }
+
+// ---- 7. prune features the spec turns off ----------------------------------
+const f = spec.features || {};
+const rmTab = (id, extraFiles = []) => {
+  edit(path.join(OUT, 'src/App.tsx'), (s) => s.replace(new RegExp(`\\s*\\{tab === '${id}' && <[^>]+>\\}`), ''));
+  edit(path.join(OUT, 'src/components/TopBar.tsx'), (s) => s.replace(new RegExp(`\\s*\\{ id: '${id}',[^\\n]*\\},`), ''));
+  for (const rel of extraFiles) { const p = path.join(OUT, rel); if (fs.existsSync(p)) fs.rmSync(p); }
+};
+if (f.inline === false) rmTab('carriers', ['src/tabs/Carriers.tsx']);
+if (f.action === false) rmTab('capacity', ['src/tabs/Capacity.tsx', 'src/components/BidModal.tsx']);
+if (f.askMode === 'fancy') rmTab('spotter', ['src/tabs/Spotter.tsx']);
+if (f.askMode === 'normal') rmTab('ask', [`src/tabs/Ask${client}.tsx`]);
+
+console.log(`\n[apply-spec] ${client} -> ${slug}-tse/ in ${((Date.now() - t0) / 1000).toFixed(2)}s`);
+if (f.action !== false)
+  console.log('[apply-spec] NOTE: the custom-action modal (BidModal) is generic — customize it by hand for a bespoke workflow.');

demoprep_app/portal/specs/triumph.json

@@ -0,0 +1,239 @@
+{
+  "client": "TriumphFinancial",
+  "slug": "triumph",
+  "host": "sebe.thoughtspotstaging.cloud",
+  "auth": "none",
+  "ids": {
+    "analyticsLiveboard": "ca684d71-dd94-4d70-a24a-efb2e443f9ea",
+    "model": "276f6862-3fe1-4df8-9ae0-12912c7f54c0",
+    "inlineLiveboard": "ca684d71-dd94-4d70-a24a-efb2e443f9ea",
+    "actionLiveboard": "ca684d71-dd94-4d70-a24a-efb2e443f9ea",
+    "actionViz": "ca684d71-dd94-4d70-a24a-efb2e443f9ea"
+  },
+  "action": {
+    "id": "request-capacity",
+    "name": "Request Capacity"
+  },
+  "columns": {
+    "inlineName": [
+      "Broker Name",
+      "Broker Tier"
+    ],
+    "inlineMetrics": [
+      {
+        "label": "Load Revenue",
+        "format": "currency",
+        "candidates": [
+          "Total Load Revenue Usd",
+          "Total Load Revenue"
+        ]
+      },
+      {
+        "label": "Broker Margin %",
+        "format": "number",
+        "candidates": [
+          "Broker Margin Pct",
+          "Broker Margin"
+        ]
+      }
+    ],
+    "filterPrimary": [
+      "Broker Name",
+      "Broker Tier"
+    ],
+    "filterSecondary": [
+      "Lane Name",
+      "Lane Tier"
+    ],
+    "date": "Load Date",
+    "hierarchy": {
+      "segment": "Broker Tier",
+      "rep": "Broker Name",
+      "cadence": "Lane Name"
+    }
+  },
+  "logo": {
+    "svgPath": "triumph-logo.svg"
+  },
+  "favicon": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 32 32\" width=\"32\" height=\"32\"><rect width=\"32\" height=\"32\" rx=\"7\" fill=\"#10294a\"/><path d=\"M8 8 H24 V13 H18.5 V26 H13.5 V13 H8 Z\" fill=\"#e8b53c\"/></svg>",
+  "theme": {
+    "default": "light",
+    "light": {
+      "--sl-cream": "#f5f8fc",
+      "--sl-cream-2": "#eaf1f8",
+      "--sl-white": "#ffffff",
+      "--sl-evergreen": "#10294a",
+      "--sl-evergreen-2": "#0a1d37",
+      "--sl-green": "#10294a",
+      "--sl-green-dark": "#1a3f6e",
+      "--sl-primary-rgb": "16, 41, 74",
+      "--intime-blue": "#2e86d1",
+      "--sl-gold": "#e8b53c",
+      "--sl-gold-ink": "#5f4708",
+      "--intime-grad": "linear-gradient(90deg, #10294a 0%, #2e86d1 100%)",
+      "--sl-mint": "#e5f2fc",
+      "--sl-mint-2": "#d2e8fa",
+      "--sl-coral": "#a2191f",
+      "--sl-warn": "#fffbe6",
+      "--sl-warn-ink": "#5f4708",
+      "--sl-ink": "#11253f",
+      "--sl-muted": "#5a6a7e",
+      "--sl-border": "#e0e8f0",
+      "--sl-border-soft": "#edf2f8",
+      "--font-serif": "'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--font-sans": "'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--topbar-h": "66px",
+      "--shadow-card": "0 1px 2px rgba(15, 23, 42, 0.05), 0 8px 24px rgba(15, 23, 42, 0.07)",
+      "--shadow-lift": "0 6px 14px rgba(15, 23, 42, 0.08), 0 18px 40px rgba(15, 23, 42, 0.12)",
+      "--brand-navy": "#10294a",
+      "--brand-accent": "#2e86d1"
+    },
+    "dark": {
+      "--sl-cream": "#0a1a2e",
+      "--sl-cream-2": "#0e2238",
+      "--sl-white": "#122c48",
+      "--sl-evergreen": "#eef4fa",
+      "--sl-evergreen-2": "#dbe6f0",
+      "--sl-green": "#7ab4e8",
+      "--sl-green-dark": "#5a9bd6",
+      "--sl-primary-rgb": "122, 180, 232",
+      "--intime-blue": "#7ab4e8",
+      "--sl-gold": "#e8b53c",
+      "--sl-gold-ink": "#0a1a2e",
+      "--intime-grad": "linear-gradient(90deg, #7ab4e8 0%, #bfe0f7 100%)",
+      "--sl-mint": "#0f2a42",
+      "--sl-mint-2": "#153450",
+      "--sl-coral": "#f0616a",
+      "--sl-warn": "#2a2413",
+      "--sl-warn-ink": "#e8b53c",
+      "--sl-ink": "#eaf1f7",
+      "--sl-muted": "#9fb2c6",
+      "--sl-border": "#22405e",
+      "--sl-border-soft": "#162e48",
+      "--font-serif": "'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--font-sans": "'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--topbar-h": "66px",
+      "--shadow-card": "0 1px 2px rgba(0, 0, 0, 0.5), 0 8px 24px rgba(0, 0, 0, 0.55)",
+      "--shadow-lift": "0 6px 14px rgba(0, 0, 0, 0.6), 0 18px 40px rgba(0, 0, 0, 0.65)",
+      "--brand-navy": "#10294a",
+      "--brand-accent": "#2e86d1"
+    }
+  },
+  "embedSwaps": [
+    ["#4F5BD5", "#2e86d1"],
+    ["#3F49B8", "#2673b8"],
+    ["#333C99", "#1f5f99"],
+    ["#22B8CF", "#7ab4e8"],
+    ["#0b0d10", "#0a1a2e"],
+    ["#14181f", "#122c48"],
+    ["#1a212b", "#153450"],
+    ["#232a34", "#22405e"],
+    ["#0b1f19", "#0e2238"],
+    ["rgba(45,164,191,0.14)", "rgba(122,180,232,0.16)"]
+  ],
+  "greenSwaps": [
+    ["#0f9a63", "#153450"],
+    ["#0c8554", "#122c48"],
+    ["#d9ecea", "#d2e8fa"],
+    ["#cfe3e2", "#c0ddf6"],
+    ["#1bb978", "#7ab4e8"],
+    ["#15ae6e", "#10294a"]
+  ],
+  "cssSwaps": [
+    ["#0e1116", "#0e2238"],
+    ["#10222a", "#0e2238"],
+    ["#123047", "#122c48"]
+  ],
+  "features": {
+    "askMode": "both",
+    "inline": false,
+    "action": false,
+    "monetize": false,
+    "tiers": false,
+    "pinning": false
+  },
+  "content": {
+    "company": "Triumph Financial",
+    "aiName": "Triumph AI",
+    "website": "https://www.triumph.io",
+    "loginTitleHtml": "Freight payments,<br />intelligently moved.",
+    "loginSubtitle": "Your broker capacity command center — loads, lanes, rates, margins and carrier capacity across the TriumphPay network, with AI-powered answers on live data. Built on Triumph Financial, powered by ThoughtSpot.",
+    "loginStats": [
+      { "value": "$30B+", "label": "Payments Network" },
+      { "value": "Live", "label": "Lane & Rate Intel" },
+      { "value": "Real-time", "label": "Capacity Signals" }
+    ],
+    "loginCardSubtitle": "Sign in with your ThoughtSpot credentials",
+    "loginDemoNoteHtml": "<strong>Sign in</strong> with your ThoughtSpot username &amp; password for <code>{host}</code>. This origin must be CORS-allowlisted on that cluster for the embeds to load.",
+    "platformNav": [
+      "Brokers",
+      "Carriers",
+      "Lanes"
+    ],
+    "tabs": {
+      "myReports": "My Reports",
+      "analytics": "Capacity Analytics",
+      "inline": "Brokers",
+      "action": "Capacity",
+      "ask": "Ask Triumph AI",
+      "spotter": "Spotter"
+    },
+    "analyticsSubtitle": "Your live broker, lane & carrier-capacity dashboard",
+    "spotterSubtitle": "Ask questions of your Triumph Financial freight & capacity data in natural language.",
+    "inline": {
+      "subtitleReady": "Revenue & margin by broker — select a broker to view its lane insights",
+      "subtitleIdle": "Revenue and margin across your broker network",
+      "loading": "Loading brokers…",
+      "errorTitle": "Couldn't load brokers",
+      "errorMsg": "Unable to load broker data.",
+      "moreLabel": "View insights",
+      "hideLabel": "Hide insights"
+    },
+    "filters": {
+      "primaryLabel": "Broker",
+      "secondaryLabel": "Lane"
+    },
+    "sampleQuestions": [
+      "Broker margin by broker tier",
+      "Load revenue by lane over time",
+      "Capacity utilization rate by month",
+      "On time delivery rate by carrier"
+    ],
+    "askWelcome": "Ask me anything about Triumph Financial — loads, lanes, rates, broker margins, carrier capacity — or ask about your own data and I will chart it live.",
+    "askEmptySub": "Ask any analytical question about your loads, lanes, broker margins, and carrier capacity. Answers come straight from your live data with an interactive chart you can drill into.",
+    "action": {
+      "label": "Request Capacity",
+      "subtitlePrefix": "Loads by broker and lane. Right-click a row and choose",
+      "subtitleSuffix": "to open a capacity request pre-filled for that lane.",
+      "modalTitle": "New Capacity Request",
+      "modalLead": "Pre-filled for the selected lane. Review and submit to request carrier capacity through TriumphPay.",
+      "submitLabel": "Submit",
+      "successTitle": "Capacity request submitted"
+    },
+    "chatbot": {
+      "welcome": "Hi! I'm Triumph AI. Ask me about loads, lanes, broker margins and carrier capacity, or ask a question about your data.",
+      "overview": "Triumph Financial is a transportation-fintech company: TBK Bank, Triumph Business Capital (invoice factoring for carriers), and TriumphPay, a freight-payments network digitizing invoice presentment, audit and settlement among brokers, carriers, shippers and factors. This portal covers broker capacity intelligence — loads, lanes, rates, margins and carrier capacity — with AI-powered answers on live data.",
+      "greetingExample": "broker margin by tier",
+      "examples": [
+        "load revenue by lane",
+        "capacity utilization by month"
+      ]
+    },
+    "trialQuestionTrigger": 3,
+    "upgradeReason": "Upgrade to TriumphPay Premium to unlock drill-down, Ask AI and downloads.",
+    "analyticsKeywords": [
+      "load",
+      "loads",
+      "lane",
+      "broker",
+      "carrier",
+      "margin",
+      "revenue",
+      "rate",
+      "capacity",
+      "utilization",
+      "dispute",
+      "on-time"
+    ]
+  }
+}

demoprep_app/portal/specs/vantage.json

@@ -0,0 +1,238 @@
+{
+  "client": "VantageCapital",
+  "slug": "vantage",
+  "host": "sebe.thoughtspotstaging.cloud",
+  "auth": "none",
+  "ids": {
+    "analyticsLiveboard": "a476cb7d-5b61-4658-86ca-5f078a212a0d",
+    "model": "d9d910d7-d3b5-4a7a-a35e-de494405055f",
+    "inlineLiveboard": "a476cb7d-5b61-4658-86ca-5f078a212a0d",
+    "actionLiveboard": "a476cb7d-5b61-4658-86ca-5f078a212a0d",
+    "actionViz": "a476cb7d-5b61-4658-86ca-5f078a212a0d"
+  },
+  "action": {
+    "id": "flag-for-review",
+    "name": "Flag for Review"
+  },
+  "columns": {
+    "inlineName": [
+      "Borrower Industry Name",
+      "Industry Sector"
+    ],
+    "inlineMetrics": [
+      {
+        "label": "Outstanding Balance",
+        "format": "currency",
+        "candidates": [
+          "Outstanding Balance Usd",
+          "Outstanding Balance"
+        ]
+      },
+      {
+        "label": "Delinquency Rate",
+        "format": "number",
+        "candidates": [
+          "Delinquency Rate"
+        ]
+      }
+    ],
+    "filterPrimary": [
+      "Borrower Industry Name",
+      "Industry Sector"
+    ],
+    "filterSecondary": [
+      "Risk Grade Name",
+      "Grade Tier"
+    ],
+    "date": "Snapshot Month",
+    "hierarchy": {
+      "segment": "Borrower Industry Name",
+      "rep": "Risk Grade Name",
+      "cadence": "Region Name"
+    }
+  },
+  "logo": {
+    "svgPath": "vantage-logo.svg"
+  },
+  "favicon": "<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 32 32\" width=\"32\" height=\"32\"><rect width=\"32\" height=\"32\" rx=\"7\" fill=\"#12283f\"/><path d=\"M8 8 L16 26 L24 8 L19 8 L16 16 L13 8 Z\" fill=\"#f4d14e\"/></svg>",
+  "theme": {
+    "default": "light",
+    "light": {
+      "--sl-cream": "#f4f7fb",
+      "--sl-cream-2": "#e9f0f7",
+      "--sl-white": "#ffffff",
+      "--sl-evergreen": "#12283f",
+      "--sl-evergreen-2": "#0c1d30",
+      "--sl-green": "#12283f",
+      "--sl-green-dark": "#1b3f63",
+      "--sl-primary-rgb": "18, 40, 63",
+      "--intime-blue": "#2f6fb0",
+      "--sl-gold": "#f4d14e",
+      "--sl-gold-ink": "#6b5410",
+      "--intime-grad": "linear-gradient(90deg, #12283f 0%, #2f6fb0 100%)",
+      "--sl-mint": "#e6f1fb",
+      "--sl-mint-2": "#d5e8f8",
+      "--sl-coral": "#a2191f",
+      "--sl-warn": "#fffbe6",
+      "--sl-warn-ink": "#6b5410",
+      "--sl-ink": "#12283f",
+      "--sl-muted": "#5a6a7e",
+      "--sl-border": "#e1e8f0",
+      "--sl-border-soft": "#eef2f7",
+      "--font-serif": "'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--font-sans": "'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--topbar-h": "66px",
+      "--shadow-card": "0 1px 2px rgba(15, 23, 42, 0.05), 0 8px 24px rgba(15, 23, 42, 0.07)",
+      "--shadow-lift": "0 6px 14px rgba(15, 23, 42, 0.08), 0 18px 40px rgba(15, 23, 42, 0.12)",
+      "--brand-navy": "#12283f",
+      "--brand-accent": "#2f6fb0"
+    },
+    "dark": {
+      "--sl-cream": "#0a1a2b",
+      "--sl-cream-2": "#0e2236",
+      "--sl-white": "#122c44",
+      "--sl-evergreen": "#eef4fa",
+      "--sl-evergreen-2": "#dbe6f0",
+      "--sl-green": "#7fb0e0",
+      "--sl-green-dark": "#5e97cc",
+      "--sl-primary-rgb": "127, 176, 224",
+      "--intime-blue": "#7fb0e0",
+      "--sl-gold": "#f4d14e",
+      "--sl-gold-ink": "#0a1a2b",
+      "--intime-grad": "linear-gradient(90deg, #7fb0e0 0%, #bfe0f5 100%)",
+      "--sl-mint": "#10293d",
+      "--sl-mint-2": "#16344c",
+      "--sl-coral": "#f0616a",
+      "--sl-warn": "#2a2413",
+      "--sl-warn-ink": "#f4d14e",
+      "--sl-ink": "#eaf1f7",
+      "--sl-muted": "#9fb2c6",
+      "--sl-border": "#23415a",
+      "--sl-border-soft": "#172f44",
+      "--font-serif": "'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--font-sans": "'Inter', system-ui, -apple-system, 'Segoe UI', sans-serif",
+      "--topbar-h": "66px",
+      "--shadow-card": "0 1px 2px rgba(0, 0, 0, 0.5), 0 8px 24px rgba(0, 0, 0, 0.55)",
+      "--shadow-lift": "0 6px 14px rgba(0, 0, 0, 0.6), 0 18px 40px rgba(0, 0, 0, 0.65)",
+      "--brand-navy": "#12283f",
+      "--brand-accent": "#2f6fb0"
+    }
+  },
+  "embedSwaps": [
+    ["#4F5BD5", "#2f6fb0"],
+    ["#3F49B8", "#285f97"],
+    ["#333C99", "#22527f"],
+    ["#22B8CF", "#7fb0e0"],
+    ["#0b0d10", "#0a1a2b"],
+    ["#14181f", "#122c44"],
+    ["#1a212b", "#16344c"],
+    ["#232a34", "#23415a"],
+    ["#0b1f19", "#0e2236"],
+    ["rgba(45,164,191,0.14)", "rgba(127,176,224,0.16)"]
+  ],
+  "greenSwaps": [
+    ["#0f9a63", "#16344c"],
+    ["#0c8554", "#122c44"],
+    ["#d9ecea", "#d5e8f8"],
+    ["#cfe3e2", "#c2ddf4"],
+    ["#1bb978", "#7fb0e0"],
+    ["#15ae6e", "#12283f"]
+  ],
+  "cssSwaps": [
+    ["#0e1116", "#0e2236"],
+    ["#10222a", "#0e2236"],
+    ["#123047", "#122c44"]
+  ],
+  "features": {
+    "askMode": "both",
+    "inline": false,
+    "action": false,
+    "monetize": false,
+    "tiers": false,
+    "pinning": false
+  },
+  "content": {
+    "company": "Vantage Capital",
+    "aiName": "Vantage AI",
+    "website": "https://www.vantagecapital.example.com",
+    "loginTitleHtml": "Credit risk,<br />seen clearly.",
+    "loginSubtitle": "Your portfolio command center for commercial lending — outstanding balances, delinquency, charge-offs and provision coverage, with AI-powered answers on live data. Built on Vantage Capital, powered by ThoughtSpot.",
+    "loginStats": [
+      { "value": "$4.2B", "label": "Portfolio Managed" },
+      { "value": "7", "label": "Risk Grades Tracked" },
+      { "value": "Real-time", "label": "Delinquency Intel" }
+    ],
+    "loginCardSubtitle": "Sign in with your ThoughtSpot credentials",
+    "loginDemoNoteHtml": "<strong>Sign in</strong> with your ThoughtSpot username &amp; password for <code>{host}</code>. This origin must be CORS-allowlisted on that cluster for the embeds to load.",
+    "platformNav": [
+      "Portfolio",
+      "Credit Risk",
+      "Reviews"
+    ],
+    "tabs": {
+      "myReports": "My Reports",
+      "analytics": "Portfolio Analytics",
+      "inline": "Industries",
+      "action": "Reviews",
+      "ask": "Ask Vantage AI",
+      "spotter": "Spotter"
+    },
+    "analyticsSubtitle": "Your live portfolio, delinquency & credit-risk dashboard",
+    "spotterSubtitle": "Ask questions of your Vantage Capital portfolio & credit-risk data in natural language.",
+    "inline": {
+      "subtitleReady": "Balance & delinquency by industry — select an industry to view its insights",
+      "subtitleIdle": "Balance and delinquency across borrower industries",
+      "loading": "Loading industries…",
+      "errorTitle": "Couldn't load industries",
+      "errorMsg": "Unable to load industry data.",
+      "moreLabel": "View insights",
+      "hideLabel": "Hide insights"
+    },
+    "filters": {
+      "primaryLabel": "Industry",
+      "secondaryLabel": "Risk Grade"
+    },
+    "sampleQuestions": [
+      "Delinquency rate by borrower industry",
+      "Outstanding balance by risk grade",
+      "Charge off rate by region over time",
+      "Provision coverage ratio by product family"
+    ],
+    "askWelcome": "Ask me anything about Vantage Capital — portfolio exposure, delinquency, charge-offs, credit reviews — or ask about your own data and I will chart it live.",
+    "askEmptySub": "Ask any analytical question about your loan portfolio, delinquency, risk grades, and credit reviews. Answers come straight from your live data with an interactive chart you can drill into.",
+    "action": {
+      "label": "Flag for Review",
+      "subtitlePrefix": "Loan exposures by industry and grade. Right-click a row and choose",
+      "subtitleSuffix": "to open a credit review pre-filled for that exposure.",
+      "modalTitle": "New Credit Review",
+      "modalLead": "Pre-filled for the selected exposure. Review and submit to open a credit review in Vantage.",
+      "submitLabel": "Submit",
+      "successTitle": "Credit review opened"
+    },
+    "chatbot": {
+      "welcome": "Hi! I'm Vantage AI. Ask me about portfolio exposure, delinquency and credit risk, or ask a question about your data.",
+      "overview": "Vantage Capital is a commercial lender. This portal covers portfolio and credit-risk analytics — outstanding balances, exposure at default, delinquency and charge-off rates, provision coverage, and credit review activity — with AI-powered answers on live portfolio data.",
+      "greetingExample": "delinquency rate by industry",
+      "examples": [
+        "outstanding balance by risk grade",
+        "charge off rate by region"
+      ]
+    },
+    "trialQuestionTrigger": 3,
+    "upgradeReason": "Upgrade to Vantage Premium to unlock drill-down, Ask AI and downloads.",
+    "analyticsKeywords": [
+      "delinquency",
+      "balance",
+      "exposure",
+      "charge-off",
+      "provision",
+      "risk",
+      "grade",
+      "industry",
+      "region",
+      "loan",
+      "portfolio",
+      "review"
+    ]
+  }
+}

demoprep_app/portal/template-tse/.gitignore

@@ -0,0 +1,20 @@
+node_modules/
+dist/
+.DS_Store
+*.local
+.env
+.env.*
+!.env.example
+.vercel/
+
+# Large local recordings
+*.mov
+
+# Unrelated demos / local tooling kept in this working dir (repo = Salesloft app only)
+clari/
+hirevue/
+hivewatch/
+simfoni/
+output/
+.vercel
+.env*

demoprep_app/portal/template-tse/README.md

@@ -0,0 +1,44 @@
+# Acme × ThoughtSpot — Embedded Analytics Portal
+
+A demo revenue-intelligence portal styled as an **Acme** product surface, with
+ThoughtSpot embedded throughout via the
+[Visual Embed SDK](https://developers.thoughtspot.com/docs). Built with Vite +
+React + TypeScript. Deploys to Vercel.
+
+This is a rebranded twin of the Salesloft demo — identical functionality, theme,
+and ThoughtSpot data; only the logo and the **Acme / Acme AI** branding differ.
+
+## What's in it
+
+| Tab | ThoughtSpot embed | Notes |
+| --- | --- | --- |
+| **My Reports** | — | Landing / overview |
+| **Analytics** | `LiveboardEmbed` | Full liveboard with a **Rep** dropdown runtime filter (`HostEvent.UpdateRuntimeFilters`) |
+| **Cadences** | — | Static view |
+| **Signals** | `LiveboardEmbed` (single viz) | One visualization from the Analytics liveboard; right-click a row → **Re-engage cadence** custom action opens a Win-Back Cadence modal |
+| **Ask Acme** | `SpotterEmbed` | "Acme AI" chat: general questions answered from a small FAQ/knowledge base, data questions routed to Spotter via `HostEvent.SpotterSearch`; playful trial-upgrade modal |
+
+## Getting started
+
+```bash
+npm install
+npm run dev      # local dev server
+npm run build    # tsc + vite build -> dist/
+npm run preview  # preview the production build
+```
+
+> If `node`/`npm` aren't on your PATH, they may live at `~/.node/bin`:
+> `export PATH="$HOME/.node/bin:$PATH"`.
+
+## Configuration
+
+- **ThoughtSpot IDs & flags** live in [`src/config.ts`](src/config.ts).
+- **ThoughtSpot host** and auth are set in [`src/lib/thoughtspot.ts`](src/lib/thoughtspot.ts).
+- **Optional LLM routing** for Ask Acme — copy `.env.example` to `.env` and set
+  `VITE_ANTHROPIC_API_KEY`. Without it, the chatbot falls back to a keyword
+  router + FAQ and still routes data questions to Spotter.
+
+## Deploy
+
+Configured for Vercel ([`vercel.json`](vercel.json)) — framework `vite`,
+output `dist/`, SPA rewrite to `/index.html`.

demoprep_app/portal/template-tse/index.html

@@ -0,0 +1,19 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="UTF-8" />
+    <link rel="icon" type="image/svg+xml" href="/northwind-icon.svg" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
+    <title>Northwind Analytics</title>
+    <link rel="preconnect" href="https://fonts.googleapis.com" />
+    <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin />
+    <link
+      href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap"
+      rel="stylesheet"
+    />
+  </head>
+  <body>
+    <div id="root"></div>
+    <script type="module" src="/src/main.tsx"></script>
+  </body>
+</html>

demoprep_app/portal/template-tse/package-lock.json

@@ -0,0 +1,1934 @@
+{
+  "name": "northwind-thoughtspot-portal",
+  "version": "1.0.0",
+  "lockfileVersion": 3,
+  "requires": true,
+  "packages": {
+    "": {
+      "name": "northwind-thoughtspot-portal",
+      "version": "1.0.0",
+      "dependencies": {
+        "@thoughtspot/visual-embed-sdk": "^1.37.0",
+        "@vercel/analytics": "^2.0.1",
+        "lucide-react": "^0.300.0",
+        "react": "^18.2.0",
+        "react-dom": "^18.2.0",
+        "react-router-dom": "^6.21.0"
+      },
+      "devDependencies": {
+        "@types/react": "^18.2.45",
+        "@types/react-dom": "^18.2.18",
+        "@vitejs/plugin-react": "^4.2.1",
+        "typescript": "^5.3.3",
+        "vite": "^5.0.10"
+      }
+    },
+    "node_modules/@babel/code-frame": {
+      "version": "7.29.0",
+      "resolved": "https://registry.npmjs.org/@babel/code-frame/-/code-frame-7.29.0.tgz",
+      "integrity": "sha512-9NhCeYjq9+3uxgdtp20LSiJXJvN0FeCtNGpJxuMFZ1Kv3cWUNb6DOhJwUvcVCzKGR66cw4njwM6hrJLqgOwbcw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/helper-validator-identifier": "^7.28.5",
+        "js-tokens": "^4.0.0",
+        "picocolors": "^1.1.1"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/compat-data": {
+      "version": "7.29.0",
+      "resolved": "https://registry.npmjs.org/@babel/compat-data/-/compat-data-7.29.0.tgz",
+      "integrity": "sha512-T1NCJqT/j9+cn8fvkt7jtwbLBfLC/1y1c7NtCeXFRgzGTsafi68MRv8yzkYSapBnFA6L3U2VSc02ciDzoAJhJg==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/core": {
+      "version": "7.29.0",
+      "resolved": "https://registry.npmjs.org/@babel/core/-/core-7.29.0.tgz",
+      "integrity": "sha512-CGOfOJqWjg2qW/Mb6zNsDm+u5vFQ8DxXfbM09z69p5Z6+mE1ikP2jUXw+j42Pf1XTYED2Rni5f95npYeuwMDQA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/code-frame": "^7.29.0",
+        "@babel/generator": "^7.29.0",
+        "@babel/helper-compilation-targets": "^7.28.6",
+        "@babel/helper-module-transforms": "^7.28.6",
+        "@babel/helpers": "^7.28.6",
+        "@babel/parser": "^7.29.0",
+        "@babel/template": "^7.28.6",
+        "@babel/traverse": "^7.29.0",
+        "@babel/types": "^7.29.0",
+        "@jridgewell/remapping": "^2.3.5",
+        "convert-source-map": "^2.0.0",
+        "debug": "^4.1.0",
+        "gensync": "^1.0.0-beta.2",
+        "json5": "^2.2.3",
+        "semver": "^6.3.1"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      },
+      "funding": {
+        "type": "opencollective",
+        "url": "https://opencollective.com/babel"
+      }
+    },
+    "node_modules/@babel/generator": {
+      "version": "7.29.1",
+      "resolved": "https://registry.npmjs.org/@babel/generator/-/generator-7.29.1.tgz",
+      "integrity": "sha512-qsaF+9Qcm2Qv8SRIMMscAvG4O3lJ0F1GuMo5HR/Bp02LopNgnZBC/EkbevHFeGs4ls/oPz9v+Bsmzbkbe+0dUw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/parser": "^7.29.0",
+        "@babel/types": "^7.29.0",
+        "@jridgewell/gen-mapping": "^0.3.12",
+        "@jridgewell/trace-mapping": "^0.3.28",
+        "jsesc": "^3.0.2"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/helper-compilation-targets": {
+      "version": "7.28.6",
+      "resolved": "https://registry.npmjs.org/@babel/helper-compilation-targets/-/helper-compilation-targets-7.28.6.tgz",
+      "integrity": "sha512-JYtls3hqi15fcx5GaSNL7SCTJ2MNmjrkHXg4FSpOA/grxK8KwyZ5bubHsCq8FXCkua6xhuaaBit+3b7+VZRfcA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/compat-data": "^7.28.6",
+        "@babel/helper-validator-option": "^7.27.1",
+        "browserslist": "^4.24.0",
+        "lru-cache": "^5.1.1",
+        "semver": "^6.3.1"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/helper-globals": {
+      "version": "7.28.0",
+      "resolved": "https://registry.npmjs.org/@babel/helper-globals/-/helper-globals-7.28.0.tgz",
+      "integrity": "sha512-+W6cISkXFa1jXsDEdYA8HeevQT/FULhxzR99pxphltZcVaugps53THCeiWA8SguxxpSp3gKPiuYfSWopkLQ4hw==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/helper-module-imports": {
+      "version": "7.28.6",
+      "resolved": "https://registry.npmjs.org/@babel/helper-module-imports/-/helper-module-imports-7.28.6.tgz",
+      "integrity": "sha512-l5XkZK7r7wa9LucGw9LwZyyCUscb4x37JWTPz7swwFE/0FMQAGpiWUZn8u9DzkSBWEcK25jmvubfpw2dnAMdbw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/traverse": "^7.28.6",
+        "@babel/types": "^7.28.6"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/helper-module-transforms": {
+      "version": "7.28.6",
+      "resolved": "https://registry.npmjs.org/@babel/helper-module-transforms/-/helper-module-transforms-7.28.6.tgz",
+      "integrity": "sha512-67oXFAYr2cDLDVGLXTEABjdBJZ6drElUSI7WKp70NrpyISso3plG9SAGEF6y7zbha/wOzUByWWTJvEDVNIUGcA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/helper-module-imports": "^7.28.6",
+        "@babel/helper-validator-identifier": "^7.28.5",
+        "@babel/traverse": "^7.28.6"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      },
+      "peerDependencies": {
+        "@babel/core": "^7.0.0"
+      }
+    },
+    "node_modules/@babel/helper-plugin-utils": {
+      "version": "7.28.6",
+      "resolved": "https://registry.npmjs.org/@babel/helper-plugin-utils/-/helper-plugin-utils-7.28.6.tgz",
+      "integrity": "sha512-S9gzZ/bz83GRysI7gAD4wPT/AI3uCnY+9xn+Mx/KPs2JwHJIz1W8PZkg2cqyt3RNOBM8ejcXhV6y8Og7ly/Dug==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/helper-string-parser": {
+      "version": "7.27.1",
+      "resolved": "https://registry.npmjs.org/@babel/helper-string-parser/-/helper-string-parser-7.27.1.tgz",
+      "integrity": "sha512-qMlSxKbpRlAridDExk92nSobyDdpPijUq2DW6oDnUqd0iOGxmQjyqhMIihI9+zv4LPyZdRje2cavWPbCbWm3eA==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/helper-validator-identifier": {
+      "version": "7.28.5",
+      "resolved": "https://registry.npmjs.org/@babel/helper-validator-identifier/-/helper-validator-identifier-7.28.5.tgz",
+      "integrity": "sha512-qSs4ifwzKJSV39ucNjsvc6WVHs6b7S03sOh2OcHF9UHfVPqWWALUsNUVzhSBiItjRZoLHx7nIarVjqKVusUZ1Q==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/helper-validator-option": {
+      "version": "7.27.1",
+      "resolved": "https://registry.npmjs.org/@babel/helper-validator-option/-/helper-validator-option-7.27.1.tgz",
+      "integrity": "sha512-YvjJow9FxbhFFKDSuFnVCe2WxXk1zWc22fFePVNEaWJEu8IrZVlda6N0uHwzZrUM1il7NC9Mlp4MaJYbYd9JSg==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/helpers": {
+      "version": "7.29.2",
+      "resolved": "https://registry.npmjs.org/@babel/helpers/-/helpers-7.29.2.tgz",
+      "integrity": "sha512-HoGuUs4sCZNezVEKdVcwqmZN8GoHirLUcLaYVNBK2J0DadGtdcqgr3BCbvH8+XUo4NGjNl3VOtSjEKNzqfFgKw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/template": "^7.28.6",
+        "@babel/types": "^7.29.0"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/parser": {
+      "version": "7.29.2",
+      "resolved": "https://registry.npmjs.org/@babel/parser/-/parser-7.29.2.tgz",
+      "integrity": "sha512-4GgRzy/+fsBa72/RZVJmGKPmZu9Byn8o4MoLpmNe1m8ZfYnz5emHLQz3U4gLud6Zwl0RZIcgiLD7Uq7ySFuDLA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/types": "^7.29.0"
+      },
+      "bin": {
+        "parser": "bin/babel-parser.js"
+      },
+      "engines": {
+        "node": ">=6.0.0"
+      }
+    },
+    "node_modules/@babel/plugin-transform-react-jsx-self": {
+      "version": "7.27.1",
+      "resolved": "https://registry.npmjs.org/@babel/plugin-transform-react-jsx-self/-/plugin-transform-react-jsx-self-7.27.1.tgz",
+      "integrity": "sha512-6UzkCs+ejGdZ5mFFC/OCUrv028ab2fp1znZmCZjAOBKiBK2jXD1O+BPSfX8X2qjJ75fZBMSnQn3Rq2mrBJK2mw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/helper-plugin-utils": "^7.27.1"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      },
+      "peerDependencies": {
+        "@babel/core": "^7.0.0-0"
+      }
+    },
+    "node_modules/@babel/plugin-transform-react-jsx-source": {
+      "version": "7.27.1",
+      "resolved": "https://registry.npmjs.org/@babel/plugin-transform-react-jsx-source/-/plugin-transform-react-jsx-source-7.27.1.tgz",
+      "integrity": "sha512-zbwoTsBruTeKB9hSq73ha66iFeJHuaFkUbwvqElnygoNbj/jHRsSeokowZFN3CZ64IvEqcmmkVe89OPXc7ldAw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/helper-plugin-utils": "^7.27.1"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      },
+      "peerDependencies": {
+        "@babel/core": "^7.0.0-0"
+      }
+    },
+    "node_modules/@babel/runtime": {
+      "version": "7.29.2",
+      "resolved": "https://registry.npmjs.org/@babel/runtime/-/runtime-7.29.2.tgz",
+      "integrity": "sha512-JiDShH45zKHWyGe4ZNVRrCjBz8Nh9TMmZG1kh4QTK8hCBTWBi8Da+i7s1fJw7/lYpM4ccepSNfqzZ/QvABBi5g==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/template": {
+      "version": "7.28.6",
+      "resolved": "https://registry.npmjs.org/@babel/template/-/template-7.28.6.tgz",
+      "integrity": "sha512-YA6Ma2KsCdGb+WC6UpBVFJGXL58MDA6oyONbjyF/+5sBgxY/dwkhLogbMT2GXXyU84/IhRw/2D1Os1B/giz+BQ==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/code-frame": "^7.28.6",
+        "@babel/parser": "^7.28.6",
+        "@babel/types": "^7.28.6"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/traverse": {
+      "version": "7.29.0",
+      "resolved": "https://registry.npmjs.org/@babel/traverse/-/traverse-7.29.0.tgz",
+      "integrity": "sha512-4HPiQr0X7+waHfyXPZpWPfWL/J7dcN1mx9gL6WdQVMbPnF3+ZhSMs8tCxN7oHddJE9fhNE7+lxdnlyemKfJRuA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/code-frame": "^7.29.0",
+        "@babel/generator": "^7.29.0",
+        "@babel/helper-globals": "^7.28.0",
+        "@babel/parser": "^7.29.0",
+        "@babel/template": "^7.28.6",
+        "@babel/types": "^7.29.0",
+        "debug": "^4.3.1"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@babel/types": {
+      "version": "7.29.0",
+      "resolved": "https://registry.npmjs.org/@babel/types/-/types-7.29.0.tgz",
+      "integrity": "sha512-LwdZHpScM4Qz8Xw2iKSzS+cfglZzJGvofQICy7W7v4caru4EaAmyUuO6BGrbyQ2mYV11W0U8j5mBhd14dd3B0A==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/helper-string-parser": "^7.27.1",
+        "@babel/helper-validator-identifier": "^7.28.5"
+      },
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/@esbuild/aix-ppc64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.21.5.tgz",
+      "integrity": "sha512-1SDgH6ZSPTlggy1yI6+Dbkiz8xzpHJEVAlF/AM1tHPLsf5STom9rwtjE4hKAF20FfXXNTFqEYXyJNWh1GiZedQ==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "aix"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/android-arm": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-arm/-/android-arm-0.21.5.tgz",
+      "integrity": "sha512-vCPvzSjpPHEi1siZdlvAlsPxXl7WbOVUBBAowWug4rJHb68Ox8KualB+1ocNvT5fjv6wpkX6o/iEpbDrf68zcg==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/android-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-arm64/-/android-arm64-0.21.5.tgz",
+      "integrity": "sha512-c0uX9VAUBQ7dTDCjq+wdyGLowMdtR/GoC2U5IYk/7D1H1JYC0qseD7+11iMP2mRLN9RcCMRcjC4YMclCzGwS/A==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/android-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/android-x64/-/android-x64-0.21.5.tgz",
+      "integrity": "sha512-D7aPRUUNHRBwHxzxRvp856rjUHRFW1SdQATKXH2hqA0kAZb1hKmi02OpYRacl0TxIGz/ZmXWlbZgjwWYaCakTA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/darwin-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/darwin-arm64/-/darwin-arm64-0.21.5.tgz",
+      "integrity": "sha512-DwqXqZyuk5AiWWf3UfLiRDJ5EDd49zg6O9wclZ7kUMv2WRFr4HKjXp/5t8JZ11QbQfUS6/cRCKGwYhtNAY88kQ==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/darwin-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/darwin-x64/-/darwin-x64-0.21.5.tgz",
+      "integrity": "sha512-se/JjF8NlmKVG4kNIuyWMV/22ZaerB+qaSi5MdrXtd6R08kvs2qCN4C09miupktDitvh8jRFflwGFBQcxZRjbw==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/freebsd-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/freebsd-arm64/-/freebsd-arm64-0.21.5.tgz",
+      "integrity": "sha512-5JcRxxRDUJLX8JXp/wcBCy3pENnCgBR9bN6JsY4OmhfUtIHe3ZW0mawA7+RDAcMLrMIZaf03NlQiX9DGyB8h4g==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/freebsd-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/freebsd-x64/-/freebsd-x64-0.21.5.tgz",
+      "integrity": "sha512-J95kNBj1zkbMXtHVH29bBriQygMXqoVQOQYA+ISs0/2l3T9/kj42ow2mpqerRBxDJnmkUDCaQT/dfNXWX/ZZCQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-arm": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-arm/-/linux-arm-0.21.5.tgz",
+      "integrity": "sha512-bPb5AHZtbeNGjCKVZ9UGqGwo8EUu4cLq68E95A53KlxAPRmUyYv2D6F0uUI65XisGOL1hBP5mTronbgo+0bFcA==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-arm64/-/linux-arm64-0.21.5.tgz",
+      "integrity": "sha512-ibKvmyYzKsBeX8d8I7MH/TMfWDXBF3db4qM6sy+7re0YXya+K1cem3on9XgdT2EQGMu4hQyZhan7TeQ8XkGp4Q==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-ia32": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-ia32/-/linux-ia32-0.21.5.tgz",
+      "integrity": "sha512-YvjXDqLRqPDl2dvRODYmmhz4rPeVKYvppfGYKSNGdyZkA01046pLWyRKKI3ax8fbJoK5QbxblURkwK/MWY18Tg==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-loong64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-loong64/-/linux-loong64-0.21.5.tgz",
+      "integrity": "sha512-uHf1BmMG8qEvzdrzAqg2SIG/02+4/DHB6a9Kbya0XDvwDEKCoC8ZRWI5JJvNdUjtciBGFQ5PuBlpEOXQj+JQSg==",
+      "cpu": [
+        "loong64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-mips64el": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-mips64el/-/linux-mips64el-0.21.5.tgz",
+      "integrity": "sha512-IajOmO+KJK23bj52dFSNCMsz1QP1DqM6cwLUv3W1QwyxkyIWecfafnI555fvSGqEKwjMXVLokcV5ygHW5b3Jbg==",
+      "cpu": [
+        "mips64el"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-ppc64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-ppc64/-/linux-ppc64-0.21.5.tgz",
+      "integrity": "sha512-1hHV/Z4OEfMwpLO8rp7CvlhBDnjsC3CttJXIhBi+5Aj5r+MBvy4egg7wCbe//hSsT+RvDAG7s81tAvpL2XAE4w==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-riscv64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-riscv64/-/linux-riscv64-0.21.5.tgz",
+      "integrity": "sha512-2HdXDMd9GMgTGrPWnJzP2ALSokE/0O5HhTUvWIbD3YdjME8JwvSCnNGBnTThKGEB91OZhzrJ4qIIxk/SBmyDDA==",
+      "cpu": [
+        "riscv64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-s390x": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-s390x/-/linux-s390x-0.21.5.tgz",
+      "integrity": "sha512-zus5sxzqBJD3eXxwvjN1yQkRepANgxE9lgOW2qLnmr8ikMTphkjgXu1HR01K4FJg8h1kEEDAqDcZQtbrRnB41A==",
+      "cpu": [
+        "s390x"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/linux-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/linux-x64/-/linux-x64-0.21.5.tgz",
+      "integrity": "sha512-1rYdTpyv03iycF1+BhzrzQJCdOuAOtaqHTWJZCWvijKD2N5Xu0TtVC8/+1faWqcP9iBCWOmjmhoH94dH82BxPQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/netbsd-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/netbsd-x64/-/netbsd-x64-0.21.5.tgz",
+      "integrity": "sha512-Woi2MXzXjMULccIwMnLciyZH4nCIMpWQAs049KEeMvOcNADVxo0UBIQPfSmxB3CWKedngg7sWZdLvLczpe0tLg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "netbsd"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/openbsd-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/openbsd-x64/-/openbsd-x64-0.21.5.tgz",
+      "integrity": "sha512-HLNNw99xsvx12lFBUwoT8EVCsSvRNDVxNpjZ7bPn947b8gJPzeHWyNVhFsaerc0n3TsbOINvRP2byTZ5LKezow==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openbsd"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/sunos-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/sunos-x64/-/sunos-x64-0.21.5.tgz",
+      "integrity": "sha512-6+gjmFpfy0BHU5Tpptkuh8+uw3mnrvgs+dSPQXQOv3ekbordwnzTVEb4qnIvQcYXq6gzkyTnoZ9dZG+D4garKg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "sunos"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/win32-arm64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-arm64/-/win32-arm64-0.21.5.tgz",
+      "integrity": "sha512-Z0gOTd75VvXqyq7nsl93zwahcTROgqvuAcYDUr+vOv8uHhNSKROyU961kgtCD1e95IqPKSQKH7tBTslnS3tA8A==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/win32-ia32": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-ia32/-/win32-ia32-0.21.5.tgz",
+      "integrity": "sha512-SWXFF1CL2RVNMaVs+BBClwtfZSvDgtL//G/smwAc5oVK/UPu2Gu9tIaRgFmYFFKrmg3SyAjSrElf0TiJ1v8fYA==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@esbuild/win32-x64": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/@esbuild/win32-x64/-/win32-x64-0.21.5.tgz",
+      "integrity": "sha512-tQd/1efJuzPC6rCFwEvLtci/xNFcTZknmXs98FYDfGE4wP9ClFV98nyKrzJKVPMhdDnjzLhdUyMX4PsQAPjwIw==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=12"
+      }
+    },
+    "node_modules/@jridgewell/gen-mapping": {
+      "version": "0.3.13",
+      "resolved": "https://registry.npmjs.org/@jridgewell/gen-mapping/-/gen-mapping-0.3.13.tgz",
+      "integrity": "sha512-2kkt/7niJ6MgEPxF0bYdQ6etZaA+fQvDcLKckhy1yIQOzaoKjBBjSj63/aLVjYE3qhRt5dvM+uUyfCg6UKCBbA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@jridgewell/sourcemap-codec": "^1.5.0",
+        "@jridgewell/trace-mapping": "^0.3.24"
+      }
+    },
+    "node_modules/@jridgewell/remapping": {
+      "version": "2.3.5",
+      "resolved": "https://registry.npmjs.org/@jridgewell/remapping/-/remapping-2.3.5.tgz",
+      "integrity": "sha512-LI9u/+laYG4Ds1TDKSJW2YPrIlcVYOwi2fUC6xB43lueCjgxV4lffOCZCtYFiH6TNOX+tQKXx97T4IKHbhyHEQ==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@jridgewell/gen-mapping": "^0.3.5",
+        "@jridgewell/trace-mapping": "^0.3.24"
+      }
+    },
+    "node_modules/@jridgewell/resolve-uri": {
+      "version": "3.1.2",
+      "resolved": "https://registry.npmjs.org/@jridgewell/resolve-uri/-/resolve-uri-3.1.2.tgz",
+      "integrity": "sha512-bRISgCIjP20/tbWSPWMEi54QVPRZExkuD9lJL+UIxUKtwVJA8wW1Trb1jMs1RFXo1CBTNZ/5hpC9QvmKWdopKw==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=6.0.0"
+      }
+    },
+    "node_modules/@jridgewell/sourcemap-codec": {
+      "version": "1.5.5",
+      "resolved": "https://registry.npmjs.org/@jridgewell/sourcemap-codec/-/sourcemap-codec-1.5.5.tgz",
+      "integrity": "sha512-cYQ9310grqxueWbl+WuIUIaiUaDcj7WOq5fVhEljNVgRfOUhY9fy2zTvfoqWsnebh8Sl70VScFbICvJnLKB0Og==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/@jridgewell/trace-mapping": {
+      "version": "0.3.31",
+      "resolved": "https://registry.npmjs.org/@jridgewell/trace-mapping/-/trace-mapping-0.3.31.tgz",
+      "integrity": "sha512-zzNR+SdQSDJzc8joaeP8QQoCQr8NuYx2dIIytl1QeBEZHJ9uW6hebsrYgbz8hJwUQao3TWCMtmfV8Nu1twOLAw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@jridgewell/resolve-uri": "^3.1.0",
+        "@jridgewell/sourcemap-codec": "^1.4.14"
+      }
+    },
+    "node_modules/@remix-run/router": {
+      "version": "1.23.2",
+      "resolved": "https://registry.npmjs.org/@remix-run/router/-/router-1.23.2.tgz",
+      "integrity": "sha512-Ic6m2U/rMjTkhERIa/0ZtXJP17QUi2CbWE7cqx4J58M8aA3QTfW+2UlQ4psvTX9IO1RfNVhK3pcpdjej7L+t2w==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=14.0.0"
+      }
+    },
+    "node_modules/@rolldown/pluginutils": {
+      "version": "1.0.0-beta.27",
+      "resolved": "https://registry.npmjs.org/@rolldown/pluginutils/-/pluginutils-1.0.0-beta.27.tgz",
+      "integrity": "sha512-+d0F4MKMCbeVUJwG96uQ4SgAznZNSq93I3V+9NHA4OpvqG8mRCpGdKmK8l/dl02h2CCDHwW2FqilnTyDcAnqjA==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/@rollup/rollup-android-arm-eabi": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm-eabi/-/rollup-android-arm-eabi-4.60.2.tgz",
+      "integrity": "sha512-dnlp69efPPg6Uaw2dVqzWRfAWRnYVb1XJ8CyyhIbZeaq4CA5/mLeZ1IEt9QqQxmbdvagjLIm2ZL8BxXv5lH4Yw==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ]
+    },
+    "node_modules/@rollup/rollup-android-arm64": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm64/-/rollup-android-arm64-4.60.2.tgz",
+      "integrity": "sha512-OqZTwDRDchGRHHm/hwLOL7uVPB9aUvI0am/eQuWMNyFHf5PSEQmyEeYYheA0EPPKUO/l0uigCp+iaTjoLjVoHg==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "android"
+      ]
+    },
+    "node_modules/@rollup/rollup-darwin-arm64": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-arm64/-/rollup-darwin-arm64-4.60.2.tgz",
+      "integrity": "sha512-UwRE7CGpvSVEQS8gUMBe1uADWjNnVgP3Iusyda1nSRwNDCsRjnGc7w6El6WLQsXmZTbLZx9cecegumcitNfpmA==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ]
+    },
+    "node_modules/@rollup/rollup-darwin-x64": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-darwin-x64/-/rollup-darwin-x64-4.60.2.tgz",
+      "integrity": "sha512-gjEtURKLCC5VXm1I+2i1u9OhxFsKAQJKTVB8WvDAHF+oZlq0GTVFOlTlO1q3AlCTE/DF32c16ESvfgqR7343/g==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ]
+    },
+    "node_modules/@rollup/rollup-freebsd-arm64": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-arm64/-/rollup-freebsd-arm64-4.60.2.tgz",
+      "integrity": "sha512-Bcl6CYDeAgE70cqZaMojOi/eK63h5Me97ZqAQoh77VPjMysA/4ORQBRGo3rRy45x4MzVlU9uZxs8Uwy7ZaKnBw==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ]
+    },
+    "node_modules/@rollup/rollup-freebsd-x64": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-freebsd-x64/-/rollup-freebsd-x64-4.60.2.tgz",
+      "integrity": "sha512-LU+TPda3mAE2QB0/Hp5VyeKJivpC6+tlOXd1VMoXV/YFMvk/MNk5iXeBfB4MQGRWyOYVJ01625vjkr0Az98OJQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "freebsd"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-arm-gnueabihf": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-gnueabihf/-/rollup-linux-arm-gnueabihf-4.60.2.tgz",
+      "integrity": "sha512-2QxQrM+KQ7DAW4o22j+XZ6RKdxjLD7BOWTP0Bv0tmjdyhXSsr2Ul1oJDQqh9Zf5qOwTuTc7Ek83mOFaKnodPjg==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-arm-musleabihf": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm-musleabihf/-/rollup-linux-arm-musleabihf-4.60.2.tgz",
+      "integrity": "sha512-TbziEu2DVsTEOPif2mKWkMeDMLoYjx95oESa9fkQQK7r/Orta0gnkcDpzwufEcAO2BLBsD7mZkXGFqEdMRRwfw==",
+      "cpu": [
+        "arm"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-arm64-gnu": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-gnu/-/rollup-linux-arm64-gnu-4.60.2.tgz",
+      "integrity": "sha512-bO/rVDiDUuM2YfuCUwZ1t1cP+/yqjqz+Xf2VtkdppefuOFS2OSeAfgafaHNkFn0t02hEyXngZkxtGqXcXwO8Rg==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-arm64-musl": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-arm64-musl/-/rollup-linux-arm64-musl-4.60.2.tgz",
+      "integrity": "sha512-hr26p7e93Rl0Za+JwW7EAnwAvKkehh12BU1Llm9Ykiibg4uIr2rbpxG9WCf56GuvidlTG9KiiQT/TXT1yAWxTA==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-loong64-gnu": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-loong64-gnu/-/rollup-linux-loong64-gnu-4.60.2.tgz",
+      "integrity": "sha512-pOjB/uSIyDt+ow3k/RcLvUAOGpysT2phDn7TTUB3n75SlIgZzM6NKAqlErPhoFU+npgY3/n+2HYIQVbF70P9/A==",
+      "cpu": [
+        "loong64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-loong64-musl": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-loong64-musl/-/rollup-linux-loong64-musl-4.60.2.tgz",
+      "integrity": "sha512-2/w+q8jszv9Ww1c+6uJT3OwqhdmGP2/4T17cu8WuwyUuuaCDDJ2ojdyYwZzCxx0GcsZBhzi3HmH+J5pZNXnd+Q==",
+      "cpu": [
+        "loong64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-ppc64-gnu": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-ppc64-gnu/-/rollup-linux-ppc64-gnu-4.60.2.tgz",
+      "integrity": "sha512-11+aL5vKheYgczxtPVVRhdptAM2H7fcDR5Gw4/bTcteuZBlH4oP9f5s9zYO9aGZvoGeBpqXI/9TZZihZ609wKw==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-ppc64-musl": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-ppc64-musl/-/rollup-linux-ppc64-musl-4.60.2.tgz",
+      "integrity": "sha512-i16fokAGK46IVZuV8LIIwMdtqhin9hfYkCh8pf8iC3QU3LpwL+1FSFGej+O7l3E/AoknL6Dclh2oTdnRMpTzFQ==",
+      "cpu": [
+        "ppc64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-riscv64-gnu": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-riscv64-gnu/-/rollup-linux-riscv64-gnu-4.60.2.tgz",
+      "integrity": "sha512-49FkKS6RGQoriDSK/6E2GkAsAuU5kETFCh7pG4yD/ylj9rKhTmO3elsnmBvRD4PgJPds5W2PkhC82aVwmUcJ7A==",
+      "cpu": [
+        "riscv64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-riscv64-musl": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-riscv64-musl/-/rollup-linux-riscv64-musl-4.60.2.tgz",
+      "integrity": "sha512-mjYNkHPfGpUR00DuM1ZZIgs64Hpf4bWcz9Z41+4Q+pgDx73UwWdAYyf6EG/lRFldmdHHzgrYyge5akFUW0D3mQ==",
+      "cpu": [
+        "riscv64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-s390x-gnu": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-s390x-gnu/-/rollup-linux-s390x-gnu-4.60.2.tgz",
+      "integrity": "sha512-ALyvJz965BQk8E9Al/JDKKDLH2kfKFLTGMlgkAbbYtZuJt9LU8DW3ZoDMCtQpXAltZxwBHevXz5u+gf0yA0YoA==",
+      "cpu": [
+        "s390x"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-x64-gnu": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-x64-gnu/-/rollup-linux-x64-gnu-4.60.2.tgz",
+      "integrity": "sha512-UQjrkIdWrKI626Du8lCQ6MJp/6V1LAo2bOK9OTu4mSn8GGXIkPXk/Vsp4bLHCd9Z9Iz2OTEaokUE90VweJgIYQ==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-linux-x64-musl": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-linux-x64-musl/-/rollup-linux-x64-musl-4.60.2.tgz",
+      "integrity": "sha512-bTsRGj6VlSdn/XD4CGyzMnzaBs9bsRxy79eTqTCBsA8TMIEky7qg48aPkvJvFe1HyzQ5oMZdg7AnVlWQSKLTnw==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "linux"
+      ]
+    },
+    "node_modules/@rollup/rollup-openbsd-x64": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-openbsd-x64/-/rollup-openbsd-x64-4.60.2.tgz",
+      "integrity": "sha512-6d4Z3534xitaA1FcMWP7mQPq5zGwBmGbhphh2DwaA1aNIXUu3KTOfwrWpbwI4/Gr0uANo7NTtaykFyO2hPuFLg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openbsd"
+      ]
+    },
+    "node_modules/@rollup/rollup-openharmony-arm64": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-openharmony-arm64/-/rollup-openharmony-arm64-4.60.2.tgz",
+      "integrity": "sha512-NetAg5iO2uN7eB8zE5qrZ3CSil+7IJt4WDFLcC75Ymywq1VZVD6qJ6EvNLjZ3rEm6gB7XW5JdT60c6MN35Z85Q==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "openharmony"
+      ]
+    },
+    "node_modules/@rollup/rollup-win32-arm64-msvc": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-arm64-msvc/-/rollup-win32-arm64-msvc-4.60.2.tgz",
+      "integrity": "sha512-NCYhOotpgWZ5kdxCZsv6Iudx0wX8980Q/oW4pNFNihpBKsDbEA1zpkfxJGC0yugsUuyDZ7gL37dbzwhR0VI7pQ==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ]
+    },
+    "node_modules/@rollup/rollup-win32-ia32-msvc": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-ia32-msvc/-/rollup-win32-ia32-msvc-4.60.2.tgz",
+      "integrity": "sha512-RXsaOqXxfoUBQoOgvmmijVxJnW2IGB0eoMO7F8FAjaj0UTywUO/luSqimWBJn04WNgUkeNhh7fs7pESXajWmkg==",
+      "cpu": [
+        "ia32"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ]
+    },
+    "node_modules/@rollup/rollup-win32-x64-gnu": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-x64-gnu/-/rollup-win32-x64-gnu-4.60.2.tgz",
+      "integrity": "sha512-qdAzEULD+/hzObedtmV6iBpdL5TIbKVztGiK7O3/KYSf+HIzU257+MX1EXJcyIiDbMAqmbwaufcYPvyRryeZtA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ]
+    },
+    "node_modules/@rollup/rollup-win32-x64-msvc": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/@rollup/rollup-win32-x64-msvc/-/rollup-win32-x64-msvc-4.60.2.tgz",
+      "integrity": "sha512-Nd/SgG27WoA9e+/TdK74KnHz852TLa94ovOYySo/yMPuTmpckK/jIF2jSwS3g7ELSKXK13/cVdmg1Z/DaCWKxA==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "win32"
+      ]
+    },
+    "node_modules/@thoughtspot/visual-embed-sdk": {
+      "version": "1.47.3",
+      "resolved": "https://registry.npmjs.org/@thoughtspot/visual-embed-sdk/-/visual-embed-sdk-1.47.3.tgz",
+      "integrity": "sha512-PJ00KD6hmC3J83Z6iLnFIcpHg9zGFRXn1EP0Vrz/2peR7WSzZbhqLDEtIJs4Kd6Ly5yQuve7XNF4LBj6EtNclA==",
+      "license": "ThoughtSpot Development Tools End User License Agreement",
+      "dependencies": {
+        "classnames": "^2.3.1",
+        "eventemitter3": "^4.0.7",
+        "lodash": "^4.18.1",
+        "mixpanel-browser": "2.47.0",
+        "ts-deepmerge": "^6.0.2",
+        "tslib": "^2.5.3",
+        "use-deep-compare-effect": "^1.8.1",
+        "yaml": "^2.5.1"
+      },
+      "peerDependencies": {
+        "react": "> 16.8.0",
+        "react-dom": "> 16.8.0"
+      }
+    },
+    "node_modules/@types/babel__core": {
+      "version": "7.20.5",
+      "resolved": "https://registry.npmjs.org/@types/babel__core/-/babel__core-7.20.5.tgz",
+      "integrity": "sha512-qoQprZvz5wQFJwMDqeseRXWv3rqMvhgpbXFfVyWhbx9X47POIA6i/+dXefEmZKoAgOaTdaIgNSMqMIU61yRyzA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/parser": "^7.20.7",
+        "@babel/types": "^7.20.7",
+        "@types/babel__generator": "*",
+        "@types/babel__template": "*",
+        "@types/babel__traverse": "*"
+      }
+    },
+    "node_modules/@types/babel__generator": {
+      "version": "7.27.0",
+      "resolved": "https://registry.npmjs.org/@types/babel__generator/-/babel__generator-7.27.0.tgz",
+      "integrity": "sha512-ufFd2Xi92OAVPYsy+P4n7/U7e68fex0+Ee8gSG9KX7eo084CWiQ4sdxktvdl0bOPupXtVJPY19zk6EwWqUQ8lg==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/types": "^7.0.0"
+      }
+    },
+    "node_modules/@types/babel__template": {
+      "version": "7.4.4",
+      "resolved": "https://registry.npmjs.org/@types/babel__template/-/babel__template-7.4.4.tgz",
+      "integrity": "sha512-h/NUaSyG5EyxBIp8YRxo4RMe2/qQgvyowRwVMzhYhBCONbW8PUsg4lkFMrhgZhUe5z3L3MiLDuvyJ/CaPa2A8A==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/parser": "^7.1.0",
+        "@babel/types": "^7.0.0"
+      }
+    },
+    "node_modules/@types/babel__traverse": {
+      "version": "7.28.0",
+      "resolved": "https://registry.npmjs.org/@types/babel__traverse/-/babel__traverse-7.28.0.tgz",
+      "integrity": "sha512-8PvcXf70gTDZBgt9ptxJ8elBeBjcLOAcOtoO/mPJjtji1+CdGbHgm77om1GrsPxsiE+uXIpNSK64UYaIwQXd4Q==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/types": "^7.28.2"
+      }
+    },
+    "node_modules/@types/estree": {
+      "version": "1.0.8",
+      "resolved": "https://registry.npmjs.org/@types/estree/-/estree-1.0.8.tgz",
+      "integrity": "sha512-dWHzHa2WqEXI/O1E9OjrocMTKJl2mSrEolh1Iomrv6U+JuNwaHXsXx9bLu5gG7BUWFIN0skIQJQ/L1rIex4X6w==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/@types/prop-types": {
+      "version": "15.7.15",
+      "resolved": "https://registry.npmjs.org/@types/prop-types/-/prop-types-15.7.15.tgz",
+      "integrity": "sha512-F6bEyamV9jKGAFBEmlQnesRPGOQqS2+Uwi0Em15xenOxHaf2hv6L8YCVn3rPdPJOiJfPiCnLIRyvwVaqMY3MIw==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/@types/react": {
+      "version": "18.3.28",
+      "resolved": "https://registry.npmjs.org/@types/react/-/react-18.3.28.tgz",
+      "integrity": "sha512-z9VXpC7MWrhfWipitjNdgCauoMLRdIILQsAEV+ZesIzBq/oUlxk0m3ApZuMFCXdnS4U7KrI+l3WRUEGQ8K1QKw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@types/prop-types": "*",
+        "csstype": "^3.2.2"
+      }
+    },
+    "node_modules/@types/react-dom": {
+      "version": "18.3.7",
+      "resolved": "https://registry.npmjs.org/@types/react-dom/-/react-dom-18.3.7.tgz",
+      "integrity": "sha512-MEe3UeoENYVFXzoXEWsvcpg6ZvlrFNlOQ7EOsvhI3CfAXwzPfO8Qwuxd40nepsYKqyyVQnTdEfv68q91yLcKrQ==",
+      "dev": true,
+      "license": "MIT",
+      "peerDependencies": {
+        "@types/react": "^18.0.0"
+      }
+    },
+    "node_modules/@vercel/analytics": {
+      "version": "2.0.1",
+      "resolved": "https://registry.npmjs.org/@vercel/analytics/-/analytics-2.0.1.tgz",
+      "integrity": "sha512-MTQG6V9qQrt1tsDeF+2Uoo5aPjqbVPys1xvnIftXSJYG2SrwXRHnqEvVoYID7BTruDz4lCd2Z7rM1BdkUehk2g==",
+      "license": "MIT",
+      "peerDependencies": {
+        "@remix-run/react": "^2",
+        "@sveltejs/kit": "^1 || ^2",
+        "next": ">= 13",
+        "nuxt": ">= 3",
+        "react": "^18 || ^19 || ^19.0.0-rc",
+        "svelte": ">= 4",
+        "vue": "^3",
+        "vue-router": "^4"
+      },
+      "peerDependenciesMeta": {
+        "@remix-run/react": {
+          "optional": true
+        },
+        "@sveltejs/kit": {
+          "optional": true
+        },
+        "next": {
+          "optional": true
+        },
+        "nuxt": {
+          "optional": true
+        },
+        "react": {
+          "optional": true
+        },
+        "svelte": {
+          "optional": true
+        },
+        "vue": {
+          "optional": true
+        },
+        "vue-router": {
+          "optional": true
+        }
+      }
+    },
+    "node_modules/@vitejs/plugin-react": {
+      "version": "4.7.0",
+      "resolved": "https://registry.npmjs.org/@vitejs/plugin-react/-/plugin-react-4.7.0.tgz",
+      "integrity": "sha512-gUu9hwfWvvEDBBmgtAowQCojwZmJ5mcLn3aufeCsitijs3+f2NsrPtlAWIR6OPiqljl96GVCUbLe0HyqIpVaoA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@babel/core": "^7.28.0",
+        "@babel/plugin-transform-react-jsx-self": "^7.27.1",
+        "@babel/plugin-transform-react-jsx-source": "^7.27.1",
+        "@rolldown/pluginutils": "1.0.0-beta.27",
+        "@types/babel__core": "^7.20.5",
+        "react-refresh": "^0.17.0"
+      },
+      "engines": {
+        "node": "^14.18.0 || >=16.0.0"
+      },
+      "peerDependencies": {
+        "vite": "^4.2.0 || ^5.0.0 || ^6.0.0 || ^7.0.0"
+      }
+    },
+    "node_modules/baseline-browser-mapping": {
+      "version": "2.10.24",
+      "resolved": "https://registry.npmjs.org/baseline-browser-mapping/-/baseline-browser-mapping-2.10.24.tgz",
+      "integrity": "sha512-I2NkZOOrj2XuguvWCK6OVh9GavsNjZjK908Rq3mIBK25+GD8vPX5w2WdxVqnQ7xx3SrZJiCiZFu+/Oz50oSYSA==",
+      "dev": true,
+      "license": "Apache-2.0",
+      "bin": {
+        "baseline-browser-mapping": "dist/cli.cjs"
+      },
+      "engines": {
+        "node": ">=6.0.0"
+      }
+    },
+    "node_modules/browserslist": {
+      "version": "4.28.2",
+      "resolved": "https://registry.npmjs.org/browserslist/-/browserslist-4.28.2.tgz",
+      "integrity": "sha512-48xSriZYYg+8qXna9kwqjIVzuQxi+KYWp2+5nCYnYKPTr0LvD89Jqk2Or5ogxz0NUMfIjhh2lIUX/LyX9B4oIg==",
+      "dev": true,
+      "funding": [
+        {
+          "type": "opencollective",
+          "url": "https://opencollective.com/browserslist"
+        },
+        {
+          "type": "tidelift",
+          "url": "https://tidelift.com/funding/github/npm/browserslist"
+        },
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/ai"
+        }
+      ],
+      "license": "MIT",
+      "dependencies": {
+        "baseline-browser-mapping": "^2.10.12",
+        "caniuse-lite": "^1.0.30001782",
+        "electron-to-chromium": "^1.5.328",
+        "node-releases": "^2.0.36",
+        "update-browserslist-db": "^1.2.3"
+      },
+      "bin": {
+        "browserslist": "cli.js"
+      },
+      "engines": {
+        "node": "^6 || ^7 || ^8 || ^9 || ^10 || ^11 || ^12 || >=13.7"
+      }
+    },
+    "node_modules/caniuse-lite": {
+      "version": "1.0.30001791",
+      "resolved": "https://registry.npmjs.org/caniuse-lite/-/caniuse-lite-1.0.30001791.tgz",
+      "integrity": "sha512-yk0l/YSrOnFZk3UROpDLQD9+kC1l4meK/wed583AXrzoarMGJcbRi2Q4RaUYbKxYAsZ8sWmaSa/DsLmdBeI1vQ==",
+      "dev": true,
+      "funding": [
+        {
+          "type": "opencollective",
+          "url": "https://opencollective.com/browserslist"
+        },
+        {
+          "type": "tidelift",
+          "url": "https://tidelift.com/funding/github/npm/caniuse-lite"
+        },
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/ai"
+        }
+      ],
+      "license": "CC-BY-4.0"
+    },
+    "node_modules/classnames": {
+      "version": "2.5.1",
+      "resolved": "https://registry.npmjs.org/classnames/-/classnames-2.5.1.tgz",
+      "integrity": "sha512-saHYOzhIQs6wy2sVxTM6bUDsQO4F50V9RQ22qBpEdCW+I+/Wmke2HOl6lS6dTpdxVhb88/I6+Hs+438c3lfUow==",
+      "license": "MIT"
+    },
+    "node_modules/convert-source-map": {
+      "version": "2.0.0",
+      "resolved": "https://registry.npmjs.org/convert-source-map/-/convert-source-map-2.0.0.tgz",
+      "integrity": "sha512-Kvp459HrV2FEJ1CAsi1Ku+MY3kasH19TFykTz2xWmMeq6bk2NU3XXvfJ+Q61m0xktWwt+1HSYf3JZsTms3aRJg==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/csstype": {
+      "version": "3.2.3",
+      "resolved": "https://registry.npmjs.org/csstype/-/csstype-3.2.3.tgz",
+      "integrity": "sha512-z1HGKcYy2xA8AGQfwrn0PAy+PB7X/GSj3UVJW9qKyn43xWa+gl5nXmU4qqLMRzWVLFC8KusUX8T/0kCiOYpAIQ==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/debug": {
+      "version": "4.4.3",
+      "resolved": "https://registry.npmjs.org/debug/-/debug-4.4.3.tgz",
+      "integrity": "sha512-RGwwWnwQvkVfavKVt22FGLw+xYSdzARwm0ru6DhTVA3umU5hZc28V3kO4stgYryrTlLpuvgI9GiijltAjNbcqA==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "ms": "^2.1.3"
+      },
+      "engines": {
+        "node": ">=6.0"
+      },
+      "peerDependenciesMeta": {
+        "supports-color": {
+          "optional": true
+        }
+      }
+    },
+    "node_modules/dequal": {
+      "version": "2.0.3",
+      "resolved": "https://registry.npmjs.org/dequal/-/dequal-2.0.3.tgz",
+      "integrity": "sha512-0je+qPKHEMohvfRTCEo3CrPG6cAzAYgmzKyxRiYSSDkS6eGJdyVJm7WaYA5ECaAD9wLB2T4EEeymA5aFVcYXCA==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=6"
+      }
+    },
+    "node_modules/electron-to-chromium": {
+      "version": "1.5.344",
+      "resolved": "https://registry.npmjs.org/electron-to-chromium/-/electron-to-chromium-1.5.344.tgz",
+      "integrity": "sha512-4MxfbmNDm+KPh066EZy+eUnkcDPcZ35wNmOWzFuh/ijvHsve6kbLTLURy88uCNK5FbpN+yk2nQY6BYh1GEt+wg==",
+      "dev": true,
+      "license": "ISC"
+    },
+    "node_modules/esbuild": {
+      "version": "0.21.5",
+      "resolved": "https://registry.npmjs.org/esbuild/-/esbuild-0.21.5.tgz",
+      "integrity": "sha512-mg3OPMV4hXywwpoDxu3Qda5xCKQi+vCTZq8S9J/EpkhB2HzKXq4SNFZE3+NK93JYxc8VMSep+lOUSC/RVKaBqw==",
+      "dev": true,
+      "hasInstallScript": true,
+      "license": "MIT",
+      "bin": {
+        "esbuild": "bin/esbuild"
+      },
+      "engines": {
+        "node": ">=12"
+      },
+      "optionalDependencies": {
+        "@esbuild/aix-ppc64": "0.21.5",
+        "@esbuild/android-arm": "0.21.5",
+        "@esbuild/android-arm64": "0.21.5",
+        "@esbuild/android-x64": "0.21.5",
+        "@esbuild/darwin-arm64": "0.21.5",
+        "@esbuild/darwin-x64": "0.21.5",
+        "@esbuild/freebsd-arm64": "0.21.5",
+        "@esbuild/freebsd-x64": "0.21.5",
+        "@esbuild/linux-arm": "0.21.5",
+        "@esbuild/linux-arm64": "0.21.5",
+        "@esbuild/linux-ia32": "0.21.5",
+        "@esbuild/linux-loong64": "0.21.5",
+        "@esbuild/linux-mips64el": "0.21.5",
+        "@esbuild/linux-ppc64": "0.21.5",
+        "@esbuild/linux-riscv64": "0.21.5",
+        "@esbuild/linux-s390x": "0.21.5",
+        "@esbuild/linux-x64": "0.21.5",
+        "@esbuild/netbsd-x64": "0.21.5",
+        "@esbuild/openbsd-x64": "0.21.5",
+        "@esbuild/sunos-x64": "0.21.5",
+        "@esbuild/win32-arm64": "0.21.5",
+        "@esbuild/win32-ia32": "0.21.5",
+        "@esbuild/win32-x64": "0.21.5"
+      }
+    },
+    "node_modules/escalade": {
+      "version": "3.2.0",
+      "resolved": "https://registry.npmjs.org/escalade/-/escalade-3.2.0.tgz",
+      "integrity": "sha512-WUj2qlxaQtO4g6Pq5c29GTcWGDyd8itL8zTlipgECz3JesAiiOKotd8JU6otB3PACgG6xkJUyVhboMS+bje/jA==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=6"
+      }
+    },
+    "node_modules/eventemitter3": {
+      "version": "4.0.7",
+      "resolved": "https://registry.npmjs.org/eventemitter3/-/eventemitter3-4.0.7.tgz",
+      "integrity": "sha512-8guHBZCwKnFhYdHr2ysuRWErTwhoN2X8XELRlrRwpmfeY2jjuUN4taQMsULKUVo1K4DvZl+0pgfyoysHxvmvEw==",
+      "license": "MIT"
+    },
+    "node_modules/fsevents": {
+      "version": "2.3.3",
+      "resolved": "https://registry.npmjs.org/fsevents/-/fsevents-2.3.3.tgz",
+      "integrity": "sha512-5xoDfX+fL7faATnagmWPpbFtwh/R77WmMMqqHGS65C3vvB0YHrgF+B1YmZ3441tMj5n63k0212XNoJwzlhffQw==",
+      "dev": true,
+      "hasInstallScript": true,
+      "license": "MIT",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": "^8.16.0 || ^10.6.0 || >=11.0.0"
+      }
+    },
+    "node_modules/gensync": {
+      "version": "1.0.0-beta.2",
+      "resolved": "https://registry.npmjs.org/gensync/-/gensync-1.0.0-beta.2.tgz",
+      "integrity": "sha512-3hN7NaskYvMDLQY55gnW3NQ+mesEAepTqlg+VEbj7zzqEMBVNhzcGYYeqFo/TlYz6eQiFcp1HcsCZO+nGgS8zg==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=6.9.0"
+      }
+    },
+    "node_modules/js-tokens": {
+      "version": "4.0.0",
+      "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-4.0.0.tgz",
+      "integrity": "sha512-RdJUflcE3cUzKiMqQgsCu06FPu9UdIJO0beYbPhHN4k6apgJtifcoCtT9bcxOpYBtpD2kCM6Sbzg4CausW/PKQ==",
+      "license": "MIT"
+    },
+    "node_modules/jsesc": {
+      "version": "3.1.0",
+      "resolved": "https://registry.npmjs.org/jsesc/-/jsesc-3.1.0.tgz",
+      "integrity": "sha512-/sM3dO2FOzXjKQhJuo0Q173wf2KOo8t4I8vHy6lF9poUp7bKT0/NHE8fPX23PwfhnykfqnC2xRxOnVw5XuGIaA==",
+      "dev": true,
+      "license": "MIT",
+      "bin": {
+        "jsesc": "bin/jsesc"
+      },
+      "engines": {
+        "node": ">=6"
+      }
+    },
+    "node_modules/json5": {
+      "version": "2.2.3",
+      "resolved": "https://registry.npmjs.org/json5/-/json5-2.2.3.tgz",
+      "integrity": "sha512-XmOWe7eyHYH14cLdVPoyg+GOH3rYX++KpzrylJwSW98t3Nk+U8XOl8FWKOgwtzdb8lXGf6zYwDUzeHMWfxasyg==",
+      "dev": true,
+      "license": "MIT",
+      "bin": {
+        "json5": "lib/cli.js"
+      },
+      "engines": {
+        "node": ">=6"
+      }
+    },
+    "node_modules/lodash": {
+      "version": "4.18.1",
+      "resolved": "https://registry.npmjs.org/lodash/-/lodash-4.18.1.tgz",
+      "integrity": "sha512-dMInicTPVE8d1e5otfwmmjlxkZoUpiVLwyeTdUsi/Caj/gfzzblBcCE5sRHV/AsjuCmxWrte2TNGSYuCeCq+0Q==",
+      "license": "MIT"
+    },
+    "node_modules/loose-envify": {
+      "version": "1.4.0",
+      "resolved": "https://registry.npmjs.org/loose-envify/-/loose-envify-1.4.0.tgz",
+      "integrity": "sha512-lyuxPGr/Wfhrlem2CL/UcnUc1zcqKAImBDzukY7Y5F/yQiNdko6+fRLevlw1HgMySw7f611UIY408EtxRSoK3Q==",
+      "license": "MIT",
+      "dependencies": {
+        "js-tokens": "^3.0.0 || ^4.0.0"
+      },
+      "bin": {
+        "loose-envify": "cli.js"
+      }
+    },
+    "node_modules/lru-cache": {
+      "version": "5.1.1",
+      "resolved": "https://registry.npmjs.org/lru-cache/-/lru-cache-5.1.1.tgz",
+      "integrity": "sha512-KpNARQA3Iwv+jTA0utUVVbrh+Jlrr1Fv0e56GGzAFOXN7dk/FviaDW8LHmK52DlcH4WP2n6gI8vN1aesBFgo9w==",
+      "dev": true,
+      "license": "ISC",
+      "dependencies": {
+        "yallist": "^3.0.2"
+      }
+    },
+    "node_modules/lucide-react": {
+      "version": "0.300.0",
+      "resolved": "https://registry.npmjs.org/lucide-react/-/lucide-react-0.300.0.tgz",
+      "integrity": "sha512-rQxUUCmWAvNLoAsMZ5j04b2+OJv6UuNLYMY7VK0eVlm4aTwUEjEEHc09/DipkNIlhXUSDn2xoyIzVT0uh7dRsg==",
+      "license": "ISC",
+      "peerDependencies": {
+        "react": "^16.5.1 || ^17.0.0 || ^18.0.0"
+      }
+    },
+    "node_modules/mixpanel-browser": {
+      "version": "2.47.0",
+      "resolved": "https://registry.npmjs.org/mixpanel-browser/-/mixpanel-browser-2.47.0.tgz",
+      "integrity": "sha512-Ldrva0fRBEIFWmEibBQO1PulfpJVF3pf28Guk09lDirDaSQqqU/xs9zQLwN2rL5VwVtsP1aD3JaCgaa98EjojQ==",
+      "license": "Apache-2.0"
+    },
+    "node_modules/ms": {
+      "version": "2.1.3",
+      "resolved": "https://registry.npmjs.org/ms/-/ms-2.1.3.tgz",
+      "integrity": "sha512-6FlzubTLZG3J2a/NVCAleEhjzq5oxgHyaCU9yYXvcLsvoVaHJq/s5xXI6/XXP6tz7R9xAOtHnSO/tXtF3WRTlA==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/nanoid": {
+      "version": "3.3.11",
+      "resolved": "https://registry.npmjs.org/nanoid/-/nanoid-3.3.11.tgz",
+      "integrity": "sha512-N8SpfPUnUp1bK+PMYW8qSWdl9U+wwNWI4QKxOYDy9JAro3WMX7p2OeVRF9v+347pnakNevPmiHhNmZ2HbFA76w==",
+      "dev": true,
+      "funding": [
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/ai"
+        }
+      ],
+      "license": "MIT",
+      "bin": {
+        "nanoid": "bin/nanoid.cjs"
+      },
+      "engines": {
+        "node": "^10 || ^12 || ^13.7 || ^14 || >=15.0.1"
+      }
+    },
+    "node_modules/node-releases": {
+      "version": "2.0.38",
+      "resolved": "https://registry.npmjs.org/node-releases/-/node-releases-2.0.38.tgz",
+      "integrity": "sha512-3qT/88Y3FbH/Kx4szpQQ4HzUbVrHPKTLVpVocKiLfoYvw9XSGOX2FmD2d6DrXbVYyAQTF2HeF6My8jmzx7/CRw==",
+      "dev": true,
+      "license": "MIT"
+    },
+    "node_modules/picocolors": {
+      "version": "1.1.1",
+      "resolved": "https://registry.npmjs.org/picocolors/-/picocolors-1.1.1.tgz",
+      "integrity": "sha512-xceH2snhtb5M9liqDsmEw56le376mTZkEX/jEb/RxNFyegNul7eNslCXP9FDj/Lcu0X8KEyMceP2ntpaHrDEVA==",
+      "dev": true,
+      "license": "ISC"
+    },
+    "node_modules/postcss": {
+      "version": "8.5.12",
+      "resolved": "https://registry.npmjs.org/postcss/-/postcss-8.5.12.tgz",
+      "integrity": "sha512-W62t/Se6rA0Az3DfCL0AqJwXuKwBeYg6nOaIgzP+xZ7N5BFCI7DYi1qs6ygUYT6rvfi6t9k65UMLJC+PHZpDAA==",
+      "dev": true,
+      "funding": [
+        {
+          "type": "opencollective",
+          "url": "https://opencollective.com/postcss/"
+        },
+        {
+          "type": "tidelift",
+          "url": "https://tidelift.com/funding/github/npm/postcss"
+        },
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/ai"
+        }
+      ],
+      "license": "MIT",
+      "dependencies": {
+        "nanoid": "^3.3.11",
+        "picocolors": "^1.1.1",
+        "source-map-js": "^1.2.1"
+      },
+      "engines": {
+        "node": "^10 || ^12 || >=14"
+      }
+    },
+    "node_modules/react": {
+      "version": "18.3.1",
+      "resolved": "https://registry.npmjs.org/react/-/react-18.3.1.tgz",
+      "integrity": "sha512-wS+hAgJShR0KhEvPJArfuPVN1+Hz1t0Y6n5jLrGQbkb4urgPE/0Rve+1kMB1v/oWgHgm4WIcV+i7F2pTVj+2iQ==",
+      "license": "MIT",
+      "dependencies": {
+        "loose-envify": "^1.1.0"
+      },
+      "engines": {
+        "node": ">=0.10.0"
+      }
+    },
+    "node_modules/react-dom": {
+      "version": "18.3.1",
+      "resolved": "https://registry.npmjs.org/react-dom/-/react-dom-18.3.1.tgz",
+      "integrity": "sha512-5m4nQKp+rZRb09LNH59GM4BxTh9251/ylbKIbpe7TpGxfJ+9kv6BLkLBXIjjspbgbnIBNqlI23tRnTWT0snUIw==",
+      "license": "MIT",
+      "dependencies": {
+        "loose-envify": "^1.1.0",
+        "scheduler": "^0.23.2"
+      },
+      "peerDependencies": {
+        "react": "^18.3.1"
+      }
+    },
+    "node_modules/react-refresh": {
+      "version": "0.17.0",
+      "resolved": "https://registry.npmjs.org/react-refresh/-/react-refresh-0.17.0.tgz",
+      "integrity": "sha512-z6F7K9bV85EfseRCp2bzrpyQ0Gkw1uLoCel9XBVWPg/TjRj94SkJzUTGfOa4bs7iJvBWtQG0Wq7wnI0syw3EBQ==",
+      "dev": true,
+      "license": "MIT",
+      "engines": {
+        "node": ">=0.10.0"
+      }
+    },
+    "node_modules/react-router": {
+      "version": "6.30.3",
+      "resolved": "https://registry.npmjs.org/react-router/-/react-router-6.30.3.tgz",
+      "integrity": "sha512-XRnlbKMTmktBkjCLE8/XcZFlnHvr2Ltdr1eJX4idL55/9BbORzyZEaIkBFDhFGCEWBBItsVrDxwx3gnisMitdw==",
+      "license": "MIT",
+      "dependencies": {
+        "@remix-run/router": "1.23.2"
+      },
+      "engines": {
+        "node": ">=14.0.0"
+      },
+      "peerDependencies": {
+        "react": ">=16.8"
+      }
+    },
+    "node_modules/react-router-dom": {
+      "version": "6.30.3",
+      "resolved": "https://registry.npmjs.org/react-router-dom/-/react-router-dom-6.30.3.tgz",
+      "integrity": "sha512-pxPcv1AczD4vso7G4Z3TKcvlxK7g7TNt3/FNGMhfqyntocvYKj+GCatfigGDjbLozC4baguJ0ReCigoDJXb0ag==",
+      "license": "MIT",
+      "dependencies": {
+        "@remix-run/router": "1.23.2",
+        "react-router": "6.30.3"
+      },
+      "engines": {
+        "node": ">=14.0.0"
+      },
+      "peerDependencies": {
+        "react": ">=16.8",
+        "react-dom": ">=16.8"
+      }
+    },
+    "node_modules/rollup": {
+      "version": "4.60.2",
+      "resolved": "https://registry.npmjs.org/rollup/-/rollup-4.60.2.tgz",
+      "integrity": "sha512-J9qZyW++QK/09NyN/zeO0dG/1GdGfyp9lV8ajHnRVLfo/uFsbji5mHnDgn/qYdUHyCkM2N+8VyspgZclfAh0eQ==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "@types/estree": "1.0.8"
+      },
+      "bin": {
+        "rollup": "dist/bin/rollup"
+      },
+      "engines": {
+        "node": ">=18.0.0",
+        "npm": ">=8.0.0"
+      },
+      "optionalDependencies": {
+        "@rollup/rollup-android-arm-eabi": "4.60.2",
+        "@rollup/rollup-android-arm64": "4.60.2",
+        "@rollup/rollup-darwin-arm64": "4.60.2",
+        "@rollup/rollup-darwin-x64": "4.60.2",
+        "@rollup/rollup-freebsd-arm64": "4.60.2",
+        "@rollup/rollup-freebsd-x64": "4.60.2",
+        "@rollup/rollup-linux-arm-gnueabihf": "4.60.2",
+        "@rollup/rollup-linux-arm-musleabihf": "4.60.2",
+        "@rollup/rollup-linux-arm64-gnu": "4.60.2",
+        "@rollup/rollup-linux-arm64-musl": "4.60.2",
+        "@rollup/rollup-linux-loong64-gnu": "4.60.2",
+        "@rollup/rollup-linux-loong64-musl": "4.60.2",
+        "@rollup/rollup-linux-ppc64-gnu": "4.60.2",
+        "@rollup/rollup-linux-ppc64-musl": "4.60.2",
+        "@rollup/rollup-linux-riscv64-gnu": "4.60.2",
+        "@rollup/rollup-linux-riscv64-musl": "4.60.2",
+        "@rollup/rollup-linux-s390x-gnu": "4.60.2",
+        "@rollup/rollup-linux-x64-gnu": "4.60.2",
+        "@rollup/rollup-linux-x64-musl": "4.60.2",
+        "@rollup/rollup-openbsd-x64": "4.60.2",
+        "@rollup/rollup-openharmony-arm64": "4.60.2",
+        "@rollup/rollup-win32-arm64-msvc": "4.60.2",
+        "@rollup/rollup-win32-ia32-msvc": "4.60.2",
+        "@rollup/rollup-win32-x64-gnu": "4.60.2",
+        "@rollup/rollup-win32-x64-msvc": "4.60.2",
+        "fsevents": "~2.3.2"
+      }
+    },
+    "node_modules/scheduler": {
+      "version": "0.23.2",
+      "resolved": "https://registry.npmjs.org/scheduler/-/scheduler-0.23.2.tgz",
+      "integrity": "sha512-UOShsPwz7NrMUqhR6t0hWjFduvOzbtv7toDH1/hIrfRNIDBnnBWd0CwJTGvTpngVlmwGCdP9/Zl/tVrDqcuYzQ==",
+      "license": "MIT",
+      "dependencies": {
+        "loose-envify": "^1.1.0"
+      }
+    },
+    "node_modules/semver": {
+      "version": "6.3.1",
+      "resolved": "https://registry.npmjs.org/semver/-/semver-6.3.1.tgz",
+      "integrity": "sha512-BR7VvDCVHO+q2xBEWskxS6DJE1qRnb7DxzUrogb71CWoSficBxYsiAGd+Kl0mmq/MprG9yArRkyrQxTO6XjMzA==",
+      "dev": true,
+      "license": "ISC",
+      "bin": {
+        "semver": "bin/semver.js"
+      }
+    },
+    "node_modules/source-map-js": {
+      "version": "1.2.1",
+      "resolved": "https://registry.npmjs.org/source-map-js/-/source-map-js-1.2.1.tgz",
+      "integrity": "sha512-UXWMKhLOwVKb728IUtQPXxfYU+usdybtUrK/8uGE8CQMvrhOpwvzDBwj0QhSL7MQc7vIsISBG8VQ8+IDQxpfQA==",
+      "dev": true,
+      "license": "BSD-3-Clause",
+      "engines": {
+        "node": ">=0.10.0"
+      }
+    },
+    "node_modules/ts-deepmerge": {
+      "version": "6.2.1",
+      "resolved": "https://registry.npmjs.org/ts-deepmerge/-/ts-deepmerge-6.2.1.tgz",
+      "integrity": "sha512-8CYSLazCyj0DJDpPIxOFzJG46r93uh6EynYjuey+bxcLltBeqZL7DMfaE5ZPzZNFlav7wx+2TDa/mBl8gkTYzw==",
+      "license": "ISC",
+      "engines": {
+        "node": ">=14.13.1"
+      }
+    },
+    "node_modules/tslib": {
+      "version": "2.8.1",
+      "resolved": "https://registry.npmjs.org/tslib/-/tslib-2.8.1.tgz",
+      "integrity": "sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w==",
+      "license": "0BSD"
+    },
+    "node_modules/typescript": {
+      "version": "5.9.3",
+      "resolved": "https://registry.npmjs.org/typescript/-/typescript-5.9.3.tgz",
+      "integrity": "sha512-jl1vZzPDinLr9eUt3J/t7V6FgNEw9QjvBPdysz9KfQDD41fQrC2Y4vKQdiaUpFT4bXlb1RHhLpp8wtm6M5TgSw==",
+      "dev": true,
+      "license": "Apache-2.0",
+      "bin": {
+        "tsc": "bin/tsc",
+        "tsserver": "bin/tsserver"
+      },
+      "engines": {
+        "node": ">=14.17"
+      }
+    },
+    "node_modules/update-browserslist-db": {
+      "version": "1.2.3",
+      "resolved": "https://registry.npmjs.org/update-browserslist-db/-/update-browserslist-db-1.2.3.tgz",
+      "integrity": "sha512-Js0m9cx+qOgDxo0eMiFGEueWztz+d4+M3rGlmKPT+T4IS/jP4ylw3Nwpu6cpTTP8R1MAC1kF4VbdLt3ARf209w==",
+      "dev": true,
+      "funding": [
+        {
+          "type": "opencollective",
+          "url": "https://opencollective.com/browserslist"
+        },
+        {
+          "type": "tidelift",
+          "url": "https://tidelift.com/funding/github/npm/browserslist"
+        },
+        {
+          "type": "github",
+          "url": "https://github.com/sponsors/ai"
+        }
+      ],
+      "license": "MIT",
+      "dependencies": {
+        "escalade": "^3.2.0",
+        "picocolors": "^1.1.1"
+      },
+      "bin": {
+        "update-browserslist-db": "cli.js"
+      },
+      "peerDependencies": {
+        "browserslist": ">= 4.21.0"
+      }
+    },
+    "node_modules/use-deep-compare-effect": {
+      "version": "1.8.1",
+      "resolved": "https://registry.npmjs.org/use-deep-compare-effect/-/use-deep-compare-effect-1.8.1.tgz",
+      "integrity": "sha512-kbeNVZ9Zkc0RFGpfMN3MNfaKNvcLNyxOAAd9O4CBZ+kCBXXscn9s/4I+8ytUER4RDpEYs5+O6Rs4PqiZ+rHr5Q==",
+      "license": "MIT",
+      "dependencies": {
+        "@babel/runtime": "^7.12.5",
+        "dequal": "^2.0.2"
+      },
+      "engines": {
+        "node": ">=10",
+        "npm": ">=6"
+      },
+      "peerDependencies": {
+        "react": ">=16.13"
+      }
+    },
+    "node_modules/vite": {
+      "version": "5.4.21",
+      "resolved": "https://registry.npmjs.org/vite/-/vite-5.4.21.tgz",
+      "integrity": "sha512-o5a9xKjbtuhY6Bi5S3+HvbRERmouabWbyUcpXXUA1u+GNUKoROi9byOJ8M0nHbHYHkYICiMlqxkg1KkYmm25Sw==",
+      "dev": true,
+      "license": "MIT",
+      "dependencies": {
+        "esbuild": "^0.21.3",
+        "postcss": "^8.4.43",
+        "rollup": "^4.20.0"
+      },
+      "bin": {
+        "vite": "bin/vite.js"
+      },
+      "engines": {
+        "node": "^18.0.0 || >=20.0.0"
+      },
+      "funding": {
+        "url": "https://github.com/vitejs/vite?sponsor=1"
+      },
+      "optionalDependencies": {
+        "fsevents": "~2.3.3"
+      },
+      "peerDependencies": {
+        "@types/node": "^18.0.0 || >=20.0.0",
+        "less": "*",
+        "lightningcss": "^1.21.0",
+        "sass": "*",
+        "sass-embedded": "*",
+        "stylus": "*",
+        "sugarss": "*",
+        "terser": "^5.4.0"
+      },
+      "peerDependenciesMeta": {
+        "@types/node": {
+          "optional": true
+        },
+        "less": {
+          "optional": true
+        },
+        "lightningcss": {
+          "optional": true
+        },
+        "sass": {
+          "optional": true
+        },
+        "sass-embedded": {
+          "optional": true
+        },
+        "stylus": {
+          "optional": true
+        },
+        "sugarss": {
+          "optional": true
+        },
+        "terser": {
+          "optional": true
+        }
+      }
+    },
+    "node_modules/yallist": {
+      "version": "3.1.1",
+      "resolved": "https://registry.npmjs.org/yallist/-/yallist-3.1.1.tgz",
+      "integrity": "sha512-a4UGQaWPH59mOXUYnAG2ewncQS4i4F43Tv3JoAM+s2VDAmS9NsK8GpDMLrCHPksFT7h3K6TOoUNn2pb7RoXx4g==",
+      "dev": true,
+      "license": "ISC"
+    },
+    "node_modules/yaml": {
+      "version": "2.8.3",
+      "resolved": "https://registry.npmjs.org/yaml/-/yaml-2.8.3.tgz",
+      "integrity": "sha512-AvbaCLOO2Otw/lW5bmh9d/WEdcDFdQp2Z2ZUH3pX9U2ihyUY0nvLv7J6TrWowklRGPYbB/IuIMfYgxaCPg5Bpg==",
+      "license": "ISC",
+      "bin": {
+        "yaml": "bin.mjs"
+      },
+      "engines": {
+        "node": ">= 14.6"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/eemeli"
+      }
+    }
+  }
+}

demoprep_app/portal/template-tse/package.json

@@ -0,0 +1,26 @@
+{
+  "name": "northwind-thoughtspot-portal",
+  "private": true,
+  "version": "1.0.0",
+  "type": "module",
+  "scripts": {
+    "dev": "vite",
+    "build": "tsc && vite build",
+    "preview": "vite preview"
+  },
+  "dependencies": {
+    "@thoughtspot/visual-embed-sdk": "^1.37.0",
+    "@vercel/analytics": "^2.0.1",
+    "lucide-react": "^0.300.0",
+    "react": "^18.2.0",
+    "react-dom": "^18.2.0",
+    "react-router-dom": "^6.21.0"
+  },
+  "devDependencies": {
+    "@types/react": "^18.2.45",
+    "@types/react-dom": "^18.2.18",
+    "@vitejs/plugin-react": "^4.2.1",
+    "typescript": "^5.3.3",
+    "vite": "^5.0.10"
+  }
+}

demoprep_app/portal/template-tse/src/App.tsx

@@ -0,0 +1,60 @@
+import { useState } from 'react';
+import { Analytics as VercelAnalytics } from '@vercel/analytics/react';
+import { useAuth } from './context/AuthContext';
+import Login from './pages/Login';
+import TopBar, { TabId } from './components/TopBar';
+import ChatBot from './components/ChatBot';
+import MyAnalytics from './tabs/MyAnalytics';
+import Analytics from './tabs/Analytics';
+import Carriers from './tabs/Carriers';
+import Capacity from './tabs/Capacity';
+import AskNorthwind from './tabs/AskNorthwind';
+import SpotterTab from './tabs/Spotter';
+import {
+  WORKSHEET_ID,
+  CADENCE_WORKSHEET_ID,
+  CHATBOT_WELCOME,
+  CHATBOT_CADENCES_WELCOME,
+} from './config';
+
+// Persist the active tab so a reload restores it instead of resetting to default.
+const TAB_KEY = 'northwind.tab';
+
+export default function App() {
+  const { isAuthenticated } = useAuth();
+  const [tab, setTab] = useState<TabId>(
+    () => (sessionStorage.getItem(TAB_KEY) as TabId) || 'analytics',
+  );
+
+  const changeTab = (next: TabId) => {
+    setTab(next);
+    try {
+      sessionStorage.setItem(TAB_KEY, next);
+    } catch {
+      /* storage unavailable */
+    }
+  };
+
+  if (!isAuthenticated) {
+    return <Login />;
+  }
+
+  return (
+    <div className="app-shell">
+      <TopBar active={tab} onChange={changeTab} />
+      <main className="app-main">
+        {tab === 'my-analytics' && <MyAnalytics />}
+        {tab === 'analytics' && <Analytics />}
+        {tab === 'carriers' && <Carriers />}
+        {tab === 'capacity' && <Capacity />}
+        {tab === 'ask' && <AskNorthwind />}
+        {tab === 'spotter' && <SpotterTab />}
+      </main>
+      <ChatBot
+        worksheetId={tab === 'carriers' ? CADENCE_WORKSHEET_ID : WORKSHEET_ID}
+        greeting={tab === 'carriers' ? CHATBOT_CADENCES_WELCOME : CHATBOT_WELCOME}
+      />
+      <VercelAnalytics />
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/BidModal.tsx

@@ -0,0 +1,133 @@
+import { useEffect, useState } from 'react';
+import { X, Gavel, Check, Send } from 'lucide-react';
+import { SignalContext } from '../lib/signalContext';
+import { CONTENT } from '../content';
+
+interface Props {
+  open: boolean;
+  context: SignalContext | null;
+  onClose: () => void;
+}
+
+// Pull a value out of the clicked-row fields by matching the column name.
+function pick(ctx: SignalContext | null, re: RegExp): string {
+  const f = ctx?.fields.find((x) => re.test(x.name) && x.value !== '—');
+  return f ? f.value : '';
+}
+
+export default function BidModal({ open, context, onClose }: Props) {
+  const [corridor, setCorridor] = useState('');
+  const [carrier, setCarrier] = useState('');
+  const [equipment, setEquipment] = useState('Dry Van');
+  const [targetRate, setTargetRate] = useState('');
+  const [notes, setNotes] = useState('');
+  const [submitted, setSubmitted] = useState(false);
+
+  // Pre-fill from the row the user right-clicked whenever a new bid is opened.
+  useEffect(() => {
+    if (!context) return;
+    setCorridor(pick(context, /corridor|lane|origin|destination/i) || '');
+    setCarrier(pick(context, /carrier/i) || '');
+    const rate = pick(context, /rate|price|\$|cost|spend/i);
+    setTargetRate(rate ? rate.replace(/[^0-9.]/g, '') : '');
+    setNotes('');
+    setSubmitted(false);
+  }, [context]);
+
+  if (!open || !context) return null;
+
+  // Everything else from the row, shown as read-only context chips.
+  const otherFields = context.fields.filter(
+    (f) => !/corridor|lane|carrier|rate|price|cost/i.test(f.name) && f.value !== '—',
+  );
+
+  return (
+    <div className="winback-overlay" onClick={onClose}>
+      <div className="winback-modal bid-modal" onClick={(e) => e.stopPropagation()}>
+        <div className="winback-header">
+          <div className="winback-title">
+            <Gavel size={18} />
+            <span>{CONTENT.action.modalTitle}</span>
+          </div>
+          <button className="winback-close" onClick={onClose} aria-label="Close">
+            <X size={18} />
+          </button>
+        </div>
+
+        {submitted ? (
+          <div className="winback-success">
+            <div className="winback-success-icon">
+              <Check size={30} />
+            </div>
+            <h3>Bid request submitted</h3>
+            <p>
+              Your bid for{' '}
+              <strong>{corridor || 'this lane'}</strong>
+              {carrier ? <> to <strong>{carrier}</strong></> : null} has been sent to the
+              Northwind capacity desk.
+            </p>
+            <button className="sl-hfilter-apply bid-done" onClick={onClose}>
+              Done
+            </button>
+          </div>
+        ) : (
+          <div className="bid-body">
+            <p className="bid-lead">
+              Pre-filled from the selected lane. Review and submit to request a carrier
+              bid through Northwind.
+            </p>
+
+            <div className="bid-grid">
+              <label className="bid-field">
+                <span>Freight corridor / lane</span>
+                <input value={corridor} onChange={(e) => setCorridor(e.target.value)} placeholder="e.g. Dallas, TX → Atlanta, GA" />
+              </label>
+              <label className="bid-field">
+                <span>Carrier</span>
+                <input value={carrier} onChange={(e) => setCarrier(e.target.value)} placeholder="Carrier name" />
+              </label>
+              <label className="bid-field">
+                <span>Equipment</span>
+                <select value={equipment} onChange={(e) => setEquipment(e.target.value)}>
+                  <option>Dry Van</option>
+                  <option>Reefer</option>
+                  <option>Flatbed</option>
+                  <option>Power Only</option>
+                </select>
+              </label>
+              <label className="bid-field">
+                <span>Target rate ($)</span>
+                <input value={targetRate} onChange={(e) => setTargetRate(e.target.value)} placeholder="2,300.00" inputMode="decimal" />
+              </label>
+            </div>
+
+            {otherFields.length > 0 && (
+              <div className="bid-context">
+                <span className="bid-context-label">From this lane</span>
+                <div className="bid-chips">
+                  {otherFields.slice(0, 6).map((f) => (
+                    <span className="bid-chip" key={f.name}>
+                      <em>{f.name}</em> {f.value}
+                    </span>
+                  ))}
+                </div>
+              </div>
+            )}
+
+            <label className="bid-field">
+              <span>Notes for the capacity desk</span>
+              <textarea rows={2} value={notes} onChange={(e) => setNotes(e.target.value)} placeholder="Any constraints, timing, or commitments…" />
+            </label>
+
+            <div className="bid-actions">
+              <button className="winback-close-text" onClick={onClose}>Cancel</button>
+              <button className="sl-hfilter-apply bid-submit" onClick={() => setSubmitted(true)}>
+                <Send size={15} /> Submit bid request
+              </button>
+            </div>
+          </div>
+        )}
+      </div>
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/ChatBot.tsx

@@ -0,0 +1,186 @@
+import { useEffect, useRef, useState } from 'react';
+import { Wand2, X, ArrowUp } from 'lucide-react';
+import { BodylessConversation } from '@thoughtspot/visual-embed-sdk';
+import { routeMessage, ChatTurn } from '../lib/chatbot';
+
+interface ChatBotProps {
+  /** Data model the chatbot routes analytics questions to (tab-dependent). */
+  worksheetId: string;
+  /** Opening greeting (tab-dependent). */
+  greeting: string;
+}
+
+interface Msg {
+  id: number;
+  role: 'user' | 'assistant';
+  text?: string;
+  /** A rendered ThoughtSpot answer (from BodylessConversation). */
+  container?: HTMLElement;
+  loading?: boolean;
+  error?: boolean;
+}
+
+let idSeq = 1;
+
+/** Mounts a ThoughtSpot answer DOM node returned by BodylessConversation. */
+function AnswerEmbed({ container }: { container: HTMLElement }) {
+  const ref = useRef<HTMLDivElement>(null);
+  useEffect(() => {
+    const node = ref.current;
+    if (node && container) node.appendChild(container);
+    return () => {
+      if (node && container && node.contains(container)) node.removeChild(container);
+    };
+  }, [container]);
+  return <div className="chat-answer" ref={ref} />;
+}
+
+export default function ChatBot({ worksheetId, greeting }: ChatBotProps) {
+  const [open, setOpen] = useState(false);
+  const [input, setInput] = useState('');
+  const [busy, setBusy] = useState(false);
+  const [msgs, setMsgs] = useState<Msg[]>([
+    { id: 0, role: 'assistant', text: greeting },
+  ]);
+  const convRef = useRef<BodylessConversation | null>(null);
+  const scrollRef = useRef<HTMLDivElement>(null);
+  const firstRun = useRef(true);
+
+  useEffect(() => {
+    scrollRef.current?.scrollTo({ top: 1e9, behavior: 'smooth' });
+  }, [msgs, open]);
+
+  // When the tab context changes (worksheet/greeting), start a fresh,
+  // context-specific conversation against the new data model.
+  useEffect(() => {
+    if (firstRun.current) {
+      firstRun.current = false;
+      return;
+    }
+    convRef.current = null;
+    setMsgs([{ id: idSeq++, role: 'assistant', text: greeting }]);
+  }, [worksheetId, greeting]);
+
+  function getConversation() {
+    if (!convRef.current) {
+      convRef.current = new BodylessConversation({ worksheetId });
+    }
+    return convRef.current;
+  }
+
+  const update = (id: number, patch: Partial<Msg>) =>
+    setMsgs((m) => m.map((x) => (x.id === id ? { ...x, ...patch } : x)));
+
+  async function send() {
+    const q = input.trim();
+    if (!q || busy) return;
+    setInput('');
+    setBusy(true);
+
+    const history: ChatTurn[] = msgs
+      .filter((m) => m.text)
+      .map((m) => ({ role: m.role, content: m.text as string }));
+
+    const userMsg: Msg = { id: idSeq++, role: 'user', text: q };
+    const reply: Msg = { id: idSeq++, role: 'assistant', loading: true };
+    setMsgs((m) => [...m, userMsg, reply]);
+
+    try {
+      const result = await routeMessage(history, q);
+      if (result.kind === 'text') {
+        update(reply.id, { text: result.text, loading: false });
+      } else {
+        // Show the preamble while Spotter renders the answer.
+        update(reply.id, { text: result.preamble, loading: true });
+        const { container, error } = await getConversation().sendMessage(result.query);
+        if (error || !container) {
+          update(reply.id, {
+            loading: false,
+            error: true,
+            text:
+              (result.preamble ? result.preamble + '\n\n' : '') +
+              'I couldn’t load that analytics result. Make sure you’re signed in to ThoughtSpot and the data model is reachable.',
+          });
+        } else {
+          update(reply.id, { text: result.preamble, container, loading: false });
+        }
+      }
+    } catch {
+      update(reply.id, {
+        loading: false,
+        error: true,
+        text: 'Something went wrong. Please try again.',
+      });
+    } finally {
+      setBusy(false);
+    }
+  }
+
+  function onKeyDown(e: React.KeyboardEvent<HTMLTextAreaElement>) {
+    if (e.key === 'Enter' && !e.shiftKey) {
+      e.preventDefault();
+      send();
+    }
+  }
+
+  if (!open) {
+    return (
+      <button
+        className="chat-fab"
+        onClick={() => setOpen(true)}
+        aria-label="Open Northwind AI"
+      >
+        <Wand2 size={22} strokeWidth={2.2} />
+      </button>
+    );
+  }
+
+  return (
+    <div className="chat-panel">
+      <div className="chat-header">
+        <div className="chat-title">
+          <Wand2 size={18} />
+          <span>Northwind AI</span>
+        </div>
+        <button className="chat-close" onClick={() => setOpen(false)} aria-label="Close">
+          <X size={18} />
+        </button>
+      </div>
+
+      <div className="chat-messages" ref={scrollRef}>
+        {msgs.map((m) => (
+          <div key={m.id} className={`chat-msg ${m.role}`}>
+            <div className={`chat-bubble ${m.error ? 'is-error' : ''}`}>
+              {m.text && <p className="chat-text">{m.text}</p>}
+              {m.loading && (
+                <span className="chat-typing">
+                  <i /> <i /> <i />
+                </span>
+              )}
+              {m.container && <AnswerEmbed container={m.container} />}
+            </div>
+          </div>
+        ))}
+      </div>
+
+      <div className="chat-input-row">
+        <textarea
+          className="chat-input"
+          rows={1}
+          placeholder="Ask Northwind AI…"
+          value={input}
+          onChange={(e) => setInput(e.target.value)}
+          onKeyDown={onKeyDown}
+        />
+        <button
+          className="chat-send"
+          onClick={send}
+          disabled={!input.trim() || busy}
+          aria-label="Send"
+        >
+          <ArrowUp size={18} />
+        </button>
+      </div>
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/ColumnFilter.tsx

@@ -0,0 +1,149 @@
+// Generic single-column dropdown filter. Fetches distinct values for the first
+// matching candidate column and lets the user pick one or more; the selection is
+// handed back to the caller as (resolvedColumn, values) so it can push a runtime
+// filter (RuntimeFilterOp.IN). Nothing is selected by default (= no filter).
+// Reuses the shared sl-hfilter / sl-tree styles.
+import { useEffect, useMemo, useRef, useState } from 'react';
+import { ChevronDown, ChevronRight, Search, X, type LucideIcon } from 'lucide-react';
+import { useAuth } from '../context/AuthContext';
+import { fetchColumnValues } from '../lib/thoughtspot';
+
+interface Props {
+  /** Label shown on the trigger, e.g. "Year" or "Scope". */
+  label: string;
+  icon: LucideIcon;
+  /** Candidate column names — first that returns data is used. */
+  candidates: string[];
+  /** Called with the resolved column + selected values (empty = show all). */
+  onApply: (column: string, values: string[]) => void;
+}
+
+export default function ColumnFilter({ label, icon: Icon, candidates, onApply }: Props) {
+  const { username, password } = useAuth();
+  const [open, setOpen] = useState(false);
+  const [column, setColumn] = useState('');
+  const [values, setValues] = useState<string[]>([]);
+  const [loading, setLoading] = useState(true);
+  const [error, setError] = useState('');
+  const [query, setQuery] = useState('');
+  const [selected, setSelected] = useState<Set<string>>(new Set());
+  const [appliedCount, setAppliedCount] = useState(0);
+  const rootRef = useRef<HTMLDivElement>(null);
+
+  useEffect(() => {
+    (async () => {
+      try {
+        const res = await fetchColumnValues(username, password, candidates);
+        setColumn(res.column);
+        setValues(res.values);
+      } catch (e) {
+        setError(e instanceof Error ? e.message : `Failed to load ${label}.`);
+      } finally {
+        setLoading(false);
+      }
+    })();
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, []);
+
+  useEffect(() => {
+    if (!open) return;
+    const h = (e: MouseEvent) => {
+      if (rootRef.current && !rootRef.current.contains(e.target as Node)) setOpen(false);
+    };
+    document.addEventListener('mousedown', h);
+    return () => document.removeEventListener('mousedown', h);
+  }, [open]);
+
+  const filtered = useMemo(() => {
+    const q = query.trim().toLowerCase();
+    return q ? values.filter((v) => v.toLowerCase().includes(q)) : values;
+  }, [values, query]);
+
+  const allSelected = values.length > 0 && values.every((v) => selected.has(v));
+
+  function toggle(v: string) {
+    setSelected((p) => {
+      const s = new Set(p);
+      s.has(v) ? s.delete(v) : s.add(v);
+      return s;
+    });
+  }
+  function toggleAll() {
+    setSelected(allSelected ? new Set() : new Set(values));
+  }
+  function apply() {
+    const list = [...selected];
+    onApply(column, list);
+    setAppliedCount(list.length);
+    setOpen(false);
+  }
+  function clearAll() {
+    setSelected(new Set());
+    onApply(column, []);
+    setAppliedCount(0);
+    setOpen(false);
+  }
+
+  const state =
+    appliedCount === 0 ? 'All' : `${appliedCount} selected`;
+
+  return (
+    <div className="sl-hfilter" ref={rootRef}>
+      <button className="sl-hfilter-trigger" onClick={() => setOpen((o) => !o)}>
+        <Icon size={15} />
+        <span>
+          {label}: {state}
+        </span>
+        {open ? <ChevronDown size={16} /> : <ChevronRight size={16} />}
+      </button>
+
+      {open && (
+        <div className="sl-hfilter-panel">
+          <div className="sl-hfilter-search">
+            <Search size={15} />
+            <input
+              placeholder={`Search ${label.toLowerCase()}…`}
+              value={query}
+              onChange={(e) => setQuery(e.target.value)}
+              autoFocus
+            />
+            {query && (
+              <button onClick={() => setQuery('')} aria-label="Clear search">
+                <X size={13} />
+              </button>
+            )}
+          </div>
+
+          {loading && <div className="sl-hfilter-state">Loading…</div>}
+          {error && <div className="sl-hfilter-state sl-hfilter-error">{error}</div>}
+
+          {!loading && !error && (
+            <>
+              <label className="sl-tree-row sl-tree-all">
+                <input type="checkbox" checked={allSelected} onChange={toggleAll} />
+                <span>Select all</span>
+              </label>
+              <div className="sl-tree">
+                {filtered.map((v) => (
+                  <label key={v} className="sl-tree-row sl-tree-l1">
+                    <input type="checkbox" checked={selected.has(v)} onChange={() => toggle(v)} />
+                    <span className="sl-tree-l1-name">{v}</span>
+                  </label>
+                ))}
+                {filtered.length === 0 && <div className="sl-hfilter-state">No matches</div>}
+              </div>
+              <div className="sl-hfilter-actions">
+                <button className="sl-hfilter-clear" onClick={clearAll}>
+                  Clear
+                </button>
+                <button className="sl-hfilter-apply" onClick={apply}>
+                  Apply
+                </button>
+              </div>
+            </>
+          )}
+        </div>
+      )}
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/CreateDashboardModal.tsx

@@ -0,0 +1,108 @@
+import { useEffect, useState } from 'react';
+import { X, LayoutDashboard, Plus, AlertCircle } from 'lucide-react';
+
+interface Props {
+  open: boolean;
+  onClose: () => void;
+  /** Creates the liveboard; should throw on failure so the modal can show it. */
+  onCreate: (name: string, description: string) => Promise<void>;
+}
+
+export default function CreateDashboardModal({ open, onClose, onCreate }: Props) {
+  const [name, setName] = useState('');
+  const [description, setDescription] = useState('');
+  const [submitting, setSubmitting] = useState(false);
+  const [error, setError] = useState('');
+
+  useEffect(() => {
+    if (open) {
+      setName('');
+      setDescription('');
+      setError('');
+      setSubmitting(false);
+    }
+  }, [open]);
+
+  if (!open) return null;
+
+  async function submit() {
+    if (!name.trim() || submitting) return;
+    setSubmitting(true);
+    setError('');
+    try {
+      await onCreate(name.trim(), description.trim());
+      // On success the parent closes this modal.
+    } catch (e) {
+      setError(e instanceof Error ? e.message : 'Could not create the dashboard.');
+      setSubmitting(false);
+    }
+  }
+
+  return (
+    <div className="winback-overlay" onClick={submitting ? undefined : onClose}>
+      <div className="winback-modal" onClick={(e) => e.stopPropagation()}>
+        <div className="winback-header">
+          <div className="winback-title">
+            <LayoutDashboard size={18} />
+            <span>Create New Dashboard</span>
+          </div>
+          <button
+            className="winback-close"
+            onClick={onClose}
+            disabled={submitting}
+            aria-label="Close"
+          >
+            <X size={18} />
+          </button>
+        </div>
+
+        <div className="winback-body">
+          {error && (
+            <div className="ma-inline-error">
+              <AlertCircle size={15} /> {error}
+            </div>
+          )}
+          <div className="winback-field" style={{ marginBottom: 16 }}>
+            <label className="winback-label">Dashboard name</label>
+            <input
+              className="login-input"
+              value={name}
+              onChange={(e) => setName(e.target.value)}
+              placeholder="e.g. Q3 Pipeline Review"
+              autoFocus
+              onKeyDown={(e) => e.key === 'Enter' && submit()}
+            />
+          </div>
+          <div className="winback-field">
+            <label className="winback-label">Description (optional)</label>
+            <input
+              className="login-input"
+              value={description}
+              onChange={(e) => setDescription(e.target.value)}
+              placeholder="What is this dashboard for?"
+              onKeyDown={(e) => e.key === 'Enter' && submit()}
+            />
+          </div>
+        </div>
+
+        <div className="winback-footer">
+          <button
+            className="winback-btn-secondary"
+            onClick={onClose}
+            disabled={submitting}
+          >
+            Cancel
+          </button>
+          <button
+            className="winback-btn-primary"
+            onClick={submit}
+            disabled={!name.trim() || submitting}
+          >
+            {submitting ? <span className="btn-spinner" /> : <Plus size={16} />}
+            {submitting ? 'Creating…' : 'Create Dashboard'}
+          </button>
+        </div>
+      </div>
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/DateRangeFilter.tsx

@@ -0,0 +1,155 @@
+// Host-side date-range filter for the Analytics tab. Presets + a custom range
+// produce an epoch-SECONDS window that the tab pushes to the liveboard as a
+// runtime filter on the Cadence Create Date column.
+import { useEffect, useRef, useState } from 'react';
+import { CalendarRange, ChevronDown, ChevronRight, X } from 'lucide-react';
+
+/** Epoch SECONDS range (either bound optional). null = no date filter. */
+export interface DateSelection {
+  start: number | null;
+  end: number | null;
+  label: string;
+}
+
+interface Props {
+  onApply: (sel: DateSelection | null) => void;
+}
+
+type PresetId = 'all' | '7d' | '30d' | '90d' | '12m' | 'ytd' | 'custom';
+
+const PRESETS: { id: PresetId; label: string }[] = [
+  { id: 'all', label: 'All time' },
+  { id: '7d', label: 'Last 7 days' },
+  { id: '30d', label: 'Last 30 days' },
+  { id: '90d', label: 'Last 90 days' },
+  { id: '12m', label: 'Last 12 months' },
+  { id: 'ytd', label: 'Year to date' },
+  { id: 'custom', label: 'Custom range' },
+];
+
+const startOfDay = (d: Date) => new Date(d.getFullYear(), d.getMonth(), d.getDate()).getTime();
+const endOfDay = (d: Date) =>
+  new Date(d.getFullYear(), d.getMonth(), d.getDate(), 23, 59, 59).getTime();
+const sec = (ms: number) => Math.floor(ms / 1000);
+
+function presetRange(id: PresetId): { start: number | null; end: number | null } {
+  const now = new Date();
+  const end = sec(endOfDay(now));
+  const daysAgo = (n: number) => {
+    const d = new Date();
+    d.setDate(d.getDate() - n);
+    return sec(startOfDay(d));
+  };
+  switch (id) {
+    case '7d':
+      return { start: daysAgo(7), end };
+    case '30d':
+      return { start: daysAgo(30), end };
+    case '90d':
+      return { start: daysAgo(90), end };
+    case '12m': {
+      const d = new Date();
+      d.setFullYear(d.getFullYear() - 1);
+      return { start: sec(startOfDay(d)), end };
+    }
+    case 'ytd':
+      return { start: sec(new Date(now.getFullYear(), 0, 1).getTime()), end };
+    default:
+      return { start: null, end: null };
+  }
+}
+
+export default function DateRangeFilter({ onApply }: Props) {
+  const [open, setOpen] = useState(false);
+  const [active, setActive] = useState<PresetId>('all');
+  const [label, setLabel] = useState('All time');
+  const [from, setFrom] = useState('');
+  const [to, setTo] = useState('');
+  const rootRef = useRef<HTMLDivElement>(null);
+
+  useEffect(() => {
+    if (!open) return;
+    const h = (e: MouseEvent) => {
+      if (rootRef.current && !rootRef.current.contains(e.target as Node)) setOpen(false);
+    };
+    document.addEventListener('mousedown', h);
+    return () => document.removeEventListener('mousedown', h);
+  }, [open]);
+
+  function choosePreset(id: PresetId) {
+    setActive(id);
+    if (id === 'custom') return; // wait for Apply on custom inputs
+    const { start, end } = presetRange(id);
+    const lbl = PRESETS.find((p) => p.id === id)!.label;
+    setLabel(lbl);
+    setOpen(false);
+    onApply(id === 'all' ? null : { start, end, label: lbl });
+  }
+
+  function applyCustom() {
+    const start = from ? sec(startOfDay(new Date(from))) : null;
+    const end = to ? sec(endOfDay(new Date(to))) : null;
+    if (start == null && end == null) {
+      setLabel('All time');
+      onApply(null);
+    } else {
+      const lbl = `${from || '…'} → ${to || '…'}`;
+      setLabel(lbl);
+      onApply({ start, end, label: lbl });
+    }
+    setOpen(false);
+  }
+
+  return (
+    <div className="sl-datefilter" ref={rootRef}>
+      <button className="sl-hfilter-trigger" onClick={() => setOpen((o) => !o)}>
+        <CalendarRange size={15} />
+        <span>Created: {label}</span>
+        {open ? <ChevronDown size={16} /> : <ChevronRight size={16} />}
+      </button>
+
+      {open && (
+        <div className="sl-hfilter-panel sl-datefilter-panel">
+          <div className="sl-date-presets">
+            {PRESETS.map((p) => (
+              <button
+                key={p.id}
+                className={`sl-date-preset${active === p.id ? ' is-active' : ''}`}
+                onClick={() => choosePreset(p.id)}
+              >
+                {p.label}
+              </button>
+            ))}
+          </div>
+
+          {active === 'custom' && (
+            <div className="sl-date-custom">
+              <label>
+                <span>From</span>
+                <input type="date" value={from} onChange={(e) => setFrom(e.target.value)} />
+              </label>
+              <label>
+                <span>To</span>
+                <input type="date" value={to} onChange={(e) => setTo(e.target.value)} />
+              </label>
+              <div className="sl-hfilter-actions">
+                <button
+                  className="sl-hfilter-clear"
+                  onClick={() => {
+                    setFrom('');
+                    setTo('');
+                  }}
+                >
+                  <X size={13} /> Reset
+                </button>
+                <button className="sl-hfilter-apply" onClick={applyCustom}>
+                  Apply
+                </button>
+              </div>
+            </div>
+          )}
+        </div>
+      )}
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/NorthwindLogo.tsx

@@ -0,0 +1,25 @@
+// Official Northwind logo (fetched from northwind.com, "on-light" variant).
+// Inlined so its ink can follow `currentColor` — dark evergreen on light
+// surfaces, light on dark surfaces — while keeping the lime accent. Color is
+// set by the host (.sl-logo → var(--sl-evergreen)); height comes from `size`.
+import logoRaw from '../assets/northwind-logo.svg?raw';
+
+interface Props {
+  className?: string;
+  /** Logo height in px (default 26). */
+  size?: number;
+  /** Kept for API compatibility; the official lockup always includes the wordmark. */
+  wordmark?: boolean;
+}
+
+export default function NorthwindLogo({ className, size = 26 }: Props) {
+  return (
+    <span
+      className={`sl-logo ${className ?? ''}`}
+      style={{ height: size }}
+      role="img"
+      aria-label="Northwind"
+      dangerouslySetInnerHTML={{ __html: logoRaw }}
+    />
+  );
+}

demoprep_app/portal/template-tse/src/components/PinModal.tsx

@@ -0,0 +1,81 @@
+// In-app "name this report" dialog used before pinning an answer to the
+// dashboard. Replaces window.prompt() (browser-native prompts are invisible in
+// screen shares / recordings) with an app-rendered, theme-aware modal.
+import { useEffect, useRef, useState } from 'react';
+import { Pin, X } from 'lucide-react';
+
+interface Props {
+  open: boolean;
+  defaultName?: string;
+  onCancel: () => void;
+  onConfirm: (name: string) => void;
+}
+
+export default function PinModal({ open, defaultName = 'New report', onCancel, onConfirm }: Props) {
+  const [name, setName] = useState(defaultName);
+  const inputRef = useRef<HTMLInputElement>(null);
+
+  useEffect(() => {
+    if (open) {
+      setName(defaultName);
+      // Focus + select so the user can type over the default immediately.
+      setTimeout(() => {
+        inputRef.current?.focus();
+        inputRef.current?.select();
+      }, 0);
+    }
+  }, [open, defaultName]);
+
+  if (!open) return null;
+
+  function confirm() {
+    const trimmed = name.trim();
+    if (trimmed) onConfirm(trimmed);
+  }
+
+  return (
+    <div className="sl-pin-overlay" onClick={onCancel}>
+      <div
+        className="sl-pin-modal"
+        onClick={(e) => e.stopPropagation()}
+        role="dialog"
+        aria-modal="true"
+      >
+        <div className="sl-pin-header">
+          <div className="sl-pin-title">
+            <Pin size={16} />
+            <span>Pin to dashboard</span>
+          </div>
+          <button className="sl-pin-close" onClick={onCancel} aria-label="Close">
+            <X size={18} />
+          </button>
+        </div>
+        <div className="sl-pin-body">
+          <label className="sl-pin-label" htmlFor="sl-pin-name">
+            Name this report
+          </label>
+          <input
+            id="sl-pin-name"
+            ref={inputRef}
+            className="sl-pin-input"
+            value={name}
+            onChange={(e) => setName(e.target.value)}
+            onKeyDown={(e) => {
+              if (e.key === 'Enter') confirm();
+              if (e.key === 'Escape') onCancel();
+            }}
+            placeholder="New report"
+          />
+        </div>
+        <div className="sl-pin-actions">
+          <button className="sl-pin-cancel" onClick={onCancel}>
+            Cancel
+          </button>
+          <button className="sl-pin-confirm" onClick={confirm} disabled={!name.trim()}>
+            <Pin size={15} /> Pin
+          </button>
+        </div>
+      </div>
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/RepFilter.tsx

@@ -0,0 +1,139 @@
+// Standalone "Rep" dropdown filter. Fetches distinct Rep Name values and lets
+// the user pick one or more; the selection is handed back to the Analytics tab,
+// which pushes it to the liveboard as a runtime filter on the Rep Name column
+// (RuntimeFilterOp.IN — confirmed via the SDK reference). Reuses the shared
+// sl-hfilter / sl-tree styles.
+import { useEffect, useMemo, useRef, useState } from 'react';
+import { ChevronDown, ChevronRight, Search, Users, X } from 'lucide-react';
+import { useAuth } from '../context/AuthContext';
+import { fetchRepNames } from '../lib/thoughtspot';
+
+interface Props {
+  /** Called with the selected rep names (empty = no filter / all reps). */
+  onApply: (reps: string[]) => void;
+}
+
+export default function RepFilter({ onApply }: Props) {
+  const { username, password } = useAuth();
+  const [open, setOpen] = useState(false);
+  const [reps, setReps] = useState<string[]>([]);
+  const [loading, setLoading] = useState(true);
+  const [error, setError] = useState('');
+  const [query, setQuery] = useState('');
+  const [selected, setSelected] = useState<Set<string>>(new Set());
+  const [appliedCount, setAppliedCount] = useState(0);
+  const rootRef = useRef<HTMLDivElement>(null);
+
+  useEffect(() => {
+    (async () => {
+      try {
+        setReps(await fetchRepNames(username, password));
+      } catch (e) {
+        setError(e instanceof Error ? e.message : 'Failed to load reps.');
+      } finally {
+        setLoading(false);
+      }
+    })();
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, []);
+
+  useEffect(() => {
+    if (!open) return;
+    const h = (e: MouseEvent) => {
+      if (rootRef.current && !rootRef.current.contains(e.target as Node)) setOpen(false);
+    };
+    document.addEventListener('mousedown', h);
+    return () => document.removeEventListener('mousedown', h);
+  }, [open]);
+
+  const filtered = useMemo(() => {
+    const q = query.trim().toLowerCase();
+    return q ? reps.filter((r) => r.toLowerCase().includes(q)) : reps;
+  }, [reps, query]);
+
+  const allSelected = reps.length > 0 && reps.every((r) => selected.has(r));
+
+  function toggle(r: string) {
+    setSelected((p) => {
+      const s = new Set(p);
+      s.has(r) ? s.delete(r) : s.add(r);
+      return s;
+    });
+  }
+  function toggleAll() {
+    setSelected(allSelected ? new Set() : new Set(reps));
+  }
+  function apply() {
+    const list = [...selected];
+    onApply(list);
+    setAppliedCount(list.length);
+    setOpen(false);
+  }
+  function clearAll() {
+    setSelected(new Set());
+    onApply([]);
+    setAppliedCount(0);
+    setOpen(false);
+  }
+
+  const label =
+    appliedCount === 0 ? 'All reps' : `${appliedCount} rep${appliedCount === 1 ? '' : 's'} selected`;
+
+  return (
+    <div className="sl-hfilter" ref={rootRef}>
+      <button className="sl-hfilter-trigger" onClick={() => setOpen((o) => !o)}>
+        <Users size={15} />
+        <span>Rep: {label}</span>
+        {open ? <ChevronDown size={16} /> : <ChevronRight size={16} />}
+      </button>
+
+      {open && (
+        <div className="sl-hfilter-panel">
+          <div className="sl-hfilter-search">
+            <Search size={15} />
+            <input
+              placeholder="Search reps…"
+              value={query}
+              onChange={(e) => setQuery(e.target.value)}
+              autoFocus
+            />
+            {query && (
+              <button onClick={() => setQuery('')} aria-label="Clear search">
+                <X size={13} />
+              </button>
+            )}
+          </div>
+
+          {loading && <div className="sl-hfilter-state">Loading…</div>}
+          {error && <div className="sl-hfilter-state sl-hfilter-error">{error}</div>}
+
+          {!loading && !error && (
+            <>
+              <label className="sl-tree-row sl-tree-all">
+                <input type="checkbox" checked={allSelected} onChange={toggleAll} />
+                <span>Select all</span>
+              </label>
+              <div className="sl-tree">
+                {filtered.map((r) => (
+                  <label key={r} className="sl-tree-row sl-tree-l1">
+                    <input type="checkbox" checked={selected.has(r)} onChange={() => toggle(r)} />
+                    <span className="sl-tree-l1-name">{r}</span>
+                  </label>
+                ))}
+                {filtered.length === 0 && <div className="sl-hfilter-state">No matches</div>}
+              </div>
+              <div className="sl-hfilter-actions">
+                <button className="sl-hfilter-clear" onClick={clearAll}>
+                  Clear
+                </button>
+                <button className="sl-hfilter-apply" onClick={apply}>
+                  Apply
+                </button>
+              </div>
+            </>
+          )}
+        </div>
+      )}
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/RepHierarchyFilter.tsx

@@ -0,0 +1,268 @@
+// Host-side 3-level tree filter for the Analytics tab:
+// Rep Segment → Rep Name → Cadence Name. Options are fetched from the model via
+// searchdata; the selection is collapsed into per-level value sets which the tab
+// pushes to the liveboard as runtime filters.
+import { useEffect, useMemo, useRef, useState } from 'react';
+import { ChevronRight, ChevronDown, Search, SlidersHorizontal, X } from 'lucide-react';
+import { useAuth } from '../context/AuthContext';
+import { fetchSegmentRepHierarchy, SegmentNode } from '../lib/thoughtspot';
+
+/** Selection collapsed into per-level value sets for runtime filters. */
+export interface HierarchySelection {
+  segments: string[];
+  reps: string[];
+  cadences: string[];
+}
+
+interface Props {
+  onApply: (selection: HierarchySelection) => void;
+}
+
+const SEP = '||';
+const leafKey = (s: string, r: string, c: string) => `${s}${SEP}${r}${SEP}${c}`;
+
+type RepLike = { rep: string; cadences: string[] };
+
+export default function RepHierarchyFilter({ onApply }: Props) {
+  const { username, password } = useAuth();
+  const [open, setOpen] = useState(false);
+  const [nodes, setNodes] = useState<SegmentNode[]>([]);
+  const [loading, setLoading] = useState(true);
+  const [error, setError] = useState('');
+  const [query, setQuery] = useState('');
+  const [expSeg, setExpSeg] = useState<Set<string>>(new Set());
+  const [expRep, setExpRep] = useState<Set<string>>(new Set());
+  const [selected, setSelected] = useState<Set<string>>(new Set()); // leaf keys
+  const [appliedCount, setAppliedCount] = useState(0);
+  const rootRef = useRef<HTMLDivElement>(null);
+
+  useEffect(() => {
+    (async () => {
+      try {
+        setNodes(await fetchSegmentRepHierarchy(username, password));
+      } catch (e) {
+        setError(e instanceof Error ? e.message : 'Failed to load filter options.');
+      } finally {
+        setLoading(false);
+      }
+    })();
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, []);
+
+  useEffect(() => {
+    if (!open) return;
+    const h = (e: MouseEvent) => {
+      if (rootRef.current && !rootRef.current.contains(e.target as Node)) setOpen(false);
+    };
+    document.addEventListener('mousedown', h);
+    return () => document.removeEventListener('mousedown', h);
+  }, [open]);
+
+  const segLeaves = (n: SegmentNode) =>
+    n.reps.flatMap((r) => r.cadences.map((c) => leafKey(n.segment, r.rep, c)));
+  const repLeaves = (seg: string, r: RepLike) => r.cadences.map((c) => leafKey(seg, r.rep, c));
+  const allLeaves = useMemo(() => nodes.flatMap(segLeaves), [nodes]);
+
+  const filtered = useMemo(() => {
+    const q = query.trim().toLowerCase();
+    if (!q) return nodes;
+    return nodes
+      .map((n) => {
+        const segMatch = n.segment.toLowerCase().includes(q);
+        const reps = n.reps
+          .map((r) => {
+            const repMatch = segMatch || r.rep.toLowerCase().includes(q);
+            const cadences = repMatch ? r.cadences : r.cadences.filter((c) => c.toLowerCase().includes(q));
+            return repMatch || cadences.length ? { rep: r.rep, cadences } : null;
+          })
+          .filter((r): r is RepLike => r != null);
+        return segMatch || reps.length ? { segment: n.segment, reps } : null;
+      })
+      .filter((n): n is SegmentNode => n != null);
+  }, [nodes, query]);
+
+  function nodeState(keys: string[]): 'none' | 'some' | 'all' {
+    if (keys.length === 0) return 'none';
+    const sel = keys.filter((k) => selected.has(k)).length;
+    if (sel === 0) return 'none';
+    return sel === keys.length ? 'all' : 'some';
+  }
+  function toggleKeys(keys: string[]) {
+    const all = nodeState(keys) === 'all';
+    setSelected((p) => {
+      const s = new Set(p);
+      keys.forEach((k) => (all ? s.delete(k) : s.add(k)));
+      return s;
+    });
+  }
+  function toggleLeaf(k: string) {
+    setSelected((p) => {
+      const s = new Set(p);
+      s.has(k) ? s.delete(k) : s.add(k);
+      return s;
+    });
+  }
+
+  const allSelected = allLeaves.length > 0 && allLeaves.every((k) => selected.has(k));
+
+  function deriveSelection(): HierarchySelection {
+    const segments = new Set<string>();
+    const reps = new Set<string>();
+    const cadences = new Set<string>();
+    for (const key of selected) {
+      const [s, r, c] = key.split(SEP);
+      segments.add(s);
+      reps.add(r);
+      cadences.add(c);
+    }
+    return { segments: [...segments], reps: [...reps], cadences: [...cadences] };
+  }
+
+  function apply() {
+    onApply(deriveSelection());
+    setAppliedCount(selected.size);
+    setOpen(false);
+  }
+  function clearAll() {
+    setSelected(new Set());
+    onApply({ segments: [], reps: [], cadences: [] });
+    setAppliedCount(0);
+    setOpen(false);
+  }
+
+  const label =
+    appliedCount === 0
+      ? 'All segments & cadences'
+      : `${appliedCount} cadence${appliedCount === 1 ? '' : 's'} selected`;
+
+  return (
+    <div className="sl-hfilter" ref={rootRef}>
+      <button className="sl-hfilter-trigger" onClick={() => setOpen((o) => !o)}>
+        <SlidersHorizontal size={15} />
+        <span>{label}</span>
+        {open ? <ChevronDown size={16} /> : <ChevronRight size={16} />}
+      </button>
+
+      {open && (
+        <div className="sl-hfilter-panel">
+          <div className="sl-hfilter-search">
+            <Search size={15} />
+            <input
+              placeholder="Search segments, reps or cadences…"
+              value={query}
+              onChange={(e) => setQuery(e.target.value)}
+              autoFocus
+            />
+            {query && (
+              <button onClick={() => setQuery('')} aria-label="Clear search">
+                <X size={13} />
+              </button>
+            )}
+          </div>
+
+          {loading && <div className="sl-hfilter-state">Loading…</div>}
+          {error && <div className="sl-hfilter-state sl-hfilter-error">{error}</div>}
+
+          {!loading && !error && (
+            <>
+              <label className="sl-tree-row sl-tree-all">
+                <input type="checkbox" checked={allSelected} onChange={() => toggleKeys(allLeaves)} />
+                <span>Select all</span>
+              </label>
+
+              <div className="sl-tree">
+                {filtered.map((n) => {
+                  const sKeys = segLeaves(n);
+                  const sState = nodeState(sKeys);
+                  const sOpen = expSeg.has(n.segment) || !!query;
+                  return (
+                    <div key={n.segment}>
+                      <div className="sl-tree-row sl-tree-l1">
+                        <button
+                          className="sl-tree-caret"
+                          onClick={() =>
+                            setExpSeg((p) => {
+                              const s = new Set(p);
+                              s.has(n.segment) ? s.delete(n.segment) : s.add(n.segment);
+                              return s;
+                            })
+                          }
+                        >
+                          {sOpen ? <ChevronDown size={14} /> : <ChevronRight size={14} />}
+                        </button>
+                        <input
+                          type="checkbox"
+                          checked={sState === 'all'}
+                          ref={(el) => el && (el.indeterminate = sState === 'some')}
+                          onChange={() => toggleKeys(sKeys)}
+                        />
+                        <span className="sl-tree-l1-name">{n.segment}</span>
+                      </div>
+
+                      {sOpen &&
+                        n.reps.map((r) => {
+                          const rKeys = repLeaves(n.segment, r);
+                          const rState = nodeState(rKeys);
+                          const rId = `${n.segment}${SEP}${r.rep}`;
+                          const rOpen = expRep.has(rId) || !!query;
+                          return (
+                            <div key={rId}>
+                              <div className="sl-tree-row sl-tree-l2">
+                                <button
+                                  className="sl-tree-caret"
+                                  onClick={() =>
+                                    setExpRep((p) => {
+                                      const s = new Set(p);
+                                      s.has(rId) ? s.delete(rId) : s.add(rId);
+                                      return s;
+                                    })
+                                  }
+                                >
+                                  {rOpen ? <ChevronDown size={14} /> : <ChevronRight size={14} />}
+                                </button>
+                                <input
+                                  type="checkbox"
+                                  checked={rState === 'all'}
+                                  ref={(el) => el && (el.indeterminate = rState === 'some')}
+                                  onChange={() => toggleKeys(rKeys)}
+                                />
+                                <span>{r.rep}</span>
+                              </div>
+                              {rOpen &&
+                                r.cadences.map((c) => {
+                                  const k = leafKey(n.segment, r.rep, c);
+                                  return (
+                                    <label key={k} className="sl-tree-row sl-tree-l3">
+                                      <input
+                                        type="checkbox"
+                                        checked={selected.has(k)}
+                                        onChange={() => toggleLeaf(k)}
+                                      />
+                                      <span>{c}</span>
+                                    </label>
+                                  );
+                                })}
+                            </div>
+                          );
+                        })}
+                    </div>
+                  );
+                })}
+                {filtered.length === 0 && <div className="sl-hfilter-state">No matches</div>}
+              </div>
+
+              <div className="sl-hfilter-actions">
+                <button className="sl-hfilter-clear" onClick={clearAll}>
+                  Clear
+                </button>
+                <button className="sl-hfilter-apply" onClick={apply}>
+                  Apply
+                </button>
+              </div>
+            </>
+          )}
+        </div>
+      )}
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/TopBar.tsx

@@ -0,0 +1,148 @@
+import {
+  BarChart3,
+  LayoutGrid,
+  Sparkles,
+  Truck,
+  Activity,
+  Bot,
+  ChevronDown,
+  LogOut,
+  Sun,
+  Moon,
+  Crown,
+  Check,
+} from 'lucide-react';
+import { useEffect, useRef, useState } from 'react';
+import { useAuth } from '../context/AuthContext';
+import { useTheme } from '../context/ThemeContext';
+import { useTier } from '../context/TierContext';
+import NorthwindLogo from './NorthwindLogo';
+import { CONTENT } from '../content';
+import { FLAGS } from '../flags';
+
+export type TabId = 'my-analytics' | 'analytics' | 'carriers' | 'capacity' | 'ask' | 'spotter';
+
+const TABS: { id: TabId; label: string; icon: typeof BarChart3 }[] = [
+  { id: 'my-analytics', label: CONTENT.tabs.myReports, icon: LayoutGrid },
+  { id: 'analytics', label: CONTENT.tabs.analytics, icon: BarChart3 },
+  { id: 'carriers', label: CONTENT.tabs.inline, icon: Truck },
+  { id: 'capacity', label: CONTENT.tabs.action, icon: Activity },
+  { id: 'ask', label: CONTENT.tabs.ask, icon: Sparkles },
+  { id: 'spotter', label: CONTENT.tabs.spotter, icon: Bot },
+];
+
+// Decorative Northwind platform nav (sets product context, non-interactive).
+const PLATFORM_NAV = [...CONTENT.platformNav];
+
+interface Props {
+  active: TabId;
+  onChange: (tab: TabId) => void;
+}
+
+export default function TopBar({ active, onChange }: Props) {
+  const { username, logout } = useAuth();
+  const { theme, toggle } = useTheme();
+  const { tier, setTier } = useTier();
+  const [menuOpen, setMenuOpen] = useState(false);
+  const [tierOpen, setTierOpen] = useState(false);
+  const initials = (username || 'U').slice(0, 2).toUpperCase();
+  const tierRef = useRef<HTMLDivElement>(null);
+  const userRef = useRef<HTMLDivElement>(null);
+
+  // Close either dropdown when clicking anywhere outside it.
+  useEffect(() => {
+    if (!menuOpen && !tierOpen) return;
+    const onDown = (e: MouseEvent) => {
+      const t = e.target as Node;
+      if (tierRef.current && !tierRef.current.contains(t)) setTierOpen(false);
+      if (userRef.current && !userRef.current.contains(t)) setMenuOpen(false);
+    };
+    document.addEventListener('mousedown', onDown);
+    return () => document.removeEventListener('mousedown', onDown);
+  }, [menuOpen, tierOpen]);
+
+  return (
+    <header className="topbar">
+      <div className="topbar-left">
+        <NorthwindLogo className="topbar-logo" />
+        <nav className="topbar-platform-nav">
+          {PLATFORM_NAV.map((item) => (
+            <span key={item} className="topbar-platform-link">
+              {item}
+            </span>
+          ))}
+        </nav>
+      </div>
+
+      <nav className="topbar-tabs">
+        {TABS.map((tab) => {
+          const Icon = tab.icon;
+          return (
+            <button
+              key={tab.id}
+              className={`topbar-tab ${active === tab.id ? 'is-active' : ''}`}
+              onClick={() => onChange(tab.id)}
+            >
+              <Icon size={17} strokeWidth={2} />
+              <span>{tab.label}</span>
+            </button>
+          );
+        })}
+      </nav>
+
+      <div className="topbar-right">
+        <button
+          className="topbar-theme-toggle"
+          onClick={toggle}
+          title={theme === 'dark' ? 'Switch to light theme' : 'Switch to dark theme'}
+          aria-label="Toggle theme"
+        >
+          {theme === 'dark' ? <Sun size={18} /> : <Moon size={18} />}
+        </button>
+
+        {/* Tier switcher — demo control to preview Premium vs Basic gating. */}
+        {FLAGS.tiers && (
+        <div className="topbar-tier" ref={tierRef} onClick={() => setTierOpen((o) => !o)}>
+          <span className={`tier-pill tier-${tier}`}>
+            <Crown size={13} />
+            {tier === 'premium' ? 'Premium' : 'Basic'}
+          </span>
+          <ChevronDown size={15} />
+          {tierOpen && (
+            <div className="topbar-menu tier-menu">
+              <button
+                className="topbar-menu-item"
+                onClick={() => { setTier('premium'); setTierOpen(false); }}
+              >
+                <Crown size={14} /> Premium
+                {tier === 'premium' && <Check size={14} className="tier-check" />}
+              </button>
+              <button
+                className="topbar-menu-item"
+                onClick={() => { setTier('basic'); setTierOpen(false); }}
+              >
+                <span style={{ width: 14 }} /> Basic
+                {tier === 'basic' && <Check size={14} className="tier-check" />}
+              </button>
+            </div>
+          )}
+        </div>
+        )}
+
+        <div className="topbar-user" ref={userRef} onClick={() => setMenuOpen((o) => !o)}>
+          <span className="topbar-avatar">{initials}</span>
+          <span className="topbar-username">{username || 'User'}</span>
+          <ChevronDown size={16} />
+          {menuOpen && (
+            <div className="topbar-menu">
+              <button className="topbar-menu-item" onClick={logout}>
+                <LogOut size={15} />
+                Sign out
+              </button>
+            </div>
+          )}
+        </div>
+      </div>
+    </header>
+  );
+}

demoprep_app/portal/template-tse/src/components/TrialModal.tsx

@@ -0,0 +1,67 @@
+// Professional upgrade prompt shown once (on the 3rd Ask Northwind AI question).
+import { X, Sparkles, Check, ArrowRight } from 'lucide-react';
+import { Northwind_UPGRADE_URL } from '../config';
+
+interface Props {
+  open: boolean;
+  remaining: number;
+  onClose: () => void;
+}
+
+const PREMIUM_FEATURES = [
+  'Unlimited Ask Northwind AI questions',
+  'Drill-down & underlying-data access',
+  'CSV, Excel & PDF exports',
+  'Priority capacity & lane insights',
+  'Dedicated onboarding & support',
+];
+
+export default function TrialModal({ open, remaining, onClose }: Props) {
+  if (!open) return null;
+
+  return (
+    <div className="sl-trial-overlay" onClick={onClose}>
+      <div className="sl-trial-modal" onClick={(e) => e.stopPropagation()} role="dialog" aria-modal="true">
+        <button className="sl-trial-close" onClick={onClose} aria-label="Close">
+          <X size={18} />
+        </button>
+
+        <div className="sl-trial-body">
+          <div className="sl-trial-icon">
+            <Sparkles size={24} />
+          </div>
+          <div className="sl-trial-eyebrow">Northwind Premium</div>
+          <h2 className="sl-trial-title">Unlock unlimited Northwind AI</h2>
+          <p className="sl-trial-sub">
+            {remaining > 0 ? (
+              <>You have <strong>{remaining}</strong> free question{remaining === 1 ? '' : 's'} left on the trial. </>
+            ) : (
+              <>You&rsquo;ve reached your free-trial limit. </>
+            )}
+            Upgrade to Premium for unlimited access across your brokerage.
+          </p>
+
+          <ul className="sl-trial-features">
+            {PREMIUM_FEATURES.map((f) => (
+              <li key={f}>
+                <span className="sl-trial-check"><Check size={14} /></span>
+                {f}
+              </li>
+            ))}
+          </ul>
+
+          <a className="sl-trial-cta" href={Northwind_UPGRADE_URL} target="_blank" rel="noopener noreferrer">
+            Upgrade to Premium <ArrowRight size={18} />
+          </a>
+          <button className="sl-trial-dismiss" onClick={onClose}>
+            Maybe later
+          </button>
+          <p className="sl-trial-fine">
+            Demo experience — no payment is taken. Talk to your Northwind account team for
+            production pricing.
+          </p>
+        </div>
+      </div>
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/components/WinBackModal.tsx

@@ -0,0 +1,158 @@
+import { useEffect, useState } from 'react';
+import { X, Zap, Check, RefreshCw } from 'lucide-react';
+import { SignalContext } from '../lib/signalContext';
+
+const CADENCE_OPTIONS = [
+  'Win-Back Cadence — High Priority',
+  'Re-engagement — 7-day Sprint',
+  'Executive Win-Back',
+];
+
+interface Props {
+  open: boolean;
+  context: SignalContext | null;
+  onClose: () => void;
+}
+
+export default function WinBackModal({ open, context, onClose }: Props) {
+  const [selected, setSelected] = useState<Record<string, boolean>>({});
+  const [cadence, setCadence] = useState(CADENCE_OPTIONS[0]);
+  const [assigned, setAssigned] = useState(false);
+
+  // Reset selection/state whenever a new signal is opened.
+  useEffect(() => {
+    if (context) {
+      setSelected(Object.fromEntries(context.contacts.map((c) => [c.id, true])));
+      setCadence(CADENCE_OPTIONS[0]);
+      setAssigned(false);
+    }
+  }, [context]);
+
+  if (!open || !context) return null;
+
+  const count = context.contacts.filter((c) => selected[c.id]).length;
+  const otherFields = context.fields.filter(
+    (f) => !/cadence|account|company|organization/i.test(f.name),
+  );
+
+  return (
+    <div className="winback-overlay" onClick={onClose}>
+      <div className="winback-modal" onClick={(e) => e.stopPropagation()}>
+        <div className="winback-header">
+          <div className="winback-title">
+            <Zap size={18} />
+            <span>Re-engage Cadence</span>
+          </div>
+          <button className="winback-close" onClick={onClose} aria-label="Close">
+            <X size={18} />
+          </button>
+        </div>
+
+        {assigned ? (
+          <div className="winback-success">
+            <div className="winback-success-icon">
+              <Check size={30} />
+            </div>
+            <h3>{count} contact{count === 1 ? '' : 's'} enrolled</h3>
+            <p>
+              {context.account ? <strong>{context.account}</strong> : 'These contacts'}{' '}
+              {count === 1 ? 'has' : 'have'} been added to <strong>{cadence}</strong>.
+            </p>
+            <button className="winback-btn-primary" onClick={onClose}>
+              Done
+            </button>
+          </div>
+        ) : (
+          <>
+            <div className="winback-body">
+              <p className="winback-lede">
+                Assign these contacts to a high-priority Win-Back Cadence to re-engage
+                this account in one click.
+              </p>
+
+              <div className="winback-context">
+                {context.cadence && (
+                  <div className="winback-ctx-item">
+                    <span className="winback-ctx-label">Source cadence</span>
+                    <span className="winback-ctx-value">{context.cadence}</span>
+                  </div>
+                )}
+                {context.account && (
+                  <div className="winback-ctx-item">
+                    <span className="winback-ctx-label">Account</span>
+                    <span className="winback-ctx-value">{context.account}</span>
+                  </div>
+                )}
+              </div>
+
+              {otherFields.length > 0 && (
+                <div className="winback-fields">
+                  {otherFields.map((f) => (
+                    <span className="winback-chip" key={f.name}>
+                      <span className="winback-chip-k">{f.name}</span>
+                      <span className="winback-chip-v">{f.value}</span>
+                    </span>
+                  ))}
+                </div>
+              )}
+
+              <div className="winback-section-title">
+                Contacts to enroll ({count})
+              </div>
+              <div className="winback-contacts">
+                {context.contacts.map((c) => (
+                  <label className="winback-contact" key={c.id}>
+                    <input
+                      type="checkbox"
+                      checked={!!selected[c.id]}
+                      onChange={(e) =>
+                        setSelected((s) => ({ ...s, [c.id]: e.target.checked }))
+                      }
+                    />
+                    <span className="winback-avatar">
+                      {c.name.slice(0, 1).toUpperCase()}
+                    </span>
+                    <span className="winback-contact-info">
+                      <span className="winback-contact-name">{c.name}</span>
+                      <span className="winback-contact-meta">
+                        {c.title}
+                        {c.email ? ` · ${c.email}` : ''}
+                      </span>
+                    </span>
+                  </label>
+                ))}
+              </div>
+
+              <div className="winback-field">
+                <label className="winback-label">Assign to cadence</label>
+                <select
+                  className="winback-select"
+                  value={cadence}
+                  onChange={(e) => setCadence(e.target.value)}
+                >
+                  {CADENCE_OPTIONS.map((c) => (
+                    <option key={c}>{c}</option>
+                  ))}
+                </select>
+              </div>
+            </div>
+
+            <div className="winback-footer">
+              <button className="winback-btn-secondary" onClick={onClose}>
+                Cancel
+              </button>
+              <button
+                className="winback-btn-primary"
+                disabled={count === 0}
+                onClick={() => setAssigned(true)}
+              >
+                <RefreshCw size={16} />
+                Assign to Win-Back Cadence
+              </button>
+            </div>
+          </>
+        )}
+      </div>
+    </div>
+  );
+}

demoprep_app/portal/template-tse/src/config.ts

@@ -0,0 +1,509 @@
+import { CONTENT } from './content';
+// ---------------------------------------------------------------------------
+// Central configuration for the Northwind Analytics mockup.
+// IDs + ThoughtSpot styling are sourced from northwind_fin.json so that every
+// embedded ThoughtSpot component shares the exact same look & feel.
+// ---------------------------------------------------------------------------
+
+/** ThoughtSpot cluster (from appConfig.thoughtspotUrl) */
+export const THOUGHTSPOT_HOST = 'https://YOUR-CLUSTER.thoughtspot.cloud';
+
+/** The Analytics liveboard embedded in Tab 2 (Analytics). */
+export const ANALYTICS_LIVEBOARD_ID = 'REPLACE_MAIN_LIVEBOARD_GUID';
+
+/** Liveboard embedded inline by "View insights" on a process, filtered by the
+ *  process name (no vizId — the whole liveboard is shown, stripped of chrome). */
+export const INLINE_INSIGHTS_LIVEBOARD_ID =
+  'REPLACE_INLINE_LIVEBOARD_GUID';
+
+/** Worksheet/model the Processes tab queries for process names + metrics. */
+export const CADENCE_WORKSHEET_ID = 'REPLACE_MODEL_GUID';
+
+/**
+ * Capacity tab embeds a single VISUALIZATION (LiveboardEmbed + liveboardId +
+ * vizId) with a code-based custom action.
+ */
+export const CAPACITY_LIVEBOARD_ID = 'REPLACE_ACTION_LIVEBOARD_GUID';
+export const CAPACITY_VIZ_ID = 'REPLACE_ACTION_VIZ_GUID';
+
+/** Code-based custom action ("Request Bid") on the Capacity viz context menu. */
+export const REQUEST_BID_ACTION_ID = 'request-bid';
+export const REQUEST_BID_ACTION_NAME = 'Request Bid';
+
+/**
+ * Candidate column names for "cadence" on the worksheet. The first one that
+ * returns data is used both to list cadence names and as the runtime-filter
+ * column name on the inline liveboard.
+ */
+export const CADENCE_COLUMN_CANDIDATES = [
+  'Dim Carrier Name',
+  'Carrier Name',
+  'Carrier',
+  'Dim Carrier',
+];
+
+/**
+ * Candidate column names for the per-cadence detail metrics shown next to each
+ * cadence in the Inline Insights UI. The first candidate that returns data for
+ * each metric is used; metrics with no match are simply omitted.
+ */
+export const CADENCE_DETAIL_COLUMNS: Record<
+  'loadsDelivered' | 'loadsOnTime' | 'carrierSize',
+  { label: string; format: 'text' | 'date' | 'number' | 'currency'; candidates: string[] }
+> = {
+  loadsDelivered: {
+    label: 'Loads Delivered',
+    format: 'number',
+    candidates: ['Loads Delivered', 'Delivered Loads', 'Total Loads Delivered', 'Loads'],
+  },
+  loadsOnTime: {
+    label: 'Loads On Time',
+    format: 'number',
+    candidates: ['Loads On Time', 'On-Time Loads', 'On Time Loads', 'On Time Deliveries'],
+  },
+  carrierSize: {
+    label: 'Carrier Size',
+    format: 'text',
+    candidates: ['Carrier Size', 'Fleet Size', 'Size', 'Carrier Tier'],
+  },
+};
+
+/**
+ * Worksheet / model used everywhere Spotter runs — the "Northwind AI" Spotter
+ * modal (Analytics tab), the Ask Northwind Spotter embed, and the host chatbot.
+ */
+export const WORKSHEET_ID = 'REPLACE_MODEL_GUID';
+
+/** Carriers list is not org-scoped for Northwind. */
+export const Northwind_ORG_ID = '';
+export const ORG_COLUMN_CANDIDATES: string[] = [];
+
+/** Analytics-tab filters — Region + Freight corridor (multiselect), both empty
+ *  by default (= no runtime filter / show all). Load Date is a date-range filter
+ *  driven by DATE_COLUMN below. Candidate column names, first match wins. */
+export const REGION_COLUMN_CANDIDATES = [
+  'Region',
+  'Dim Region',
+  'Sales Region',
+  'Geo Region',
+];
+export const CORRIDOR_COLUMN_CANDIDATES = [
+  'Freight Corridor',
+  'Corridor',
+  'Lane',
+  'Freight Lane',
+];
+
+// ---------------------------------------------------------------------------
+// ThoughtSpot embed styling — taken verbatim from
+// northwind_fin.json > stylingConfig.embeddedContent
+// ---------------------------------------------------------------------------
+
+/** CSS variables applied to every embedded ThoughtSpot surface. */
+// Embed font — the family the ThoughtSpot iframes request. Overwritten by the
+// codemod from spec.appFont; pair it with TS_FONT_URL (below), which loads that
+// font INTO the iframe (a Google-Fonts URL, or a self-hosted /embed-font.css).
+export const EMBED_FONT_FAMILY = '"Inter", system-ui, "Segoe UI", sans-serif';
+
+export const TS_CSS_VARIABLES: Record<string, string> = {
+  '--ts-var-root-background': '#FCFCFD',
+  '--ts-var-root-color': '#182338',
+  '--ts-var-root-font-family': EMBED_FONT_FAMILY,
+  '--ts-var-application-color': '#26324e',
+  '--ts-var-nav-background': '#FFFFFF',
+  '--ts-var-nav-color': '#26324e',
+  '--ts-var-search-data-button-background': '#4F5BD5',
+  '--ts-var-search-data-button-font-color': '#FFFFFF',
+  '--ts-var-search-bar-background': '#F5F6F8',
+  '--ts-var-search-bar-text-font-color': '#26324e',
+  '--ts-var-button-border-radius': '10px',
+  '--ts-var-button--icon-border-radius': '8px',
+  '--ts-var-button--primary-background': '#4F5BD5',
+  '--ts-var-button--primary-color': '#FFFFFF',
+  '--ts-var-button--primary--hover-background': '#3F49B8',
+  '--ts-var-button--primary--active-background': '#333C99',
+  '--ts-var-button--secondary-background': '#EEF0F4',
+  '--ts-var-button--secondary-color': '#26324e',
+  // See dark-theme note: the SDK reads the typo'd "--hovers-background".
+  '--ts-var-button--secondary--hover-background': '#E3E6EC',
+  '--ts-var-button--secondary--hovers-background': '#E3E6EC',
+  '--ts-var-button--secondary--active-background': '#D7DBE3',
+  '--ts-var-viz-title-color': '#182338',
+  '--ts-var-viz-title-font-family': EMBED_FONT_FAMILY,
+  '--ts-var-viz-description-color': '#3a4a6b',
+  '--ts-var-viz-border-radius': '18px',
+  '--ts-var-viz-box-shadow': '0 1px 2px rgba(7, 59, 58, 0.06), 0 16px 44px rgba(7, 59, 58, 0.16)',
+  '--ts-var-viz-background': '#FFFFFF',
+  '--ts-var-chip-border-radius': '999px',
+  '--ts-var-chip-background': '#EEF0F5',
+  '--ts-var-chip-color': '#26324e',
+  '--ts-var-chip--hover-background': '#E3E6EC',
+  '--ts-var-chip--hover-color': '#182338',
+  '--ts-var-chip--active-background': '#4F5BD5',
+  '--ts-var-chip--active-color': '#FFFFFF',
+  '--ts-var-menu-background': '#FFFFFF',
+  '--ts-var-menu-color': '#26324e',
+  '--ts-var-menu--hover-background': '#F1F2F6',
+  '--ts-var-dialog-body-background': '#FFFFFF',
+  '--ts-var-dialog-body-color': '#26324e',
+  '--ts-var-dialog-header-background': '#FFFFFF',
+  '--ts-var-dialog-header-color': '#182338',
+  '--ts-var-list-hover-background': '#F1F2F6',
+  '--ts-var-list-selected-background': '#E8EAF1',
+  '--ts-var-liveboard-layout-background': '#FCFCFD',
+  '--ts-var-liveboard-header-background': '#FFFFFF',
+  '--ts-var-liveboard-header-font-color': '#182338',
+  '--ts-var-liveboard-tile-background': '#FFFFFF',
+  '--ts-var-liveboard-tile-border-color': '#E2E4EC',
+  '--ts-var-liveboard-tile-border-radius': '18px',
+  '--ts-var-liveboard-tile-padding': '12px',
+  '--ts-var-liveboard-tile-table-header-background': '#F5F6F9',
+  '--ts-var-liveboard-tab-active-border-color': '#4F5BD5',
+  '--ts-var-liveboard-tab-hover-color': '#4F5BD5',
+  '--ts-var-liveboard-header-action-button-background': '#EEF0F5',
+  '--ts-var-liveboard-header-action-button-font-color': '#26324e',
+  // NOTE: despite "-color", these are the button's hover/active BACKGROUND.
+  // Label is dark here, so keep the hover background LIGHT for contrast.
+  '--ts-var-liveboard-header-action-button-hover-color': '#E3E6EC',
+  '--ts-var-liveboard-header-action-button-active-color': '#D7DBE3',
+  '--ts-var-parameter-chip-background': '#EEF0F5',
+  '--ts-var-parameter-chip-text-color': '#26324e',
+  '--ts-var-parameter-chip-hover-background': '#E3E6EC',
+  '--ts-var-parameter-chip-hover-text-color': '#182338',
+  '--ts-var-parameter-chip-active-background': '#4F5BD5',
+  '--ts-var-parameter-chip-active-text-color': '#FFFFFF',
+  '--ts-var-axis-title-color': '#3a4a6b',
+  '--ts-var-axis-data-label-color': '#5A6A7E',
+  '--ts-var-kpi-hero-color': '#182338',
+  '--ts-var-kpi-comparison-color': '#3a4a6b',
+  '--ts-var-kpi-positive-change-color': '#1E8E5A',
+  '--ts-var-kpi-negative-change-color': '#C84B3A',
+};
+
+export type ThemeName = 'light' | 'dark';
+
+/**
+ * Dark counterpart to TS_CSS_VARIABLES — a deep-evergreen dark theme that keeps
+ * Northwind's green/coral accents. Passed per-embed for the active theme so the
+ * embedded Liveboard / Search / Spotter track the host app's light/dark toggle.
+ *   CSS variables: https://developers.thoughtspot.com/docs/css-variables-reference
+ */
+export const TS_VARS_DARK: Record<string, string> = {
+  '--ts-var-root-background': '#0b0d10',
+  '--ts-var-root-color': '#e7ecf2',
+  '--ts-var-root-secondary-color': '#93a0b4',
+  '--ts-var-root-font-family': EMBED_FONT_FAMILY,
+  '--ts-var-application-color': '#e7ecf2',
+  '--ts-var-nav-background': '#0b0d10',
+  '--ts-var-nav-color': '#e7ecf2',
+  '--ts-var-search-data-button-background': '#4F5BD5',
+  '--ts-var-search-data-button-font-color': '#FFFFFF',
+  '--ts-var-search-bar-background': '#14181f',
+  '--ts-var-search-bar-text-font-color': '#e7ecf2',
+  '--ts-var-search-auto-complete-background': '#14181f',
+  '--ts-var-search-auto-complete-font-color': '#e7ecf2',
+  '--ts-var-button-border-radius': '10px',
+  '--ts-var-button--icon-border-radius': '8px',
+  // Deeper green + white text so labels stay readable (incl. on hover).
+  '--ts-var-button--primary-background': '#4F5BD5',
+  '--ts-var-button--primary-color': '#ffffff',
+  '--ts-var-button--primary--hover-background': '#3F49B8',
+  '--ts-var-button--primary--active-background': '#333C99',
+  '--ts-var-button--secondary-background': '#1a212b',
+  '--ts-var-button--secondary-color': '#e7ecf2',
+  // NOTE: the SDK's *effective* var name is the typo'd "--hovers-background"
+  // (default #aac2f8, a light blue) — the clean "--hover-background" is only in
+  // the docs. Set both so the label stays readable on hover.
+  '--ts-var-button--secondary--hover-background': '#4F5BD5',
+  '--ts-var-button--secondary--hovers-background': '#4F5BD5',
+  '--ts-var-button--secondary--active-background': '#3F49B8',
+  '--ts-var-button--tertiary-background': 'transparent',
+  '--ts-var-button--tertiary-color': '#c3cbd6',
+  '--ts-var-button--tertiary--hover-background': '#1a212b',
+  '--ts-var-viz-title-color': '#e7ecf2',
+  '--ts-var-viz-title-font-family': EMBED_FONT_FAMILY,
+  '--ts-var-viz-description-color': '#93a0b4',
+  '--ts-var-viz-border-radius': '18px',
+  '--ts-var-viz-box-shadow':
+    '0 1px 0 rgba(255,255,255,0.05) inset, 0 14px 44px rgba(0,0,0,0.55), 0 0 0 1px rgba(45,164,191,0.14)',
+  '--ts-var-viz-background': '#14181f',
+  '--ts-var-viz-legend-hover-background': '#1a212b',
+  '--ts-var-chip-border-radius': '999px',
+  '--ts-var-chip-background': '#1a212b',
+  '--ts-var-chip-color': '#c3cbd6',
+  '--ts-var-chip--hover-background': '#232a34',
+  '--ts-var-chip--hover-color': '#e7ecf2',
+  '--ts-var-chip--active-background': '#4F5BD5',
+  '--ts-var-chip--active-color': '#FFFFFF',
+  '--ts-var-menu-background': '#14181f',
+  '--ts-var-menu-color': '#e7ecf2',
+  '--ts-var-menu--hover-background': '#1a212b',
+  '--ts-var-menu-selected-text-color': '#4F5BD5',
+  '--ts-var-menu-separator-background': '#232a34',
+  '--ts-var-dialog-body-background': '#14181f',
+  '--ts-var-dialog-body-color': '#e7ecf2',
+  '--ts-var-dialog-header-background': '#14181f',
+  '--ts-var-dialog-header-color': '#e7ecf2',
+  '--ts-var-dialog-footer-background': '#14181f',
+  '--ts-var-list-hover-background': '#1a212b',
+  '--ts-var-list-selected-background': '#232a34',
+  '--ts-var-liveboard-layout-background': '#0b0d10',
+  '--ts-var-liveboard-header-background': '#0b0d10',
+  '--ts-var-liveboard-header-font-color': '#e7ecf2',
+  '--ts-var-liveboard-edit-bar-background': '#14181f',
+  '--ts-var-liveboard-cross-filter-layout-background': '#14181f',
+  '--ts-var-liveboard-tile-background': '#14181f',
+  '--ts-var-liveboard-tile-border-color': '#232a34',
+  '--ts-var-liveboard-tile-border-radius': '18px',
+  '--ts-var-liveboard-tile-padding': '12px',
+  '--ts-var-liveboard-tile-table-header-background': '#1a212b',
+  '--ts-var-liveboard-tab-active-border-color': '#4F5BD5',
+  '--ts-var-liveboard-tab-hover-color': '#4F5BD5',
+  '--ts-var-liveboard-header-action-button-background': '#1a212b',
+  '--ts-var-liveboard-header-action-button-font-color': '#e7ecf2',
+  // NOTE: despite "-color", these are the button's hover/active BACKGROUND.
+  // Label is light here, so keep the hover background DARK green for contrast.
+  '--ts-var-liveboard-header-action-button-hover-color': '#4F5BD5',
+  '--ts-var-liveboard-header-action-button-active-color': '#3F49B8',
+  // Masterpieces grouping/styling (active because isLiveboardMasterpiecesEnabled)
+  '--ts-var-liveboard-group-background': '#0b0d10',
+  '--ts-var-liveboard-group-title-font-color': '#e7ecf2',
+  '--ts-var-liveboard-group-border-color': '#232a34',
+  '--ts-var-liveboard-group-description-font-color': '#93a0b4',
+  '--ts-var-liveboard-group-tile-title-font-color': '#e7ecf2',
+  '--ts-var-liveboard-group-tile-description-font-color': '#93a0b4',
+  '--ts-var-parameter-chip-background': '#1a212b',
+  '--ts-var-parameter-chip-text-color': '#e7ecf2',
+  '--ts-var-parameter-chip-hover-background': '#232a34',
+  '--ts-var-parameter-chip-hover-text-color': '#ffffff',
+  '--ts-var-parameter-chip-active-background': '#4F5BD5',
+  '--ts-var-parameter-chip-active-text-color': '#FFFFFF',
+  '--ts-var-axis-title-color': '#93a0b4',
+  '--ts-var-axis-data-label-color': '#93a0b4',
+  '--ts-var-answer-data-panel-background-color': '#0b0d10',
+  '--ts-var-answer-edit-panel-background-color': '#0b0d10',
+  '--ts-var-kpi-hero-color': '#e7ecf2',
+  '--ts-var-kpi-comparison-color': '#93a0b4',
+  '--ts-var-kpi-positive-change-color': '#35c98a',
+  '--ts-var-kpi-negative-change-color': '#ff7a6b',
+  // Spotter / Sage conversational surfaces
+  '--ts-var-spotter-input-background': '#14181f',
+  '--ts-var-spotter-prompt-background': '#1a212b',
+  '--ts-var-sage-embed-background-color': '#0b0d10',
+  '--ts-var-sage-bar-header-background-color': '#0b0d10',
+  '--ts-var-sage-search-box-background-color': '#14181f',
+  '--ts-var-sage-search-box-font-color': '#e7ecf2',
+  '--ts-var-source-selector-background-color': '#14181f',
+  '--ts-var-sage-seed-questions-background': '#14181f',
+  '--ts-var-sage-seed-questions-font-color': '#c3cbd6',
+  '--ts-var-sage-seed-questions-hover-background': '#1a212b',
+};
+
+/** Return the CSS-variable set for the active theme. */
+export function tsVarsFor(theme: ThemeName): Record<string, string> {
+  return theme === 'dark' ? TS_VARS_DARK : TS_CSS_VARIABLES;
+}
+
+/**
+ * Dark-only color overrides for surfaces with no dedicated CSS variable (answer
+ * cards, floating chrome), plus the KPI headline in the Northwind green. Applied
+ * via rules_UNSTABLE only in dark mode.
+ *   Docs: https://developers.thoughtspot.com/docs/css-rules
+ */
+const DARK_SURFACE = '#14181f';
+const DARK_INK = '#e7ecf2';
+const surfaceRule = { 'background-color': `${DARK_SURFACE} !important`, color: `${DARK_INK} !important` };
+export const TS_RULES_DARK: Record<string, Record<string, string>> = {
+  // Text inputs / textareas / field wrappers.
+  [[
+    'input:not([type="checkbox"]):not([type="radio"])',
+    'textarea',
+    '[class*="input" i]',
+    '[class*="field" i]',
+    '[class*="searchbox" i]',
+  ].join(',')]: { 'background-color': '#0b1f19 !important', color: `${DARK_INK} !important` },
+  // Segmented controls / toggles / switchers (e.g. Quick Search | Deep Analysis).
+  [[
+    '[class*="toggle" i]',
+    '[class*="switch" i]',
+    '[class*="switcher" i]',
+    '[class*="segment" i]',
+  ].join(',')]: surfaceRule,
+  // Dropdown / select header (dimension & filter pickers).
+  '[data-testid="select-dropdown-header"]': { 'background-color': '#0b1f19 !important', color: `${DARK_INK} !important` },
+  // Answer / conversation / Spotter / Sage surfaces + the conversations sidebar.
+  [[
+    '[class*="answer" i]',
+    '[class*="conversation" i]',
+    '[class*="message" i]',
+    '[class*="spotter" i]',
+    '[class*="sage" i]',
+    '[class*="sidebar" i]',
+    '[class*="history" i]',
+  ].join(',')]: surfaceRule,
+  // Floating chrome — panels / modals / menus / footers.
+  [[
+    '[class*="modal" i]',
+    '[class*="dialog" i]',
+    '[class*="popover" i]',
+    '[class*="flyout" i]',
+    '[class*="drawer" i]',
+    '[class*="data-panel" i]',
+    '[class*="dataPanel" i]',
+    '[class*="footer" i]',
+  ].join(',')]: surfaceRule,
+};
+
+/**
+ * Welcome message for the host-owned Northwind AI chatbot
+ * (from northwind_fin.json > appConfig.chatbot.welcomeMessage, lightly adapted).
+ */
+export const CHATBOT_WELCOME = CONTENT.chatbot.welcome;
+
+/** Greeting + data model used when the chatbot is opened on the Carriers tab. */
+export const CHATBOT_CADENCES_WELCOME =
+  'What would you like to know about the carriers shown here?';
+
+/** Spotter icon sprite (magician/spotter icon) from northwind_fin.json. */
+export const TS_ICON_SPRITE_URL =
+  'https://cdn.jsdelivr.net/gh/thoughtspot/tse-demo-builders-pre-built/icons/spotter/generic-02.svg';
+
+/**
+ * String customizations — replace "Spotter" with "Northwind AI" everywhere,
+ * plus the landing-page description override (from
+ * northwind_fin.json > stylingConfig.embeddedContent.strings / stringIDs).
+ */
+export const TS_STRINGS: Record<string, string> = {
+  Spotter: CONTENT.aiName,
+};
+
+export const TS_STRING_IDS: Record<string, string> = {
+  'convAssist.landingpage.description2': 'Ask a question about sales.',
+};
+
+// ---------------------------------------------------------------------------
+// Embed flags — from northwind_fin.json > stylingConfig.embedFlags
+// ---------------------------------------------------------------------------
+
+export const LIVEBOARD_EMBED_FLAGS = {
+  enable2ColumnLayout: true,
+  isLiveboardStylingAndGroupingEnabled: true,
+  // Masterpiece layout — richer KPI/headline tile styling & grouping.
+  // https://developers.thoughtspot.com/docs/Interface_LiveboardViewConfig#_isliveboardmasterpiecesenabled
+  isLiveboardMasterpiecesEnabled: true,
+};
+
+export const SPOTTER_EMBED_FLAGS = {
+  updatedSpotterChatPrompt: true,
+  spotterSidebarConfig: {
+    enablePastConversationsSidebar: true,
+    spotterSidebarTitle: 'My Conversations',
+    // Start collapsed so the answer canvas gets full width when Spotter opens.
+    spotterSidebarDefaultExpanded: false,
+  },
+};
+
+/**
+ * Extra customCSS rules injected into every embed so ThoughtSpot's own buttons
+ * carry the Northwind blue↔green gradient and the KPI (headline) tiles pop with the
+ * brand gradient. ThoughtSpot class names are version-dependent — these use
+ * broad candidate selectors; if a surface isn't picked up, Inspect the iframe
+ * and tighten the selector to the real class name.
+ */
+export const Northwind_EMBED_RULES: Record<string, Record<string, string>> = {
+  // No gradient on KPI tiles / viz titles by default — embeds use ThoughtSpot's
+  // native styling, which already follows the theme via the CSS variables above.
+  // Only brand the AI / Spotter header button (an off-brand green pill on some
+  // clusters) with a SOLID color. To opt into brand gradients on KPIs/titles,
+  // add background linear-gradient rules here (buttons must stay solid).
+  '[class*="aiHighlight" i], [class*="ai-highlight" i], [class*="AIHighlights" i], [data-testid*="AIHighlight" i], [class*="spotterLaunch" i], [class*="spotter-button" i], [class*="spotter" i][class*="btn" i], [class*="spotter" i][class*="button" i], [class*="spotter" i][class*="pill" i], [class*="spotter" i][class*="nudge" i], [class*="sage-button" i], [class*="sage" i][class*="btn" i], [class*="sage" i][class*="button" i], [class*="sage" i][class*="pill" i], [class*="askSage" i], [class*="ask-sage" i], [class*="nudge" i]':
+    {
+      background: '#4F5BD5 !important',
+      'background-image': 'none !important',
+      color: '#ffffff !important',
+      'border-color': 'transparent !important',
+    },
+};
+
+/** Google-Fonts stylesheet loaded INTO embed iframes so embedded ThoughtSpot
+ * surfaces render in the same faces as the host UI (Plus Jakarta Sans for text,
+ * Fraunces for titles) — iframes can't see the host page's fonts. customCSSUrl.
+ */
+export const TS_FONT_URL =
+  'https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap';
+
+/** "Watch video" link shown in the Northwind AI pane. */
+export const Northwind_VIDEO_URL = 'https://www.northwind.com/platform/conversations';
+
+/** Sample questions shown on the custom Northwind AI landing pane. */
+export const Northwind_SAMPLE_QUESTIONS = [...CONTENT.sampleQuestions];
+
+/** Trial / upgrade prompt shown in the Northwind AI pane after the 2nd query. */
+export const Northwind_TRIAL_QUESTIONS = 20;
+export const TRIAL_TRIGGER = CONTENT.trialQuestionTrigger;
+export const Northwind_UPGRADE_URL = CONTENT.website + '/pricing';
+
+/**
+ * Hide the Spotter embed's own composer / input bar on the Northwind AI screen
+ * — questions are driven from the host-side pane on the right, so the in-embed
+ * "Enter your question" box (and its Quick/Deep toggle) is redundant. Injected
+ * via rules_UNSTABLE. Class names are version-fragile; verify via Inspect if a
+ * release renames them.
+ */
+const HIDE = { display: 'none !important' };
+export const HIDE_SPOTTER_INPUT_RULES: Record<string, Record<string, string>> = {
+  '[class*="composer" i]': HIDE,
+  '[class*="promptInput" i]': HIDE,
+  '[class*="prompt-input" i]': HIDE,
+  '[class*="chatInput" i]': HIDE,
+  '[class*="chat-input" i]': HIDE,
+  '[class*="conversationInput" i]': HIDE,
+  '[class*="conversation-input" i]': HIDE,
+  '[class*="conversationFooter" i]': HIDE,
+  '[class*="conversation-footer" i]': HIDE,
+  '[class*="bottomBar" i]': HIDE,
+  '[class*="searchInputContainer" i]': HIDE,
+  '[data-testid*="conversation-input" i]': HIDE,
+  '[data-testid*="spotter-input" i]': HIDE,
+};
+
+// ---------------------------------------------------------------------------
+// Host-side filters on the Analytics liveboard.
+//   Hierarchy: OWNER_COLUMN → CADENCE_NAME_COLUMN. Date: DATE_COLUMN.
+// Option lists are fetched from FILTER_SOURCE_ID via the searchdata REST API;
+// selections are pushed to the liveboard as runtime filters. Column names must
+// match the data source's column display names for the runtime filter to bind.
+//   Runtime filters: https://developers.thoughtspot.com/docs/runtime-filters
+// ---------------------------------------------------------------------------
+
+/**
+ * Model queried for the filter option lists (same model the liveboard uses).
+ * Hierarchy: SEGMENT_COLUMN → REP_COLUMN → CADENCE_NAME_COLUMN. Date: DATE_COLUMN.
+ */
+export const FILTER_SOURCE_ID = WORKSHEET_ID;
+export const SEGMENT_COLUMN = 'Rep Segment';
+export const REP_COLUMN = 'Rep Name';
+export const CADENCE_NAME_COLUMN = 'Cadence Name';
+export const DATE_COLUMN = 'Load Date';
+
+/**
+ * Hide the native runtime-filter pills / filter bar ThoughtSpot renders on the
+ * liveboard header — we drive filters from the host UI instead. Injected into
+ * the liveboard iframe via rules_UNSTABLE. Class names are version-fragile;
+ * verify/adjust via right-click → Inspect if a release renames them.
+ *   Docs: https://developers.thoughtspot.com/docs/css-rules
+ */
+const PILL_HIDE = { display: 'none !important' };
+export const HIDE_FILTER_PILL_RULES: Record<string, Record<string, string>> = {
+  '[class*="filterChip" i]': PILL_HIDE,
+  '[class*="filter-chip" i]': PILL_HIDE,
+  '[class*="appliedFilter" i]': PILL_HIDE,
+  '[class*="applied-filter" i]': PILL_HIDE,
+  '[class*="pinboardFilter" i]': PILL_HIDE,
+  '[class*="pinboard-filter" i]': PILL_HIDE,
+  '[class*="filterBar" i]': PILL_HIDE,
+  '[class*="filter-bar" i]': PILL_HIDE,
+  '[data-testid*="filter-chip" i]': PILL_HIDE,
+  '[data-testid*="pinboard-filters" i]': PILL_HIDE,
+};
+

demoprep_app/portal/template-tse/src/content.ts

@@ -0,0 +1,111 @@
+// ---------------------------------------------------------------------------
+// content.ts — the SINGLE source of truth for every human-readable string in
+// the portal. A reskin edits this file (or the codemod `scripts/apply-spec.mjs`
+// writes it from a spec.json); nothing else under src/ hard-codes brand or
+// domain copy. IDs / GUIDs / host / columns / theme live in config.ts.
+//
+// Defaults below are the neutral "Northwind" freight-brokerage template.
+// ---------------------------------------------------------------------------
+
+export interface LoginStat {
+  value: string;
+  label: string;
+}
+
+export const CONTENT = {
+  // ---- identity ----------------------------------------------------------
+  company: 'Northwind',
+  aiName: 'Northwind AI', // shown as the fancy-chat brand + "Ask <aiName>" tab
+  website: 'https://www.northwind.com',
+
+  // ---- login hero --------------------------------------------------------
+  // loginTitleHtml may contain a single <br /> for the line break.
+  loginTitleHtml: 'Where freight<br />meets finance.',
+  loginSubtitle:
+    'Your brokerage command center — carrier capacity, freight corridors, and loads-on-time, with AI-powered answers on live data. Built on Northwind, powered by ThoughtSpot.',
+  loginStats: [
+    { value: '1.4M+', label: 'Loads Analyzed' },
+    { value: '98%', label: 'On-Time Visibility' },
+    { value: 'Real-time', label: 'Capacity Intel' },
+  ] as LoginStat[],
+  loginCardSubtitle: 'Enter any username and password to continue',
+  // {host} is substituted with THOUGHTSPOT_HOST's bare hostname at render time.
+  loginDemoNoteHtml:
+    "<strong>Sign in</strong> with your ThoughtSpot username &amp; password for <code>{host}</code>. This origin must be CORS-allowlisted on that cluster for the embeds to load.",
+
+  // ---- decorative platform nav (product-context words, non-interactive) ---
+  platformNav: ['Broker', 'Shipper', 'Factor'],
+
+  // ---- tab labels --------------------------------------------------------
+  tabs: {
+    myReports: 'My Reports',
+    analytics: 'Analytics',
+    inline: 'Carriers', // the inline-insights list tab
+    action: 'Capacity', // the custom-action tab
+    ask: 'Ask Northwind AI',
+    spotter: 'Spotter',
+  },
+
+  // ---- page subtitles / states ------------------------------------------
+  analyticsSubtitle: 'Your live freight & capacity performance dashboard',
+  spotterSubtitle: 'Ask questions of your Northwind data in natural language.',
+  inline: {
+    subtitleReady:
+      'Carrier activity for your organization — select a carrier to view its insights',
+    subtitleIdle: 'Your carriers and their activity records',
+    loading: 'Loading carriers…',
+    errorTitle: "Couldn't load carriers",
+    errorMsg: 'Unable to load carrier data.',
+    moreLabel: 'View insights',
+    hideLabel: 'Hide insights',
+  },
+
+  // ---- analytics host filters -------------------------------------------
+  filters: { primaryLabel: 'Region', secondaryLabel: 'Freight corridor' },
+
+  // ---- Ask-AI (fancy chat) ----------------------------------------------
+  sampleQuestions: [
+    'Loads delivered by corridor',
+    'Top carriers by on-time rate',
+    'Capacity by region',
+    'Loads per carrier over time',
+  ],
+  askWelcome:
+    'Ask me anything about Northwind — freight, carriers, capacity, best practices — or ask about your own data and I will chart it live.',
+  askEmptySub:
+    'Ask any analytical question about your freight, carriers, capacity, and coverage. Answers come straight from your live data with an interactive chart you can drill into.',
+
+  // ---- custom-action tab + modal ----------------------------------------
+  action: {
+    label: 'Request Bid', // the context-menu action name
+    subtitlePrefix: 'Lane capacity & carrier opportunities. Right-click a row and choose',
+    subtitleSuffix: 'to launch a request pre-filled from that row.',
+    modalTitle: 'Request Bid',
+    modalLead: 'Pre-filled from the selected row. Review and submit through Northwind.',
+    submitLabel: 'Submit',
+    successTitle: 'Request submitted',
+  },
+
+  // ---- floating chatbot + fallback brain --------------------------------
+  chatbot: {
+    welcome:
+      "Hi! I'm Northwind AI. Ask me about freight, carriers and capacity, or ask a question about your data.",
+    overview:
+      'Northwind is an AI-powered logistics platform that helps brokerages manage carrier capacity, freight corridors, and on-time delivery, with answers on live data.',
+    greetingExample: 'loads delivered by corridor',
+    examples: ['loads delivered by corridor', 'top carriers by on-time rate'],
+  },
+
+  // ---- monetization ------------------------------------------------------
+  // The paywall fires once, on the Nth Ask-AI question.
+  trialQuestionTrigger: 3,
+  upgradeReason:
+    'Upgrade to Northwind Premium to unlock drill-down, Ask AI and downloads.',
+
+  // Extra domain terms that mark a chat message as a DATA question. Folded into
+  // the chatbot's analytics router alongside the generic "by X" / "top N" / metric
+  // patterns, so a client's own vocabulary routes to Spotter instead of the FAQ.
+  analyticsKeywords: ['loads', 'carriers', 'capacity', 'corridor', 'lane', 'on-time', 'delivered', 'freight'],
+} as const;
+
+export type Content = typeof CONTENT;