Spaces:
Sleeping
Sleeping
Commit History
-
feat: the guidance record fills the half of Track Record that never worked
The panel already asked two questions and answered one. Its right half scored the consensus record; its left half read brief["guidance_history"], where verdict is always None, so it showed "no scoreable guidance found" for every company while the ranges sat unused in metrics one table away. The left half now scores from those ranges and uses the same two devices as the right -- score card and verdict strip -- so the two halves finally read alike: did they keep their own word, did they clear the street's bar. Beneath them sits the one device this dashboard did not have. A verdict strip can say a streak happened; it cannot say the cushion behind it fell from +20.4% to +3.7%. history_table() shows the figures per row, right-aligned and tabular, in its own horizontal scroller so a wide table never makes the page scroll sideways. Two things were removed rather than added. The Beat / In-line / Missed / Pending counters in the guidance bridge read the same empty verdict field and so displa…da30423 -
data: GOOGL Q12023 recovered, 88 of 89 rows now verified
Re-extracted GOOGL and NVDA against the two fixes. GOOGL's April 2023 10-Q was reachable for the first time -- it had been sitting in the filings.files archive the ingestion never opened -- so its row gained SEC provenance and its six core statement lines. All eight acceptance criteria now pass: zero LEGACY_UNVERIFIED, zero rows without a source URL, zero core statement-line gaps. Status is 88 VERIFIED / 1 CHECK_REQUIRED, the remaining row being NVDA Q32024, whose prior year-to-date capex the company never tagged. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014WiLV6MBBYAD5sWnv5Xqc5e2df2dc
-
data: AMD brief generated locally on Opus 5 against the repaired tables
First run over the re-extracted data, and it exercises exactly what the sub-project unlocked. The brief's headline compares the reported quarter to management's own guide -- a sentence that was not computable this morning, when structured guidance was 0 of 89 rows. The Financials header reads VERIFIED where AMD previously showed METRICS HIDDEN over twelve masked rows, and the balance-sheet warnings render off the statement lines that were 56/89 before. data/threads.db appears here for the first time: it is declared in storage/threads_db.py but was never created until a brief run touched it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014WiLV6MBBYAD5sWnv5Xqc5f30fb09 -
data: refreshed MD&A and risk-factor sections from the re-extraction
The re-extraction had each filing in hand anyway, so it rewrote the two section types it fetches. Same 312 rows, current text. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014WiLV6MBBYAD5sWnv5Xqc55f2135c -
data: re-extract 88 periods, guidance from the calls, segments on six tickers
Row status moved from 41 VERIFIED / 15 CHECK_REQUIRED / 33 LEGACY_UNVERIFIED to 87 / 1 / 1. Structured guidance went from 0 rows to 53, each carrying the passage it came from, with zero parse errors across 88 periods. Segment rows went from 275 on one ticker to 1,288 across all six. Every fallback_non_sec warning is gone. Run record in docs/superpowers/plans/2026-09-01-data-foundation-run.md names every residual gap and labels it. Two are defects with follow-ups: GOOGL Q12023 was never reached because EDGAR caps submissions.filings.recent at ~1000 entries and Alphabet's Form 4 volume pushes that array's floor to 2023-06-27, so the April filing lives in the filings.files archive the ingestion never reads; and NVDA Q32024 cannot de-accumulate its prior year-to-date capex. The 84 statement-line absences are filer facts -- Apple reports no separate goodwill line, AMD tags no current contract liability. GOOGL and TSLA returned 28 rows of NO_GUIDANCE_ISSUED, which is the correct reading of…1f17471 -
feat: decompose the EPS headline and read the statements behind it
Two blocks on The Brief, both answering the question the headline raises and never settles: where did the number come from. analytics/eps_bridge.py partitions the year-on-year change in earnings per share into revenue, operating margin, interest, tax rate, share count and a residual carried openly as "other, non-operating". The identity holds by construction and is re-verified per computation; a bridge that fails its own checksum does not render. Five refusals keep it honest -- no prior-year comparable, a period without SEC lineage, a missing input, a diluted share count that moved more than 25% (a split, not dilution), and a residual larger than half the move. That last gate caught a real defect: Alphabet's quarterly EPS rows hold cumulative figures, so the panel withdraws for GOOGL rather than explain 8% of a move and invent the rest. analytics/statement_signals.py adds three figures that are always worth seeing -- incremental margin, two-year stack, cash conversion -- and a list o…bb3ad6b -
data: segment tables, the Readthrough brief and the run cost log
metrics.db gains the segments table: 396 per-product and per-region result lines backfilled from twelve AAPL filings' XBRL instances, which is what the new breakdown and impact panels read. briefs.db carries the regenerated AAPL brief — the press-style article with its scorecard, section headings and the disclosed margin bridge. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>dfe80e8
-
fix: key backfill snapshot_id on fiscalDateEnding, not the derived label
snapshot_id was derived from target_period, which is DERIVED, not observed: it is Alpha Vantage's fiscalDateEnding date string until a matching quarterly row lands in metrics.db, then it flips to "Q{n}{YYYY}". ingest.py adds such rows every quarter, so the next re-run of this backfill would have minted new ids for those observations and appended them a second time under the other label -- exactly the duplication Ruling 13 had to clean by hand, in an archive whose only asset is being trustworthy. Per Ruling 18 the fix is structural rather than a docstring caveat: identity now comes from fiscalDateEnding, an observed provider fact that never moves, which leaves target_period free to improve without forking the record. The module docstring's idempotency claim is rewritten to match, including the consequence -- a re-run after an ingest writes nothing, so an archived label stays as first written and a consumer wanting the current label joins on period_end_date. The backfill half is regene…4d76839 -
fix: restrict Alpha Vantage backfill matching to quarterly rows only
Ruling 11's join to metrics.report_date, unrestricted by row type, matched 18 quarterly Alpha Vantage estimates to annual (FY{YYYY}) rows whenever a fiscal Q4's period end fell within 7 days of the fiscal year end -- a category error labelling a quarterly estimate with an annual period. Restrict the join to rows whose period matches ^Q[1-4]\d{4}$; a fiscal Q4 has no 10-Q row in this repo's scheme and correctly falls back to the date-string identifier. Also regenerates the archive's backfill portion: unlike live_snapshot records (irrecoverable once the source cache rolls), backfill_at_report records can always be re-fetched, so the append-only protection that matters for the former doesn't apply to the latter. Stripped all backfill_at_report records (byte-for-byte preserving the 12 live_snapshot records, verified against a backup before and after) and re-ran the corrected backfill from scratch, avoiding the duplicate-observation problem the unrestricted join introduced. Result: 38 mat…4ae956f -
fix: join Alpha Vantage backfill on fiscalDateEnding, not reportedDate
metrics.report_date holds a fiscal period-end date despite its name (it comes from SEC's filings.recent.reportDate), not an announcement date. Alpha Vantage's reportedDate is an announcement date ~30 days later, so joining on it matched 0 of 617 real records. Switch the join key to Alpha Vantage's fiscalDateEnding (also a period-end date, 0-5 days of drift measured across the six covered tickers) and widen the match window from 3 to 7 days. target_period_source string is left unchanged for consistency with records already written under it. Re-running the backfill appends 56 newly-matched records (quarter and FY periods) alongside the existing date-string fallback records for the same underlying observations — expected append-only behavior, not deduplicated here per controller instruction.5a03a83 -
feat: backfill historical consensus EPS from Alpha Vantage
One-off import of ~15-20 years of quarterly EPS estimates per covered ticker into the append-only consensus archive, marked capture_kind="backfill_at_report" so aligned_consensus refuses them (no capture timestamp means none can be proven pre-release). Each Alpha Vantage quarter is labelled by joining its reportedDate against metrics.report_date (exact, else within 3 days) and borrowing that row's period verbatim; otherwise it falls back to Alpha Vantage's own fiscalDateEnding string as the identifier, which aligned_consensus can never resolve to a reported row but which still preserves the (estimate, actual, report date) triple for the calibration backtest. Adds unit tests for the matching/fallback logic against a tmp_path metrics DB (never the network).5b40465 -
fix: derive target_period for the daily consensus capture
Controller ruling on the gap found while implementing Task 4: the daily job called fetch_analyst_estimates(ticker) with no target_period, so every captured payload carried target_period=None and record_consensus_snapshot's guard silently dropped every record — the archive stayed permanently empty. - analytics/consensus.py: next_unreported_period(ticker) derives the quarter after the most recently reported one from metrics.period labels (Q{n}{YYYY} / FY{YYYY}), taking the max on parsed (year, quarter) rather than lexicographically, since "Q12027" < "Q42026" as strings. Returns None when nothing parses. - scripts/snapshot_consensus.py: calls next_unreported_period first and skips (rather than fetches with a None period) when it returns None. Also warns, per ticker, when the latest reported filing is more than 120 days old (dashboard/provenance.py::filing_staleness's threshold) since the derived period is then probably wrong. - ingestion/analyst.py: every snapshot now car…6bdcc4d
-
chore: backfill the last seven missing earnings transcripts
Completes transcript coverage for every ingested ticker. All seven were recorded as unavailable when first requested; the provider has since backfilled its archive, and the marker correctly stopped delta ingestion from asking again. AMD Q2 2023, Q1 2023, Q3 2022, Q2 2022 NVDA Q3 2023, Q2 2023 MSFT FY 2026 Microsoft's is the one with an immediate effect: FY 2026 is its most recent call and was missing outright, so it now enters the four-call window the Q&A analysis reads. The rest are 2022-2023 and sit outside that window — their value is depth of archive for semantic search and for the multi-quarter drift detectors. scripts/backfill_transcripts.py now reports zero gaps across all six tickers. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>ac07e2d -
chore: warm every runtime cache shipped to the Space
Produced by scripts/warm_runtime_cache.py across all six covered tickers, so production serves a complete snapshot rather than whatever a developer session happened to touch. Newly warm for every ticker: post-earnings price reaction, which the Space cannot compute itself, and the balance-sheet and filing-event sets for MSFT, TSLA and AMD, which previously shipped cold. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>1ec90e0 -
chore: refresh data snapshot shipped to the Space
Backfills four earnings-call transcripts that were missing for reasons a delta ingest could not repair on its own: GOOGL Q1 2023 — the 10-Q had aged out of EDGAR's recent-filings index, so ingest.py never revisited the period and never asked for its transcript. TSLA Q2 2023, Q1 2023, Q3 2022 — recorded as unavailable when first requested. The provider has since backfilled its own archive, but the marker correctly stopped delta runs from asking again. All three requested tickers now have a transcript for every stored period. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>1c943d0 -
chore: refresh data snapshot shipped to the Space
Adds data/company_profiles.db, which storage/company_profiles.py reads and which had never been committed, plus refreshed chroma, metrics, sections and briefs snapshots. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>a50f9ca
-
chore: purge stale broken NVDA brief so next Generate rebuilds it
The only persisted brief (NVDA, 2026-07-16) had verified_count=0 from a run predating the evidence_ref enforcement rework, and was auto-loaded by app.py's sidebar Coverage panel before any user could trigger a fresh run. Deleted from briefs, brief_snapshots, and brief_translations. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>5666f5d
-
feat: per-fact evidence_ref enforcement + AAPL delta re-ingestion
Require a full evidence_ref (evidence_id/source/content_hash/document_id/ chunk_id/source_url/as_of) on every sourced fact in the brief and company profile prompts, so verify_fact can bind each claim to its exact supporting record instead of a reused/memorable one. Loosen _annual_trends to accept CHECK_REQUIRED rows whose only warnings are non-SEC fallback metrics, nulling just the affected fields instead of dropping the whole period. Improve the segment/geography section regex in edgar.py to also match Item 15-style notes when Item 8 has no match. Re-run delta ingestion for AAPL (new 10-Q Q3 2026 filed 2026-07-31) so the committed data/ snapshot — which the Viney/amplegest HF Space serves as-is via Git LFS — reflects it; AMD/GOOGL/MSFT/NVDA/TSLA had no new filings. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>bd4975b
-
feat: multi-provider LLM support, prominent chat, design pass, and new analytics
- Add agent/llm.py provider factory (Anthropic/OpenAI) with a sidebar model picker so visitors can paste their own API key instead of relying solely on the server's; thread the resulting config through generation, chat, and translation. - Promote Chat to primary nav (was buried under a secondary "Evidence tool" heading) with suggested-question chips and a "sources searched" line. - Fix a pre-existing bug where every nav/CTA color accent silently no-oped (Streamlit exposes st.button(key=...) as a `st-key-{key}` wrapper class, not a `key` attribute) and tighten the palette to three families (emerald brand, violet AI, semantic bull/bear). - Add a "Sources analyzed" provenance card, quarter/annual basis split on the financials grid, and cross-category signal dedup. - Add valuation + peer comparison, next-catalysts, brief-vs-brief delta, and one-page PDF export. - Rename the product from Primer to Amplegest across the app, docs, and the Hugging Face Space (old URL redire…7880373
-
chore: data artifacts from end-to-end generation validation run (NVDA)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>be5e5ac -
chore: checkpoint sidebar nav + i18n work-in-progress before Decision Stack redesign
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>e6496c0
-
deploy: add MSFT + AMD data (sections.db updated, 3 tickers total)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>cbd1881 -
deploy: refresh NVDA data (sections.db recreated locally)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>25e64af -
deploy: add sections.db + updated data for NVDA (15 filings ingested)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>fa8600a
-
feat: analytical edge section + fix HF Spaces fragment polling race
- Replace run_every=0.4 fragment with server-side sleep loop to eliminate browser-timer/proxy race causing 'SessionInfo before initialized' on HF Spaces - Add analytical_tensions, earnings_quality_signals, non_obvious_takeaway to BriefOutput schema and synthesis prompt - Render Analytical Edge section on Verdict tab (tension cards, quality signal chips, non-obvious takeaway) - Prune analytical_tensions that duplicate bull/bear point evidence (Jaccard) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>05d403b -
deploy: Primer app for HF Spaces — clean orphan history
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>35676b4