v2.6 generator update

.gitignore

@@ -1,7 +1,14 @@
 latexdocs/
+docs/*.aux
+docs/*.fdb_latexmk
+docs/*.fls
+docs/*.log
+docs/*.out
+docs/*.toc
+docs/*.synctex.gz
 generator.py
 */uv.lock
 **/.venv
 *.h5ad
 **/__pycache__/
-derived/**/*
+derived/**/*

README.md

@@ -36,11 +36,12 @@ machinery condition monitoring. It represents 180 calendar days for a fleet of
 All measurements are synthetic. DRMHB is not field data and is not a substitute
 for validation on real machinery.
 
-The ZIPs currently under `data/` are the 2.6 compact campaign and its separate
-paired controls. The conversion examples use
+The ZIPs under `data/` are the **2.6.0 compact campaign** and its separate
+512 paired controls. The conversion examples use
 [`vibframe-anndata==0.3.0`](https://413hq.github.io/vibframe-anndata-docs/0.3.0/)
-with complete evaluation preservation. The published 2.5.0 EDA results below
-refer to an earlier fixture and have not been rerun on these ZIPs.
+with complete evaluation preservation. The complete EDA, all report figures and
+baseline tables were rerun on the current campaign and checked on 1 October 2026.
+Measured identities and checks are in [the dataset audit](docs/provenance/current_dataset.json).
 
 **Start here:** [dataset report](docs/DRMHB.pdf), [conversion examples](examples/README.md)
 and [complete ground-truth guide](docs/GROUND_TRUTH.md).
@@ -96,22 +97,24 @@ The 46,907 observations are divided as follows:
 
 | Split | Rows | Intended use |
 |---|---:|---|
-| `calibration` | 7,984 | Learn healthy normality |
-| `normal_holdout` | 3,657 | Calibrate thresholds and false-positive behaviour |
-| `future_stream` | 35,266 | Causal evaluation with faults, transitions and OOD |
+| `calibration` | 8,179 | Learn healthy normality |
+| `normal_holdout` | 3,723 | Calibrate thresholds and false-positive behaviour |
+| `future_stream` | 35,005 | Causal evaluation with faults, transitions and OOD |
 
 Acquisition is deliberately non-uniform:
 
 | Context | Policy | Realized rows |
 |---|---|---:|
-| Startup transition | Every 5 minutes | 10,752 |
-| Shutdown transition | Every 5 minutes | 10,752 |
-| Internal regime transition | Every 10-15 minutes | 18,571 |
-| Stable running | Every 2-4 hours | 3,424 |
+| Startup transition | Reference: every 5 minutes | 8,729 |
+| Shutdown transition | Reference: every 5 minutes | 8,784 |
+| Internal regime transition | Reference: every 10-15 minutes | 14,562 |
+| Stable running | Reference: every 2-4 hours, plus reallocated captures | 11,424 |
 | OFF | Every 8 hours | 3,408 |
 
-There are 40,075 transition observations because changing regions are sampled
-densely. Observation count is therefore not proportional to elapsed time.
+The compact profile reallocates 8,000 reference transition observations to steady
+operation while protecting endpoints, central progress and fast motor edges.
+There are 32,075 transition observations. Realized steady spacing has a median
+of 55 minutes; reference cadences are not strict spacing rules after reallocation. Observation count is therefore not proportional to elapsed time.
 Temporal plots and persistence metrics must use timestamps rather than row index.
 
 ## Signal acquisition and storage
@@ -123,7 +126,10 @@ that complete source realization:
 - `VEL_1K`: velocity spectrum from 0 to 1,000 Hz;
 - `ACC_10K`: acceleration spectrum from 0 to 10,000 Hz.
 
-Each spectrum has 1,601 bins. With three directions and two processing modes,
+`VEL_1K` stores native peak lines at 0.625 Hz spacing. `ACC_10K` stores the
+root-sum-square of native peak lines in 6.25 Hz bands, rather than a maximum line
+amplitude. Under its periodic-Hann convention, RMS is `sqrt(sum(data**2)/3)`;
+apply the correction once. Each spectrum has 1,601 bins. With three directions and two processing modes,
 the dataset contains 281,442 spectral signals.
 
 The persisted waveform branch is anti-alias filtered and decimated to 25.6 kHz.
@@ -132,12 +138,14 @@ Its length is time-based rather than revolution-normalized:
 | Stored window | Samples/channel | Role |
 |---|---:|---|
 | 0.25 s | 6,400 | Ordinary OFF observation |
-| 0.50 s | 12,800 | Ordinary active or transition observation |
-| 1.60 s | 40,960 | Stratified anchor or localized event |
+| 0.40 s | 10,240 | Ordinary active or transition observation |
+| 0.75 s | 19,200 | Localized event |
+| 1.60 s | 40,960 | Stratified anchor |
 
-The current fixture contains 3,240 short OFF windows, 40,551 ordinary windows
-and 3,116 full-length windows. Each observation has H/V/A waveforms, giving
-140,721 waveform signals and 2,002,260,480 persisted samples in total.
+The current fixture contains 3,240 short OFF windows, 41,125 ordinary windows,
+198 event windows and 2,344 full-length anchors. Each observation has H/V/A
+waveforms, giving 140,721 waveform signals and 1,625,003,520 persisted samples.
+The separate paired controls use 0.50 s / 12,800-sample windows.
 
 Waveforms are quantized to an int16-equivalent grid and reconstructed as
 physical-unit float32 values to remain compatible with VibFrame 0.2.
@@ -147,7 +155,8 @@ physical-unit float32 values to remain compatible with VibFrame 0.2.
 Healthy active observations may contain low-probability disturbances such as
 RPM drift, load transients, electrical interference, sensor-noise bursts and
 transient impacts. The current fixture contains 1,756 such rows. Three healthy
-machines also receive short operational-OOD windows, totaling 88 observations.
+machines each receive eight operational-OOD episodes, totaling 24 episodes and
+1,057 observations. The report lists every configured UTC interval, RPM and valve command.
 
 Fault distribution:
 
@@ -194,7 +203,8 @@ Snapshot joins use **source, machine and integer UTC-microsecond `snap_t`**, not
 row order. Per-waveform joins also identify the raw channel. Capture `t` can be
 later than `snap_t` because of cropping; it is not a different snapshot.
 
-The reference snapshot view had 42 columns. Code must inspect actual fields
+The current original snapshot-truth tables have 49 columns; the aligned view
+also contains conversion identity fields. Code must inspect actual fields
 instead of hard-coding a fixed schema. Do not use `split`, `fault_*`, severity,
 physical states, `ood_operational`, nuisance labels or DiagGT as unsupervised
 model inputs. Operational-state knowledge must be an explicitly declared
@@ -228,8 +238,7 @@ python examples/05_read_evaluation.py derived/DRMHB-compact.v0.3.0.features.h5ad
 
 Example 01 explicitly enables `ground_truth.scope="all"`, stores numerical
 values as `float32`, and ingests in raw-data blocks of 8 MiB. The featureless base
-has `X.shape == (46907, 0)` for the current longitudinal ZIP. The reference
-feature requests materialized 468 machine/source-scoped columns in the reference
+has `X.shape == (46907, 0)` for the current longitudinal ZIP. The feature requests materialized 468 machine/source-scoped columns in the current
 EDA, with structural missing values for other machines. They are not 468 dense
 fleet-wide descriptors; the notebook explains the logical-column alignment.
 
@@ -256,11 +265,12 @@ in waveform truth.
 
 ### Storage and memory
 
-The current longitudinal ZIP is **7,635,324,956 bytes (7.111 GiB)**. The reference 0.2.2
-run measured about 9.35 GiB for its featureless H5AD and 9.44 GiB after reference
-features. Those are historical snapshot-only sizes, **not measured 0.3.0 sizes**.
-Full annotations add original sidecar bytes, aligned views and capture-time
-matrices. Examples report their actual output size rather than extrapolating.
+The longitudinal ZIP is **7,635,324,956 bytes (7.111 GiB)**. The measured
+0.3.0 featureless H5AD is **8,771,810,437 bytes (8.169 GiB)** and the feature H5AD
+is **8,874,970,373 bytes (8.265 GiB)**. Both retain 97 original annotation files,
+totaling 26,874,052 bytes. The audit verifies all original annotation bytes
+against the ZIP, snapshot identities, waveform lengths and identical raw buffers
+in both H5ADs. These sizes include full annotations and capture-time matrices.
 
 The default 512 MiB sidecar budget covers original encoded annotation files,
 not decoded tables or total process RAM. Inspect exact archived sizes with
@@ -304,56 +314,71 @@ hours; an observation gap longer than 4.5 hours breaks the run.
 
 ## Compact reference baseline
 
-The values below come from an earlier EDA run on the 2.5.0 fixture with
-converter 0.2.2. Updating the annotation API does not
-constitute rerunning those experiments.
-
-The reference notebook fits four deliberately small unsupervised baselines.
-Thresholds are the 99th percentile of normal-holdout scores.
+The values below come from the complete current 2.6.0 EDA using converter 0.3.0.
+Four deliberately small unsupervised baselines use 39 logical features, fitted
+on calibration only. Thresholds are the 99th percentile of normal-holdout scores.
 
 | Model | AUROC | AUPRC | F1 | Healthy FPR | Transition FPR | OOD FPR | Median delay |
 |---|---:|---:|---:|---:|---:|---:|---:|
-| Robust distance | 0.739 | 0.758 | 0.245 | 1.08% | 1.16% | 4.55% | 35.78 d |
-| PCA reconstruction | 0.693 | 0.778 | 0.531 | 1.07% | 1.11% | 4.55% | 1.75 d |
-| Isolation Forest | **0.814** | **0.843** | 0.485 | 1.23% | 1.40% | 3.41% | 7.77 d |
-| One-Class SVM | 0.665 | 0.765 | **0.544** | 1.15% | 1.29% | **1.14%** | **0.88 d** |
+| Robust distance | 0.741 | 0.771 | 0.237 | 0.616% | 0.668% | **0.000%** | 38.44 d |
+| PCA reconstruction | 0.735 | 0.768 | 0.329 | 1.017% | 1.142% | 3.784% | **13.48 d** |
+| Isolation Forest | **0.815** | **0.843** | 0.485 | 1.371% | 1.675% | 5.866% | 53.25 d |
+| One-Class SVM | 0.687 | 0.774 | **0.508** | 0.842% | 1.049% | 2.744% | 36.51 d |
+
+These are sanity checks, not leaderboard targets. Median delay excludes
+undetected machines: robust distance detects 2/9 faulty machines, PCA 3/9,
+and Isolation Forest and One-Class SVM each 5/9. Report coverage with delay.
+OOD false positives exceed ordinary healthy FPR for three models.
 
-These values are sanity checks, not leaderboard targets. Several machines are
-not detected by any compact baseline, and OOD false positives remain materially
-higher than ordinary healthy FPR for three models.
+Healthy calibration requires 6, 9 and 13 PCA components for 80%, 90% and 95%
+cumulative variance, respectively. The TwoNN intrinsic-dimension estimate is 9.17.
 
-Healthy calibration also remains non-trivial in the reference features: PCA
-requires 7, 10 and 14 components for 80%, 90% and 95% cumulative variance,
-respectively, and the TwoNN intrinsic-dimension estimate is approximately 8.80.
+The 512 paired controls contain 256 damage interventions and 256 healthy
+operating-state interventions (1,024 snapshots). Their timestamps are variant
+identifiers, not a degradation trajectory. `fault_present` and
+`diagnostic_eligible` (RPM >= 120) are separate truth fields; eligibility does
+not guarantee detectability. No RUL or failure-time target is defined.
 
 ## Repository contents
 
 ```text
 DRMHB/
 |- data/DRMHB-compact.vibframe.zip          current longitudinal signal dataset
-|- data/DRMHB-compact-controls.vibframe.zip separate paired controls
-|- derived/                      local H5ADs (not versioned)
+|- data/DRMHB-compact-controls.vibframe.zip separate 512-pair controls
+|- derived/                      local H5ADs and EDA exports (not versioned)
 |- examples/                     import, inspect, features, retrofit, evaluation readers
-|- notebooks/DRMHB_EDA.ipynb       editable 0.3.0 EDA workflow, outputs cleared
-|- notebooks/reference/          immutable executed source of published EDA results
+|- notebooks/DRMHB_EDA.ipynb       complete executed 0.3.0 EDA on the 2.6.0 campaign
 |- latexdocs/DRMHB.tex            editable report source
-|- docs/DRMHB.pdf                 compiled dataset report
-|- docs/figures/                  EDA images and editable vector diagrams
-|- docs/provenance/               figure/table and source identities
-|- tools/                        report build and reference-asset recovery
-`- tests/                        small, synthetic example-integration checks
+|- latexdocs/figures/             promoted EDA images and editable TikZ diagrams
+|- latexdocs/provenance/          figure/table and report input identities
+|- docs/DRMHB.pdf                 compiled current dataset report
+|- docs/provenance/               measured current dataset audit
+`- tools/                        audit and report build
+```
+
+The [report](docs/DRMHB.pdf) explains the installation, schedules, acquisitions,
+degradation, paired controls, evaluation and complete AnnData workflow.
+The 37 figures and two baseline tables come from the complete current notebook
+execution; their [manifest](latexdocs/provenance/eda_assets.json) records hashes
+and notebook identity. The TikZ week was verified against the current full-horizon
+process recipe.
+
+Run the audit with the notebook's Python environment:
+
+```bash
+python tools/audit_dataset.py --deep
+python tools/build_report.py --refresh-assets
 ```
 
-The [report](docs/DRMHB.pdf) explains the installation, daily/weekly state
-schedule, acquisitions, degradation, evaluation and complete AnnData workflow.
-Build it with `python tools/build_report.py` once a LaTeX distribution with
-`latexmk` and TikZ is installed. The report's 37 EDA figures and two baseline
-tables are recovered from the original executed notebook, not redrawn from
-invented data. Their [manifest](docs/provenance/eda_assets.json) records source
-cells and hashes. The updated notebook exports new plots/tables to
-`derived/eda_v0.3.0/`, separately from the report reference assets. Review and
-promote a complete new run together with its numerical narrative and provenance;
-partial execution cannot silently replace published figures.
+`--deep` checks every raw numerical buffer in the base and feature H5ADs.
+The build requires a complete sequential notebook execution and promotes all
+report figures/tables together. Rebuild without promotion using
+`python tools/build_report.py`; a LaTeX distribution with `latexmk` and TikZ is
+required because the report has multiple input files and image assets.
+
+The source VibFrame validation reports have `ok=true`, `strict_ok=false`:
+the warning concerns the open node type `CentrifugalPump`. The consistency audit
+does not re-synthesise signals or independently validate the physical model.
 
 ## Limitations
 

docs/DRMHB.pdf

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:12b5c7e09cfa3b57bdc71792b5a29fe921d52cdf0d37176a320beb57f32b8a85
-size 4502462
+oid sha256:2884161a3b8ec3091794fb2946b53db4f825627b52a73ed0fa1f6ad50ebb7c99
+size 7657654

docs/GROUND_TRUTH.md

@@ -110,8 +110,14 @@ truth still provides offsets/durations for audit.
 The H5AD grows by original sidecar bytes, aligned labels, the manifest and timing
 companions. `list_evaluation_files()` reports exact original sizes. The 512 MiB
 budget refers to encoded sidecars, not total RAM. No raw compression is changed.
-The current longitudinal ZIP is 7,635,324,956 bytes; full-annotation H5AD sizes must be
-measured after conversion rather than reusing the older snapshot-only sizes.
+The current longitudinal ZIP is 7,635,324,956 bytes. Measured complete H5ADs are
+8,771,810,437 bytes (base) and 8,874,970,373 bytes (features), preserving 97
+original files totaling 26,874,052 bytes. The [current audit](provenance/current_dataset.json)
+verifies their source annotation bytes, identities and waveform lengths, plus
+complete raw-buffer equality between base and feature files. The current original
+snapshot-truth schema has 49 columns. Campaign windows are 0.25/0.4/0.75/1.6 s;
+paired controls use 0.5 s. `fault_present` and `diagnostic_eligible` are distinct;
+no RUL or failure-time label is defined.
 
 Authoritative API and full coverage rules:
 https://413hq.github.io/vibframe-anndata-docs/0.3.0/guides/ground-truth/

docs/provenance/current_dataset.json

@@ -0,0 +1,564 @@
+{
+  "checked_at_utc": "2026-10-01T06:55:18.716273+00:00",
+  "archives": [
+    {
+      "path": "data/DRMHB-compact.vibframe.zip",
+      "size_bytes": 7635324956,
+      "sha256": "58d3fd1042cfac189ba9a464954fbcf3e6f990b75327ddaf1986d13e7583eb1b",
+      "generator_version": "2.6.0",
+      "dataset_role": "campaign",
+      "run_fingerprint": "52077ee4d777e4636ebcb334dc5550c09bdb0aae71129ae9d5caf226484f1e13",
+      "snapshots": 46907,
+      "snapshot_truth_columns": 49,
+      "spectral_signals": 281442,
+      "spectral_samples": 450588642,
+      "waveform_signals": 140721,
+      "waveform_samples": 1625003520,
+      "splits": {
+        "future_stream": 35005,
+        "calibration": 8179,
+        "normal_holdout": 3723
+      },
+      "acquisition_counts": {
+        "regime_transition": 14562,
+        "steady": 11424,
+        "shutdown_transition": 8784,
+        "startup_transition": 8729,
+        "off": 3408
+      },
+      "nuisance_snapshots": 1756,
+      "ood_snapshots": 1057,
+      "waveform_windows": {
+        "0.25": 3240,
+        "0.4000000059604645": 41125,
+        "0.75": 198,
+        "1.600000023841858": 2344
+      },
+      "storage_roles": {
+        "temporal_window": 41125,
+        "stationary_off_window": 3240,
+        "full_stratified_anchor": 2344,
+        "localized_event_window": 198
+      },
+      "stored_waveform_policy": {
+        "active_seconds": 0.4,
+        "off_seconds": 0.25,
+        "event_seconds": 0.75,
+        "anchor_seconds": 1.6,
+        "anchor_fraction": 0.05,
+        "control_seconds": 0.5,
+        "dtype": "list<float32>",
+        "quantization": "per-row int16 grid reconstructed to physical-g float32"
+      },
+      "ood_episodes": {
+        "M01": [],
+        "M02": [
+          {
+            "id": "M02-OOD-01",
+            "start": "2026-03-04 08:00:00+00:00",
+            "end": "2026-03-05 22:30:00+00:00",
+            "rpm": 1800.0,
+            "valve": 0.35
+          },
+          {
+            "id": "M02-OOD-02",
+            "start": "2026-03-17 08:00:00+00:00",
+            "end": "2026-03-18 22:30:00+00:00",
+            "rpm": 1100.0,
+            "valve": 0.92
+          },
+          {
+            "id": "M02-OOD-03",
+            "start": "2026-04-01 08:00:00+00:00",
+            "end": "2026-04-02 22:30:00+00:00",
+            "rpm": 1580.0,
+            "valve": 0.25
+          },
+          {
+            "id": "M02-OOD-04",
+            "start": "2026-04-16 08:00:00+00:00",
+            "end": "2026-04-17 18:00:00+00:00",
+            "rpm": 1850.0,
+            "valve": 0.68
+          },
+          {
+            "id": "M02-OOD-05",
+            "start": "2026-05-04 08:00:00+00:00",
+            "end": "2026-05-05 22:30:00+00:00",
+            "rpm": 950.0,
+            "valve": 0.9
+          },
+          {
+            "id": "M02-OOD-06",
+            "start": "2026-05-18 08:00:00+00:00",
+            "end": "2026-05-19 22:30:00+00:00",
+            "rpm": 1650.0,
+            "valve": 0.3
+          },
+          {
+            "id": "M02-OOD-07",
+            "start": "2026-06-01 08:00:00+00:00",
+            "end": "2026-06-02 22:30:00+00:00",
+            "rpm": 1200.0,
+            "valve": 0.98
+          },
+          {
+            "id": "M02-OOD-08",
+            "start": "2026-06-16 08:00:00+00:00",
+            "end": "2026-06-17 22:30:00+00:00",
+            "rpm": 1780.0,
+            "valve": 0.48
+          }
+        ],
+        "M03": [],
+        "M04": [],
+        "M05": [],
+        "M06": [],
+        "M07": [
+          {
+            "id": "M07-OOD-01",
+            "start": "2026-03-05 08:00:00+00:00",
+            "end": "2026-03-06 18:00:00+00:00",
+            "rpm": 1100.0,
+            "valve": 0.92
+          },
+          {
+            "id": "M07-OOD-02",
+            "start": "2026-03-18 08:00:00+00:00",
+            "end": "2026-03-19 22:30:00+00:00",
+            "rpm": 1580.0,
+            "valve": 0.25
+          },
+          {
+            "id": "M07-OOD-03",
+            "start": "2026-04-02 08:00:00+00:00",
+            "end": "2026-04-03 18:00:00+00:00",
+            "rpm": 1850.0,
+            "valve": 0.68
+          },
+          {
+            "id": "M07-OOD-04",
+            "start": "2026-04-20 08:00:00+00:00",
+            "end": "2026-04-21 22:30:00+00:00",
+            "rpm": 950.0,
+            "valve": 0.9
+          },
+          {
+            "id": "M07-OOD-05",
+            "start": "2026-05-04 08:00:00+00:00",
+            "end": "2026-05-05 22:30:00+00:00",
+            "rpm": 1650.0,
+            "valve": 0.3
+          },
+          {
+            "id": "M07-OOD-06",
+            "start": "2026-05-18 08:00:00+00:00",
+            "end": "2026-05-19 22:30:00+00:00",
+            "rpm": 1200.0,
+            "valve": 0.98
+          },
+          {
+            "id": "M07-OOD-07",
+            "start": "2026-06-01 08:00:00+00:00",
+            "end": "2026-06-02 22:30:00+00:00",
+            "rpm": 1780.0,
+            "valve": 0.48
+          },
+          {
+            "id": "M07-OOD-08",
+            "start": "2026-06-17 08:00:00+00:00",
+            "end": "2026-06-18 22:30:00+00:00",
+            "rpm": 1800.0,
+            "valve": 0.35
+          }
+        ],
+        "M08": [],
+        "M09": [],
+        "M10": [],
+        "M11": [
+          {
+            "id": "M11-OOD-01",
+            "start": "2026-03-09 08:00:00+00:00",
+            "end": "2026-03-10 22:30:00+00:00",
+            "rpm": 1580.0,
+            "valve": 0.25
+          },
+          {
+            "id": "M11-OOD-02",
+            "start": "2026-03-19 08:00:00+00:00",
+            "end": "2026-03-20 18:00:00+00:00",
+            "rpm": 1850.0,
+            "valve": 0.68
+          },
+          {
+            "id": "M11-OOD-03",
+            "start": "2026-04-06 08:00:00+00:00",
+            "end": "2026-04-07 22:30:00+00:00",
+            "rpm": 950.0,
+            "valve": 0.9
+          },
+          {
+            "id": "M11-OOD-04",
+            "start": "2026-04-20 08:00:00+00:00",
+            "end": "2026-04-21 22:30:00+00:00",
+            "rpm": 1650.0,
+            "valve": 0.3
+          },
+          {
+            "id": "M11-OOD-05",
+            "start": "2026-05-04 08:00:00+00:00",
+            "end": "2026-05-05 22:30:00+00:00",
+            "rpm": 1200.0,
+            "valve": 0.98
+          },
+          {
+            "id": "M11-OOD-06",
+            "start": "2026-05-18 08:00:00+00:00",
+            "end": "2026-05-19 22:30:00+00:00",
+            "rpm": 1780.0,
+            "valve": 0.48
+          },
+          {
+            "id": "M11-OOD-07",
+            "start": "2026-06-02 08:00:00+00:00",
+            "end": "2026-06-03 22:30:00+00:00",
+            "rpm": 1800.0,
+            "valve": 0.35
+          },
+          {
+            "id": "M11-OOD-08",
+            "start": "2026-06-18 08:00:00+00:00",
+            "end": "2026-06-19 18:00:00+00:00",
+            "rpm": 1100.0,
+            "valve": 0.92
+          }
+        ],
+        "M12": []
+      },
+      "validation_ok": true,
+      "strict_validation_ok": false,
+      "validation_warnings": [
+        "definition.undocumented-node-type"
+      ]
+    },
+    {
+      "path": "data/DRMHB-compact-controls.vibframe.zip",
+      "size_bytes": 185975818,
+      "sha256": "b193177effafaf17089f753e57cd94b5ad969873d7cedfdceaab81171fdee128",
+      "generator_version": "2.6.0",
+      "dataset_role": "controls",
+      "run_fingerprint": "27931f2e6d0a1d8ad6b7b83c8e2d7af7df2dbf62848fdf6bb07553322e904cf4",
+      "snapshots": 1024,
+      "snapshot_truth_columns": 49,
+      "spectral_signals": 6144,
+      "spectral_samples": 9836544,
+      "waveform_signals": 3072,
+      "waveform_samples": 39321600,
+      "splits": {
+        "controlled_pairs": 1024
+      },
+      "acquisition_counts": {
+        "controlled_pair": 1024
+      },
+      "nuisance_snapshots": 0,
+      "ood_snapshots": 0,
+      "waveform_windows": {
+        "0.5": 1024
+      },
+      "storage_roles": {
+        "paired_control_window": 1024
+      },
+      "stored_waveform_policy": {
+        "active_seconds": 0.4,
+        "off_seconds": 0.25,
+        "event_seconds": 0.75,
+        "anchor_seconds": 1.6,
+        "anchor_fraction": 0.05,
+        "control_seconds": 0.5,
+        "dtype": "list<float32>",
+        "quantization": "per-row int16 grid reconstructed to physical-g float32"
+      },
+      "ood_episodes": {
+        "M01": [],
+        "M02": [
+          {
+            "id": "M02-OOD-01",
+            "start": "2026-03-04 08:00:00+00:00",
+            "end": "2026-03-05 22:30:00+00:00",
+            "rpm": 1800.0,
+            "valve": 0.35
+          },
+          {
+            "id": "M02-OOD-02",
+            "start": "2026-03-17 08:00:00+00:00",
+            "end": "2026-03-18 22:30:00+00:00",
+            "rpm": 1100.0,
+            "valve": 0.92
+          },
+          {
+            "id": "M02-OOD-03",
+            "start": "2026-04-01 08:00:00+00:00",
+            "end": "2026-04-02 22:30:00+00:00",
+            "rpm": 1580.0,
+            "valve": 0.25
+          },
+          {
+            "id": "M02-OOD-04",
+            "start": "2026-04-16 08:00:00+00:00",
+            "end": "2026-04-17 18:00:00+00:00",
+            "rpm": 1850.0,
+            "valve": 0.68
+          },
+          {
+            "id": "M02-OOD-05",
+            "start": "2026-05-04 08:00:00+00:00",
+            "end": "2026-05-05 22:30:00+00:00",
+            "rpm": 950.0,
+            "valve": 0.9
+          },
+          {
+            "id": "M02-OOD-06",
+            "start": "2026-05-18 08:00:00+00:00",
+            "end": "2026-05-19 22:30:00+00:00",
+            "rpm": 1650.0,
+            "valve": 0.3
+          },
+          {
+            "id": "M02-OOD-07",
+            "start": "2026-06-01 08:00:00+00:00",
+            "end": "2026-06-02 22:30:00+00:00",
+            "rpm": 1200.0,
+            "valve": 0.98
+          },
+          {
+            "id": "M02-OOD-08",
+            "start": "2026-06-16 08:00:00+00:00",
+            "end": "2026-06-17 22:30:00+00:00",
+            "rpm": 1780.0,
+            "valve": 0.48
+          }
+        ],
+        "M03": [],
+        "M04": [],
+        "M05": [],
+        "M06": [],
+        "M07": [
+          {
+            "id": "M07-OOD-01",
+            "start": "2026-03-05 08:00:00+00:00",
+            "end": "2026-03-06 18:00:00+00:00",
+            "rpm": 1100.0,
+            "valve": 0.92
+          },
+          {
+            "id": "M07-OOD-02",
+            "start": "2026-03-18 08:00:00+00:00",
+            "end": "2026-03-19 22:30:00+00:00",
+            "rpm": 1580.0,
+            "valve": 0.25
+          },
+          {
+            "id": "M07-OOD-03",
+            "start": "2026-04-02 08:00:00+00:00",
+            "end": "2026-04-03 18:00:00+00:00",
+            "rpm": 1850.0,
+            "valve": 0.68
+          },
+          {
+            "id": "M07-OOD-04",
+            "start": "2026-04-20 08:00:00+00:00",
+            "end": "2026-04-21 22:30:00+00:00",
+            "rpm": 950.0,
+            "valve": 0.9
+          },
+          {
+            "id": "M07-OOD-05",
+            "start": "2026-05-04 08:00:00+00:00",
+            "end": "2026-05-05 22:30:00+00:00",
+            "rpm": 1650.0,
+            "valve": 0.3
+          },
+          {
+            "id": "M07-OOD-06",
+            "start": "2026-05-18 08:00:00+00:00",
+            "end": "2026-05-19 22:30:00+00:00",
+            "rpm": 1200.0,
+            "valve": 0.98
+          },
+          {
+            "id": "M07-OOD-07",
+            "start": "2026-06-01 08:00:00+00:00",
+            "end": "2026-06-02 22:30:00+00:00",
+            "rpm": 1780.0,
+            "valve": 0.48
+          },
+          {
+            "id": "M07-OOD-08",
+            "start": "2026-06-17 08:00:00+00:00",
+            "end": "2026-06-18 22:30:00+00:00",
+            "rpm": 1800.0,
+            "valve": 0.35
+          }
+        ],
+        "M08": [],
+        "M09": [],
+        "M10": [],
+        "M11": [
+          {
+            "id": "M11-OOD-01",
+            "start": "2026-03-09 08:00:00+00:00",
+            "end": "2026-03-10 22:30:00+00:00",
+            "rpm": 1580.0,
+            "valve": 0.25
+          },
+          {
+            "id": "M11-OOD-02",
+            "start": "2026-03-19 08:00:00+00:00",
+            "end": "2026-03-20 18:00:00+00:00",
+            "rpm": 1850.0,
+            "valve": 0.68
+          },
+          {
+            "id": "M11-OOD-03",
+            "start": "2026-04-06 08:00:00+00:00",
+            "end": "2026-04-07 22:30:00+00:00",
+            "rpm": 950.0,
+            "valve": 0.9
+          },
+          {
+            "id": "M11-OOD-04",
+            "start": "2026-04-20 08:00:00+00:00",
+            "end": "2026-04-21 22:30:00+00:00",
+            "rpm": 1650.0,
+            "valve": 0.3
+          },
+          {
+            "id": "M11-OOD-05",
+            "start": "2026-05-04 08:00:00+00:00",
+            "end": "2026-05-05 22:30:00+00:00",
+            "rpm": 1200.0,
+            "valve": 0.98
+          },
+          {
+            "id": "M11-OOD-06",
+            "start": "2026-05-18 08:00:00+00:00",
+            "end": "2026-05-19 22:30:00+00:00",
+            "rpm": 1780.0,
+            "valve": 0.48
+          },
+          {
+            "id": "M11-OOD-07",
+            "start": "2026-06-02 08:00:00+00:00",
+            "end": "2026-06-03 22:30:00+00:00",
+            "rpm": 1800.0,
+            "valve": 0.35
+          },
+          {
+            "id": "M11-OOD-08",
+            "start": "2026-06-18 08:00:00+00:00",
+            "end": "2026-06-19 18:00:00+00:00",
+            "rpm": 1100.0,
+            "valve": 0.92
+          }
+        ],
+        "M12": []
+      },
+      "validation_ok": true,
+      "strict_validation_ok": false,
+      "validation_warnings": [
+        "definition.undocumented-node-type"
+      ],
+      "control_pairs": 512,
+      "control_families": {
+        "damage": 256,
+        "operating_state": 256
+      }
+    }
+  ],
+  "derivatives": [
+    {
+      "path": "derived/DRMHB-compact.v0.3.0.raw.h5ad",
+      "size_bytes": 8771810437,
+      "shape": [
+        46907,
+        0
+      ],
+      "counts": {
+        "n_obs": 46907,
+        "waveform_signals": 140721,
+        "waveform_samples": 1625003520,
+        "spectrum_signals": 281442,
+        "spectrum_samples": 450588642,
+        "ground_truth_rows": 46907,
+        "waveform_ground_truth_snapshots": 46907
+      },
+      "evaluation_files": 97,
+      "evaluation_bytes": 26874052,
+      "source_sidecar_bytes_verified": true,
+      "validation": {
+        "path": "/Users/agzapico/Documents/Repos/DRMHB/derived/DRMHB-compact.v0.3.0.raw.h5ad",
+        "n_obs": 46907,
+        "n_vars": 0,
+        "machines": 12,
+        "has_waveform": 46907,
+        "has_spectrum": 46907,
+        "ground_truth_rows": 46907,
+        "ground_truth_columns": 51,
+        "evaluation_files": 97,
+        "evaluation_bytes": 26874052,
+        "waveform_ground_truth_snapshots": 46907,
+        "waveform_signals": 140721,
+        "waveform_samples": 1625003520,
+        "waveform_compression": "none",
+        "waveform_compression_opts": null,
+        "spectrum_signals": 281442,
+        "spectrum_samples": 450588642,
+        "spectrum_compression": "none",
+        "spectrum_compression_opts": null
+      }
+    },
+    {
+      "path": "derived/DRMHB-compact.v0.3.0.features.h5ad",
+      "size_bytes": 8874970373,
+      "shape": [
+        46907,
+        468
+      ],
+      "counts": {
+        "n_obs": 46907,
+        "waveform_signals": 140721,
+        "waveform_samples": 1625003520,
+        "spectrum_signals": 281442,
+        "spectrum_samples": 450588642,
+        "ground_truth_rows": 46907,
+        "waveform_ground_truth_snapshots": 46907
+      },
+      "evaluation_files": 97,
+      "evaluation_bytes": 26874052,
+      "source_sidecar_bytes_verified": true,
+      "validation": {
+        "path": "/Users/agzapico/Documents/Repos/DRMHB/derived/DRMHB-compact.v0.3.0.features.h5ad",
+        "n_obs": 46907,
+        "n_vars": 468,
+        "machines": 12,
+        "has_waveform": 46907,
+        "has_spectrum": 46907,
+        "ground_truth_rows": 46907,
+        "ground_truth_columns": 51,
+        "evaluation_files": 97,
+        "evaluation_bytes": 26874052,
+        "waveform_ground_truth_snapshots": 46907,
+        "waveform_signals": 140721,
+        "waveform_samples": 1625003520,
+        "waveform_compression": "none",
+        "waveform_compression_opts": null,
+        "spectrum_signals": 281442,
+        "spectrum_samples": 450588642,
+        "spectrum_compression": "none",
+        "spectrum_compression_opts": null
+      }
+    }
+  ],
+  "all_raw_buffers_identical_base_vs_features": true,
+  "scope": "ZIP identities, generation/validation reports, all construction sidecars, all archived H5AD annotation bytes, snapshot joins and all waveform lengths; optional complete base/feature raw-buffer comparison. No re-synthesis or independent physical-model validation."
+}

examples/README.md

@@ -41,7 +41,7 @@ The thirteen feature requests remain the same as in the published EDA:
 velocity/acceleration RMS, subsynchronous/1X/low-harmonic bands, 1X/2X/BPFO/BPFI
 amplitudes, VPF-family RMS, and waveform RMS/crest factor/kurtosis.
 
-For the reference fleet they materialized 468 machine/source-scoped columns.
+For the current 2.6.0 fleet they materialized 468 machine/source-scoped columns.
 Each machine populates its relevant descriptors; structural NaNs are not zeros.
 The EDA constructs a common logical descriptor matrix without averaging different
 machines. Preserve `var` units, processing mode, point and algorithm/options
@@ -103,9 +103,9 @@ current longitudinal ZIP. Use a distinct output basename for their H5AD.
 
 ## Resources and verification
 
-No raw compression settings or VibFrame bytes are changed. Historical 0.2.2
-H5AD sizes are not estimates of the full-annotation 0.3.0 outputs: measure the
-actual results. Metadata decoding also consumes RAM, so the raw block size and
+No raw compression settings or VibFrame bytes are changed. The current complete 0.3.0 base and feature H5ADs measure 8.169 and
+8.265 GiB respectively; both retain 97 original files (26,874,052 bytes).
+Verify local derivatives with `python tools/audit_dataset.py --deep`. Metadata decoding also consumes RAM, so the raw block size and
 512 MiB encoded-sidecar budget are not process-memory caps.
 
 For a quick end-to-end check using the smaller controls ZIP:

notebooks/DRMHB_EDA.ipynb

@@ -32,13 +32,12 @@
     "The notebook derives claims from the generated data rather than copying the generator plan. Selected figures are exported as PDF and PNG under `derived/eda_v0.3.0/figures/`, with a manifest describing the question answered by each figure.\n",
     "### Reproducible use\n",
     "\n",
-    "This editable workflow uses `vibframe-anndata==0.3.0` with complete annotations.\n",
-    "The earlier EDA run used the 2.5.0 fixture and converter 0.2.2; its executed\n",
-    "notebook is not included here. The input below is the current 2.6 compact ZIP.\n",
-    "This notebook starts with cleared outputs: an updated API is not a new experiment.\n",
-    "Fresh execution writes plots to `derived/eda_v0.3.0/figures/` and tables to\n",
-    "`derived/eda_v0.3.0/tables/`, with their provenance. It never overwrites the archived\n",
-    "report figures: promote a complete reviewed run, its results and captions together.\n"
+    "This notebook uses `vibframe-anndata==0.3.0` with complete annotations and the\n",
+    "current 2.6.0 compact campaign. A complete execution exports plots and tables to\n",
+    "`derived/eda_v0.3.0/`; the reviewed exports are promoted into the dataset report\n",
+    "with `python tools/build_report.py --refresh-assets`. Source identities and\n",
+    "checks are recorded in `docs/provenance/current_dataset.json` and\n",
+    "`latexdocs/provenance/eda_assets.json`. Paired controls remain a separate corpus.\n"
    ]
   },
   {
@@ -62,10 +61,24 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 1,
    "id": "8341546f",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:55:53.580307Z",
+     "iopub.status.busy": "2026-10-01T06:55:53.580033Z",
+     "iopub.status.idle": "2026-10-01T06:56:08.165794Z",
+     "shell.execute_reply": "2026-10-01T06:56:08.165345Z"
+    }
+   },
    "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Matplotlib is building the font cache; this may take a moment.\n"
+     ]
+    },
     {
      "name": "stdout",
      "output_type": "stream",
@@ -154,19 +167,24 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 2,
    "id": "f1ee8d23",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:08.167054Z",
+     "iopub.status.busy": "2026-10-01T06:56:08.166876Z",
+     "iopub.status.idle": "2026-10-01T06:56:09.694925Z",
+     "shell.execute_reply": "2026-10-01T06:56:09.694116Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "[ingest] Completed in 119.47 s.\n",
-      "[features] Completed in 236.46 s.\n",
       "Base: /Users/agzapico/Documents/Repos/DRMHB/derived/DRMHB-compact.v0.3.0.raw.h5ad (8.169 GiB)\n",
       "Features: /Users/agzapico/Documents/Repos/DRMHB/derived/DRMHB-compact.v0.3.0.features.h5ad (8.265 GiB)\n",
-      "Free disk: 190.419 GiB; document assets: /Users/agzapico/Documents/Repos/DRMHB/derived/eda_v0.3.0/figures\n"
+      "Free disk: 170.072 GiB; document assets: /Users/agzapico/Documents/Repos/DRMHB/derived/eda_v0.3.0/figures\n"
      ]
     }
    ],
@@ -279,9 +297,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 3,
    "id": "ee763b7b",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:09.696366Z",
+     "iopub.status.busy": "2026-10-01T06:56:09.696269Z",
+     "iopub.status.idle": "2026-10-01T06:56:09.700813Z",
+     "shell.execute_reply": "2026-10-01T06:56:09.700321Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -307,7 +332,7 @@
        " 'spectrum_compression_opts': None}"
       ]
      },
-     "execution_count": 5,
+     "execution_count": 3,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -332,9 +357,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 4,
    "id": "9b0d60c6",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:09.701898Z",
+     "iopub.status.busy": "2026-10-01T06:56:09.701831Z",
+     "iopub.status.idle": "2026-10-01T06:56:10.743825Z",
+     "shell.execute_reply": "2026-10-01T06:56:10.743431Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -1674,9 +1706,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 5,
    "id": "5a556ac4",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:10.744912Z",
+     "iopub.status.busy": "2026-10-01T06:56:10.744819Z",
+     "iopub.status.idle": "2026-10-01T06:56:11.397807Z",
+     "shell.execute_reply": "2026-10-01T06:56:11.397259Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -2093,9 +2132,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": 6,
    "id": "f24738df",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:11.398992Z",
+     "iopub.status.busy": "2026-10-01T06:56:11.398892Z",
+     "iopub.status.idle": "2026-10-01T06:56:14.184932Z",
+     "shell.execute_reply": "2026-10-01T06:56:14.184145Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -2589,9 +2635,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": 7,
    "id": "af98f5df",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:14.189131Z",
+     "iopub.status.busy": "2026-10-01T06:56:14.188974Z",
+     "iopub.status.idle": "2026-10-01T06:56:14.813280Z",
+     "shell.execute_reply": "2026-10-01T06:56:14.812525Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -2647,9 +2700,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": 8,
    "id": "b0bdf9ca",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:14.815029Z",
+     "iopub.status.busy": "2026-10-01T06:56:14.814845Z",
+     "iopub.status.idle": "2026-10-01T06:56:14.959117Z",
+     "shell.execute_reply": "2026-10-01T06:56:14.958690Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -2765,9 +2825,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": 9,
    "id": "f03201ce",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:14.960302Z",
+     "iopub.status.busy": "2026-10-01T06:56:14.960226Z",
+     "iopub.status.idle": "2026-10-01T06:56:15.223281Z",
+     "shell.execute_reply": "2026-10-01T06:56:15.222885Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -3008,9 +3075,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
+   "execution_count": 10,
    "id": "f3d0dc45",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:15.224668Z",
+     "iopub.status.busy": "2026-10-01T06:56:15.224579Z",
+     "iopub.status.idle": "2026-10-01T06:56:16.427999Z",
+     "shell.execute_reply": "2026-10-01T06:56:16.427543Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -3137,9 +3211,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 13,
+   "execution_count": 11,
    "id": "fea4b819",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:16.429448Z",
+     "iopub.status.busy": "2026-10-01T06:56:16.429339Z",
+     "iopub.status.idle": "2026-10-01T06:56:17.734308Z",
+     "shell.execute_reply": "2026-10-01T06:56:17.733783Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -3230,9 +3311,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
+   "execution_count": 12,
    "id": "93bf7b65",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:17.735516Z",
+     "iopub.status.busy": "2026-10-01T06:56:17.735427Z",
+     "iopub.status.idle": "2026-10-01T06:56:18.125968Z",
+     "shell.execute_reply": "2026-10-01T06:56:18.125268Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -3304,9 +3392,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 15,
+   "execution_count": 13,
    "id": "beb50cfb",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:18.127292Z",
+     "iopub.status.busy": "2026-10-01T06:56:18.127198Z",
+     "iopub.status.idle": "2026-10-01T06:56:18.143829Z",
+     "shell.execute_reply": "2026-10-01T06:56:18.143406Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -3354,9 +3449,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 16,
+   "execution_count": 14,
    "id": "b95e77b5",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:18.145194Z",
+     "iopub.status.busy": "2026-10-01T06:56:18.145119Z",
+     "iopub.status.idle": "2026-10-01T06:56:18.499445Z",
+     "shell.execute_reply": "2026-10-01T06:56:18.499047Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -3371,11 +3473,11 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/948014552.py:12: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/948014552.py:12: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  .gt(pd.Timedelta(minutes=30))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/948014552.py:24: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/948014552.py:24: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  lo = example[\"timestamp\"].min() - pd.Timedelta(hours=1)\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/948014552.py:25: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/948014552.py:25: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  hi = example[\"timestamp\"].max() + pd.Timedelta(hours=1)\n"
      ]
     },
@@ -3504,9 +3606,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 17,
+   "execution_count": 15,
    "id": "f801a856",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:18.500977Z",
+     "iopub.status.busy": "2026-10-01T06:56:18.500881Z",
+     "iopub.status.idle": "2026-10-01T06:56:18.661145Z",
+     "shell.execute_reply": "2026-10-01T06:56:18.660790Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -3591,9 +3700,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 18,
+   "execution_count": 16,
    "id": "08a1c195",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:18.662562Z",
+     "iopub.status.busy": "2026-10-01T06:56:18.662468Z",
+     "iopub.status.idle": "2026-10-01T06:56:19.670962Z",
+     "shell.execute_reply": "2026-10-01T06:56:19.670528Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -3728,9 +3844,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 19,
+   "execution_count": 17,
    "id": "4c105742",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:19.672030Z",
+     "iopub.status.busy": "2026-10-01T06:56:19.671953Z",
+     "iopub.status.idle": "2026-10-01T06:56:19.676898Z",
+     "shell.execute_reply": "2026-10-01T06:56:19.676590Z"
+    }
+   },
    "outputs": [],
    "source": [
     "def find_channel(channels, *, machine, point_id, proc_mode_id):\n",
@@ -3809,9 +3932,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 20,
+   "execution_count": 18,
    "id": "19688b01",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:19.677909Z",
+     "iopub.status.busy": "2026-10-01T06:56:19.677837Z",
+     "iopub.status.idle": "2026-10-01T06:56:20.064610Z",
+     "shell.execute_reply": "2026-10-01T06:56:20.064196Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -3903,9 +4033,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 21,
+   "execution_count": 19,
    "id": "41604920",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:20.065673Z",
+     "iopub.status.busy": "2026-10-01T06:56:20.065604Z",
+     "iopub.status.idle": "2026-10-01T06:56:20.069025Z",
+     "shell.execute_reply": "2026-10-01T06:56:20.068612Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -4039,9 +4176,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 22,
+   "execution_count": 20,
    "id": "5693c86c",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:20.070032Z",
+     "iopub.status.busy": "2026-10-01T06:56:20.069963Z",
+     "iopub.status.idle": "2026-10-01T06:56:20.072235Z",
+     "shell.execute_reply": "2026-10-01T06:56:20.071842Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -4067,7 +4211,7 @@
        " 'spectrum_compression_opts': None}"
       ]
      },
-     "execution_count": 22,
+     "execution_count": 20,
      "metadata": {},
      "output_type": "execute_result"
     }
@@ -4090,9 +4234,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 23,
+   "execution_count": 21,
    "id": "0f0b10a3",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:20.073139Z",
+     "iopub.status.busy": "2026-10-01T06:56:20.073078Z",
+     "iopub.status.idle": "2026-10-01T06:56:20.563358Z",
+     "shell.execute_reply": "2026-10-01T06:56:20.562944Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -4728,9 +4879,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 24,
+   "execution_count": 22,
    "id": "06a3d8a0",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:20.564429Z",
+     "iopub.status.busy": "2026-10-01T06:56:20.564351Z",
+     "iopub.status.idle": "2026-10-01T06:56:20.658493Z",
+     "shell.execute_reply": "2026-10-01T06:56:20.658075Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -5034,9 +5192,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 25,
+   "execution_count": 23,
    "id": "713fded0",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:20.659691Z",
+     "iopub.status.busy": "2026-10-01T06:56:20.659617Z",
+     "iopub.status.idle": "2026-10-01T06:56:20.828799Z",
+     "shell.execute_reply": "2026-10-01T06:56:20.828394Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -5072,9 +5237,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 26,
+   "execution_count": 24,
    "id": "ce56a659",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:20.829963Z",
+     "iopub.status.busy": "2026-10-01T06:56:20.829877Z",
+     "iopub.status.idle": "2026-10-01T06:56:20.833786Z",
+     "shell.execute_reply": "2026-10-01T06:56:20.833136Z"
+    }
+   },
    "outputs": [],
    "source": [
     "if \"split\" in truth_f:\n",
@@ -5085,9 +5257,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 27,
+   "execution_count": 25,
    "id": "2f5498f8",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:20.834973Z",
+     "iopub.status.busy": "2026-10-01T06:56:20.834907Z",
+     "iopub.status.idle": "2026-10-01T06:56:21.297657Z",
+     "shell.execute_reply": "2026-10-01T06:56:21.297239Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -5149,9 +5328,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 28,
+   "execution_count": 26,
    "id": "015a6297",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:21.298795Z",
+     "iopub.status.busy": "2026-10-01T06:56:21.298706Z",
+     "iopub.status.idle": "2026-10-01T06:56:21.621832Z",
+     "shell.execute_reply": "2026-10-01T06:56:21.621431Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -5204,9 +5390,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 29,
+   "execution_count": 27,
    "id": "af6e2df9",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:21.622917Z",
+     "iopub.status.busy": "2026-10-01T06:56:21.622822Z",
+     "iopub.status.idle": "2026-10-01T06:56:21.728583Z",
+     "shell.execute_reply": "2026-10-01T06:56:21.728201Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -5267,9 +5460,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 30,
+   "execution_count": 28,
    "id": "b0b6668e",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:21.729765Z",
+     "iopub.status.busy": "2026-10-01T06:56:21.729695Z",
+     "iopub.status.idle": "2026-10-01T06:56:22.193939Z",
+     "shell.execute_reply": "2026-10-01T06:56:22.193468Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -5287,7 +5487,13 @@
      "text": [
       "Components for 80%: 6\n",
       "Components for 90%: 9\n",
-      "Components for 95%: 13\n",
+      "Components for 95%: 13\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
       "TwoNN intrinsic-dimension estimate: 9.17\n"
      ]
     }
@@ -5350,9 +5556,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 31,
+   "execution_count": 29,
    "id": "64845ed2",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:22.195634Z",
+     "iopub.status.busy": "2026-10-01T06:56:22.195500Z",
+     "iopub.status.idle": "2026-10-01T06:56:23.259570Z",
+     "shell.execute_reply": "2026-10-01T06:56:23.259090Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -5403,16 +5616,29 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 32,
+   "execution_count": 30,
    "id": "4d2fd903",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:23.261255Z",
+     "iopub.status.busy": "2026-10-01T06:56:23.261131Z",
+     "iopub.status.idle": "2026-10-01T06:56:49.509528Z",
+     "shell.execute_reply": "2026-10-01T06:56:49.508608Z"
+    }
+   },
    "outputs": [
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
       "/Users/agzapico/Documents/Repos/DRMHB/.venv/lib/python3.12/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
-      "  from .autonotebook import tqdm as notebook_tqdm\n",
+      "  from .autonotebook import tqdm as notebook_tqdm\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
       "/Users/agzapico/Documents/Repos/DRMHB/.venv/lib/python3.12/site-packages/umap/umap_.py:1952: UserWarning: n_jobs value 1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
       "  warn(\n"
      ]
@@ -5474,9 +5700,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 33,
+   "execution_count": 31,
    "id": "9d5ee546",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:49.511989Z",
+     "iopub.status.busy": "2026-10-01T06:56:49.511903Z",
+     "iopub.status.idle": "2026-10-01T06:56:49.597599Z",
+     "shell.execute_reply": "2026-10-01T06:56:49.596907Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -5561,9 +5794,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 34,
+   "execution_count": 32,
    "id": "c45a61f4",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:49.599183Z",
+     "iopub.status.busy": "2026-10-01T06:56:49.599064Z",
+     "iopub.status.idle": "2026-10-01T06:56:49.605093Z",
+     "shell.execute_reply": "2026-10-01T06:56:49.604435Z"
+    }
+   },
    "outputs": [],
    "source": [
     "def first_truth_column(keywords, *, exclude=()):\n",
@@ -5661,9 +5901,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 35,
+   "execution_count": 33,
    "id": "85c6aab6",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:49.606828Z",
+     "iopub.status.busy": "2026-10-01T06:56:49.606676Z",
+     "iopub.status.idle": "2026-10-01T06:56:49.797845Z",
+     "shell.execute_reply": "2026-10-01T06:56:49.796909Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -5847,9 +6094,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 36,
+   "execution_count": 34,
    "id": "25e525c8",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:49.799303Z",
+     "iopub.status.busy": "2026-10-01T06:56:49.799218Z",
+     "iopub.status.idle": "2026-10-01T06:56:52.077516Z",
+     "shell.execute_reply": "2026-10-01T06:56:52.076594Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -5980,9 +6234,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 37,
+   "execution_count": 35,
    "id": "43265db8",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:52.079334Z",
+     "iopub.status.busy": "2026-10-01T06:56:52.079227Z",
+     "iopub.status.idle": "2026-10-01T06:56:53.925811Z",
+     "shell.execute_reply": "2026-10-01T06:56:53.925380Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -6064,9 +6325,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 38,
+   "execution_count": 36,
    "id": "7692dd90",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:53.927453Z",
+     "iopub.status.busy": "2026-10-01T06:56:53.927365Z",
+     "iopub.status.idle": "2026-10-01T06:56:54.391032Z",
+     "shell.execute_reply": "2026-10-01T06:56:54.390708Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -6156,9 +6424,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 39,
+   "execution_count": 37,
    "id": "a0ffe60e",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:54.392350Z",
+     "iopub.status.busy": "2026-10-01T06:56:54.392264Z",
+     "iopub.status.idle": "2026-10-01T06:56:56.037780Z",
+     "shell.execute_reply": "2026-10-01T06:56:56.037323Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -6476,9 +6751,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 40,
+   "execution_count": 38,
    "id": "1892d117",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:56.039246Z",
+     "iopub.status.busy": "2026-10-01T06:56:56.039138Z",
+     "iopub.status.idle": "2026-10-01T06:56:56.046688Z",
+     "shell.execute_reply": "2026-10-01T06:56:56.046272Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -6579,9 +6861,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 41,
+   "execution_count": 39,
    "id": "9387cc16",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:56.047621Z",
+     "iopub.status.busy": "2026-10-01T06:56:56.047556Z",
+     "iopub.status.idle": "2026-10-01T06:56:57.306286Z",
+     "shell.execute_reply": "2026-10-01T06:56:57.305908Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -6732,9 +7021,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 42,
+   "execution_count": 40,
    "id": "d1790db6",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:57.307481Z",
+     "iopub.status.busy": "2026-10-01T06:56:57.307400Z",
+     "iopub.status.idle": "2026-10-01T06:56:57.621453Z",
+     "shell.execute_reply": "2026-10-01T06:56:57.621139Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -6780,7 +7076,13 @@
       " - rms | pump_DE_V\n",
       " - vel_overall_rms | pump_DE_A\n",
       " - vel_overall_rms | pump_DE_H\n",
-      " - vel_overall_rms | pump_DE_V\n",
+      " - vel_overall_rms | pump_DE_V\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
       "X_train : (8179, 39)\n",
       "X_hold  : (3723, 39)\n",
       "X_future: (35005, 39)\n",
@@ -6889,17 +7191,24 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 43,
+   "execution_count": 41,
    "id": "8794ae1b",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:57.622767Z",
+     "iopub.status.busy": "2026-10-01T06:56:57.622693Z",
+     "iopub.status.idle": "2026-10-01T06:56:57.630251Z",
+     "shell.execute_reply": "2026-10-01T06:56:57.629527Z"
+    }
+   },
    "outputs": [
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:1: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:1: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  ALARM_PERSISTENCE = pd.Timedelta(hours=6)\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:2: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:2: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  MAX_ALARM_GAP = pd.Timedelta(hours=4.5)\n"
      ]
     }
@@ -7004,43 +7313,62 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 44,
+   "execution_count": 42,
    "id": "31344640",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:57.631326Z",
+     "iopub.status.busy": "2026-10-01T06:56:57.631262Z",
+     "iopub.status.idle": "2026-10-01T06:56:58.870983Z",
+     "shell.execute_reply": "2026-10-01T06:56:58.870555Z"
+    }
+   },
    "outputs": [
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
-      "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
-      "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n",
-      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_27534/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
+      "/var/folders/bv/rdwx1w692714why48mh5j1m80000gn/T/ipykernel_13171/10933082.py:49: DeprecationWarning: The 'generic' unit for NumPy timedelta is deprecated, and will raise an error in the future. This includes implicit conversion of bare integers (e.g. `+ 1`). Please use a specific unit instead.\n",
       "  delay = float((pd.Timestamp(alarm_time) - pd.Timestamp(onset_time)) / pd.Timedelta(days=1))\n"
      ]
     },
@@ -7256,9 +7584,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 45,
+   "execution_count": 43,
    "id": "7b3cda77",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:58.872176Z",
+     "iopub.status.busy": "2026-10-01T06:56:58.872104Z",
+     "iopub.status.idle": "2026-10-01T06:56:59.184150Z",
+     "shell.execute_reply": "2026-10-01T06:56:59.183720Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -7324,9 +7659,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 46,
+   "execution_count": 44,
    "id": "e517fbfc",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:59.185128Z",
+     "iopub.status.busy": "2026-10-01T06:56:59.185057Z",
+     "iopub.status.idle": "2026-10-01T06:56:59.498139Z",
+     "shell.execute_reply": "2026-10-01T06:56:59.497682Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -7382,9 +7724,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 47,
+   "execution_count": 45,
    "id": "9d828732",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:59.499190Z",
+     "iopub.status.busy": "2026-10-01T06:56:59.499101Z",
+     "iopub.status.idle": "2026-10-01T06:56:59.670203Z",
+     "shell.execute_reply": "2026-10-01T06:56:59.669760Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -7457,9 +7806,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 48,
+   "execution_count": 46,
    "id": "9f19d556",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:59.671426Z",
+     "iopub.status.busy": "2026-10-01T06:56:59.671273Z",
+     "iopub.status.idle": "2026-10-01T06:56:59.677179Z",
+     "shell.execute_reply": "2026-10-01T06:56:59.676780Z"
+    }
+   },
    "outputs": [
     {
      "name": "stdout",
@@ -7544,9 +7900,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 49,
+   "execution_count": 47,
    "id": "4ffc8b64",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:56:59.678129Z",
+     "iopub.status.busy": "2026-10-01T06:56:59.678053Z",
+     "iopub.status.idle": "2026-10-01T06:57:00.182882Z",
+     "shell.execute_reply": "2026-10-01T06:57:00.182473Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -7629,9 +7992,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 50,
+   "execution_count": 48,
    "id": "007b32ad",
-   "metadata": {},
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-10-01T06:57:00.184346Z",
+     "iopub.status.busy": "2026-10-01T06:57:00.184243Z",
+     "iopub.status.idle": "2026-10-01T06:57:00.972525Z",
+     "shell.execute_reply": "2026-10-01T06:57:00.971989Z"
+    }
+   },
    "outputs": [
     {
      "data": {
@@ -7737,17 +8107,19 @@
     "without first reading the generator implementation: fleet chronology, operating envelope,\n",
     "transition connectivity/hysteresis, waveform-storage policy, machine/regime fingerprints,\n",
     "fault chronology/signatures, and baseline difficulty are all visualized directly from the\n",
-    "persisted dataset and evaluation sidecar. Document-ready figures are written in both PDF and PNG form to `derived/eda_v0.3.0/figures/`; `manifest.csv` records the question answered by each export. The published `.tex` uses the archived reference assets. Promote a reviewed new run and update its numerical narrative and provenance together, rather than mixing partial new outputs with the reference report.\n"
+    "persisted dataset and evaluation sidecar. Document-ready figures are written in both PDF and PNG form to `derived/eda_v0.3.0/figures/`; `manifest.csv` records the question answered by each export. The report uses the complete reviewed 2.6.0 execution. Promote new figures and tables together with their numerical narrative using `tools/build_report.py --refresh-assets`; the build rejects incomplete executions and records hashes.\n"
    ]
   }
  ],
  "metadata": {
   "drmhb_provenance": {
-   "execution_status": "not_executed_on_full_campaign",
-   "note": "Historical figures and numerical results came from the 2.5.0 fixture; this cleared notebook targets the current 2.6 compact ZIP.",
-   "reference_hub_revision": "d218f6e1dc7505d7c92c0b23fc1d27808a26c107",
-   "reference_results_conversion_version": "0.2.2",
-   "workflow_package_version": "0.3.0"
+   "execution_status": "complete_current_campaign_execution",
+   "dataset_generator_version": "2.6.0",
+   "workflow_package_version": "0.3.0",
+   "source_archive": "data/DRMHB-compact.vibframe.zip",
+   "source_archive_sha256": "58d3fd1042cfac189ba9a464954fbcf3e6f990b75327ddaf1986d13e7583eb1b",
+   "dataset_audit": "docs/provenance/current_dataset.json",
+   "note": "All code cells executed sequentially without errors; complete current EDA promoted to the report."
   },
   "kernelspec": {
    "display_name": "vibframe-fleet-eda (3.12.14)",

tools/audit_dataset.py

@@ -0,0 +1,155 @@
+"""Audit local canonical ZIPs and H5AD derivatives without loading raw signals.
+
+Run from the repository root with the examples' Python environment. The JSON is
+measured evidence, not a generator storage estimate. --deep additionally checks
+that every numerical raw buffer is identical in the base and feature H5ADs.
+"""
+from __future__ import annotations
+
+import argparse
+import hashlib
+import io
+import json
+import sys
+import zipfile
+from datetime import datetime, timezone
+from pathlib import Path
+
+import h5py
+import numpy as np
+import pandas as pd
+import pyarrow.parquet as pq
+
+ROOT = Path(__file__).resolve().parents[1]
+sys.path.insert(0, str(ROOT / "examples"))
+from drmhb_workflow import preservation_signature, read_metadata, require_complete_drmhb
+from vibframe_anndata import get_snapshot_ground_truth, read_evaluation_file
+
+
+def sha256(path):
+    digest = hashlib.sha256()
+    with Path(path).open("rb") as handle:
+        for block in iter(lambda: handle.read(16 * 1024**2), b""):
+            digest.update(block)
+    return digest.hexdigest()
+
+
+def archive_summary(path):
+    with zipfile.ZipFile(path) as archive:
+        names = [n for n in archive.namelist() if not n.startswith("__MACOSX/") and not n.endswith("/")]
+        root = next(n.rsplit("/", 1)[0] for n in names if n.endswith("/scenario_config.json"))
+        config = json.loads(archive.read(root + "/scenario_config.json"))
+        frames = {}
+        for kind in ("snapshot", "waveform"):
+            parts = [pq.read_table(io.BytesIO(archive.read(n))).to_pandas() for n in names
+                     if f"/evaluation/{kind}_truth/" in n and n.endswith(".parquet")]
+            frames[kind] = pd.concat(parts, ignore_index=True)
+        truth, wave = frames["snapshot"], frames["waveform"]
+        snapshot_columns = len(truth.columns)
+        truth["snap_t"] = truth["timestamp"]
+        keys = ["machine_id", "snap_t"]
+        assert not truth.duplicated(keys).any()
+        assert not wave.duplicated(keys + ["point_id"]).any()
+        assert len(wave) == 3 * len(truth)
+        assert wave.groupby(keys).waveform_n_samples.nunique().eq(1).all()
+        assert wave.groupby(keys).waveform_start_offset_s.nunique().eq(1).all()
+        validation = json.loads(archive.read(root + "/validation_report.json"))
+        quality = json.loads(archive.read(root + "/quality_report.json"))
+        assert validation["ok"] and quality["ok"]
+        assert quality["total_snapshots"] == len(truth)
+        assert quality["total_waveform_samples"] == int(wave.waveform_n_samples.sum())
+        result = {
+            "path": str(path.relative_to(ROOT)), "size_bytes": path.stat().st_size,
+            "sha256": sha256(path), "generator_version": config["script_version"],
+            "dataset_role": config["dataset_role"], "run_fingerprint": config["run_fingerprint"],
+            "snapshots": len(truth), "snapshot_truth_columns": snapshot_columns,
+            "spectral_signals": 6 * len(truth), "spectral_samples": 6 * 1601 * len(truth),
+            "waveform_signals": len(wave), "waveform_samples": int(wave.waveform_n_samples.sum()),
+            "splits": truth.split.value_counts().to_dict(),
+            "acquisition_counts": truth.acquisition_kind.value_counts().to_dict(),
+            "nuisance_snapshots": int(truth.nuisance_type.fillna("").ne("").sum()),
+            "ood_snapshots": int(truth.ood_operational.sum()),
+            "waveform_windows": {str(k): int(v // 3) for k, v in wave.waveform_duration_s.value_counts().sort_index().items()},
+            "storage_roles": {str(k): int(v // 3) for k, v in wave.waveform_storage_mode.value_counts().items()},
+            "stored_waveform_policy": config["stored_waveform_policy"],
+            "ood_episodes": config["ood_episodes"],
+            "validation_ok": validation["ok"], "strict_validation_ok": validation["strict_ok"],
+            "validation_warnings": sorted({i["code"] for i in validation["issues"] if i["severity"] == "warning"}),
+        }
+        if config["dataset_role"] == "controls":
+            pairs = truth.groupby("control_pair_id")
+            assert pairs.size().eq(2).all()
+            assert pairs.control_variant.nunique().eq(2).all()
+            result["control_pairs"] = len(pairs)
+            result["control_families"] = truth.drop_duplicates("control_pair_id").control_family.value_counts().to_dict()
+        return result, truth, wave, root
+
+
+def audit(deep=False):
+    archives, source_truth, source_wave = [], None, None
+    for filename in ("DRMHB-compact.vibframe.zip", "DRMHB-compact-controls.vibframe.zip"):
+        summary, truth, wave, root = archive_summary(ROOT / "data" / filename)
+        archives.append(summary)
+        if summary["dataset_role"] == "campaign":
+            source_truth, source_wave, campaign_root = truth, wave, root
+    derivatives, signatures = [], []
+    with zipfile.ZipFile(ROOT / archives[0]["path"]) as archive:
+        for suffix in ("raw", "features"):
+            path = ROOT / "derived" / f"DRMHB-compact.v0.3.0.{suffix}.h5ad"
+            validation = require_complete_drmhb(path)
+            counts, inventory = preservation_signature(path)
+            signatures.append((counts, inventory))
+            meta = read_metadata(path)
+            gt = get_snapshot_ground_truth(path)
+            expected = source_truth.set_index(["machine_id", "snap_t"]).sort_index()
+            actual = gt.set_index(["machine_id", "snap_t"]).sort_index()
+            pd.testing.assert_index_equal(expected.index, actual.index)
+            for column in ("split", "acquisition_kind", "fault_present", "ood_operational", "waveform_rms_h"):
+                pd.testing.assert_series_equal(expected[column], actual[column], check_dtype=False)
+            assert counts["waveform_samples"] == archives[0]["waveform_samples"]
+            assert counts["spectrum_samples"] == archives[0]["spectral_samples"]
+            assert gt.index.equals(meta["obs"].index)
+            for row in inventory.itertuples():
+                original = archive.read(campaign_root + "/" + row.path)
+                assert len(original) == row.size_bytes
+                assert hashlib.sha256(original).hexdigest() == row.sha256
+                assert read_evaluation_file(path, row.path, source=row.source) == original
+            with h5py.File(path, "r") as handle:
+                lengths = handle["obsm/raw_waveforms_lengths"][:]
+                expected_lengths = source_wave.groupby(["machine_id", "snap_t"]).waveform_n_samples.first()
+                row_keys = pd.MultiIndex.from_frame(meta["obs"][["machine", "snap_t"]])
+                want = expected_lengths.reindex(row_keys).to_numpy()
+                assert (lengths > 0).sum(axis=1).tolist() == [3] * len(want)
+                assert np.array_equal(lengths.max(axis=1), want)
+            derivatives.append({"path": str(path.relative_to(ROOT)), "size_bytes": path.stat().st_size,
+                                "shape": list(meta["shape"]), "counts": counts,
+                                "evaluation_files": len(inventory), "evaluation_bytes": int(inventory.size_bytes.sum()),
+                                "source_sidecar_bytes_verified": True, "validation": validation})
+    assert signatures[0][0] == signatures[1][0]
+    pd.testing.assert_frame_equal(signatures[0][1], signatures[1][1])
+    if deep:
+        with h5py.File(ROOT / derivatives[0]["path"], "r") as base, h5py.File(ROOT / derivatives[1]["path"], "r") as features:
+            def compare(name, obj):
+                if not name.startswith("raw_") or not isinstance(obj, h5py.Dataset):
+                    return
+                other = features["obsm/" + name]
+                assert obj.shape == other.shape and obj.dtype == other.dtype
+                # At most 32 MiB per input for the large one-dimensional buffers.
+                block = max(1, 32 * 1024**2 // max(obj.dtype.itemsize, 1) // int(np.prod(obj.shape[1:])))
+                for start in range(0, obj.shape[0], block):
+                    assert np.array_equal(obj[start:start + block], other[start:start + block], equal_nan=True), name
+            base["obsm"].visititems(compare)
+    return {"checked_at_utc": datetime.now(timezone.utc).isoformat(), "archives": archives,
+            "derivatives": derivatives, "all_raw_buffers_identical_base_vs_features": deep,
+            "scope": "ZIP identities, generation/validation reports, all construction sidecars, all archived H5AD annotation bytes, snapshot joins and all waveform lengths; optional complete base/feature raw-buffer comparison. No re-synthesis or independent physical-model validation."}
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--deep", action="store_true")
+    args = parser.parse_args()
+    result = audit(args.deep)
+    output = ROOT / "docs/provenance/current_dataset.json"
+    output.parent.mkdir(parents=True, exist_ok=True)
+    output.write_text(json.dumps(result, indent=2) + "\n")
+    print(f"Verified current ZIPs and derivatives: {output}")

tools/build_report.py

@@ -0,0 +1,115 @@
+"""Build the multi-file LaTeX report; optionally promote a complete EDA run."""
+from __future__ import annotations
+
+import argparse
+import hashlib
+import json
+import re
+import shutil
+import subprocess
+import tempfile
+from pathlib import Path
+
+import pandas as pd
+
+ROOT = Path(__file__).resolve().parents[1]
+
+
+def digest(path):
+    return hashlib.sha256(path.read_bytes()).hexdigest()
+
+
+def refresh_assets():
+    notebook = ROOT / "notebooks/DRMHB_EDA.ipynb"
+    nb = json.loads(notebook.read_text())
+    code = [c for c in nb["cells"] if c["cell_type"] == "code" and "".join(c["source"]).strip()]
+    counts = [c.get("execution_count") for c in code]
+    if any(v is None for v in counts) or counts != list(range(counts[0], counts[0] + len(counts))):
+        raise RuntimeError("Promote a complete, sequential notebook execution")
+    if any(o.get("output_type") == "error" for c in code for o in c.get("outputs", [])):
+        raise RuntimeError("Notebook contains execution errors")
+    audit = json.loads((ROOT / "docs/provenance/current_dataset.json").read_text())
+    for entry in audit["archives"] + audit["derivatives"]:
+        if (ROOT / entry["path"]).stat().st_size != entry["size_bytes"]:
+            raise RuntimeError("Dataset changed since audit: " + entry["path"])
+    nb["metadata"]["drmhb_provenance"] = {
+        "execution_status": "complete_current_campaign_execution",
+        "dataset_generator_version": "2.6.0",
+        "workflow_package_version": "0.3.0",
+        "source_archive": audit["archives"][0]["path"],
+        "source_archive_sha256": audit["archives"][0]["sha256"],
+        "dataset_audit": "docs/provenance/current_dataset.json",
+        "note": "All code cells executed sequentially without errors; complete current EDA promoted to the report.",
+    }
+    notebook.write_text(json.dumps(nb, ensure_ascii=False, indent=1) + "\n")
+    source = ROOT / "derived/eda_v0.3.0"
+    manifest = pd.read_csv(source / "figures/manifest.csv").set_index("slug")
+    tex = (ROOT / "latexdocs/DRMHB.tex").read_text()
+    slugs = re.findall(r"\\edaplot\{[^}]+\}\{([^}]+)\}", tex)
+    records = []
+    for slug in slugs:
+        row = manifest.loc[slug]
+        if row["package_version"] != "0.3.0" or Path(row["source_vibframe"]).name != "DRMHB-compact.vibframe.zip":
+            raise RuntimeError("Unexpected EDA source: " + slug)
+        origin = ROOT / row["png"]
+        target = ROOT / "latexdocs/figures/eda" / origin.name
+        shutil.copyfile(origin, target)
+        cell = next((i for i, c in enumerate(nb["cells"]) if c["cell_type"] == "code" and f'"{slug}"' in "".join(c["source"])), None)
+        if cell is None:
+            # Mechanism-specific chronology plots use a dynamic fault-type slug.
+            cell = next(i for i, c in enumerate(nb["cells"]) if "fault_chronology_" in "".join(c["source"]))
+        records.append({"slug": slug, "path": str(target.relative_to(ROOT)), "sha256": digest(target),
+                        "notebook_cell": cell, "execution_count": nb["cells"][cell]["execution_count"],
+                        "question": row["question"], "origin": row["png"]})
+    for name in ("baseline_ml_summary", "baseline_detection_delay"):
+        origin = source / "tables" / (name + ".tex")
+        target = ROOT / "latexdocs" / origin.name
+        # Preserve the report's narrow headers while updating all numerical rows.
+        old, new = target.read_text(), origin.read_text()
+        rows = new.split("\\midrule\n", 1)[1].split("\\bottomrule", 1)[0]
+        target.write_text(old.split("\\midrule\n", 1)[0] + "\\midrule\n" + rows + "\\bottomrule" + old.split("\\bottomrule", 1)[1])
+        records.append({"slug": name, "path": str(target.relative_to(ROOT)), "sha256": digest(target),
+                        "origin": str(origin.relative_to(ROOT))})
+    evidence = {"dataset_generator_version": "2.6.0", "executed_conversion_version": "0.3.0",
+                "notebook": str(notebook.relative_to(ROOT)), "notebook_sha256": digest(notebook),
+                "recomputed": True, "dataset_audit": "docs/provenance/current_dataset.json",
+                "source_archives": audit["archives"], "assets": records}
+    path = ROOT / "latexdocs/provenance/eda_assets.json"
+    path.write_text(json.dumps(evidence, indent=2) + "\n")
+    print(f"Promoted {len(slugs)} figures and two baseline tables")
+
+
+def build():
+    report = ROOT / "docs/DRMHB.pdf"
+    # A failed build must leave the previous public PDF intact.
+    with tempfile.TemporaryDirectory(prefix="drmhb-latex-") as staging:
+        result = subprocess.run(["latexmk", "-pdf", "-interaction=nonstopmode", "-halt-on-error", f"-outdir={staging}", "DRMHB.tex"],
+                                cwd=ROOT / "latexdocs", stdout=subprocess.PIPE, stderr=subprocess.STDOUT, text=True)
+        if result.returncode:
+            raise RuntimeError(result.stdout)
+        shutil.copyfile(Path(staging) / "DRMHB.log", ROOT / "docs/DRMHB.log")
+        temporary = report.with_suffix(".pdf.tmp")
+        shutil.copyfile(Path(staging) / "DRMHB.pdf", temporary)
+        temporary.replace(report)
+    inputs = list((ROOT / "latexdocs").rglob("*.tex")) + list((ROOT / "latexdocs/figures").rglob("*.png"))
+    inputs = [p for p in inputs if "provenance" not in p.parts]
+    audit = json.loads((ROOT / "docs/provenance/current_dataset.json").read_text())
+    evidence = {"dataset_generator_version": "2.6.0", "conversion_package": "vibframe-anndata==0.3.0",
+                "dataset_audit": "docs/provenance/current_dataset.json",
+                "dataset_audit_sha256": digest(ROOT / "docs/provenance/current_dataset.json"),
+                "eda_manifest_sha256": digest(ROOT / "latexdocs/provenance/eda_assets.json"),
+                "archives": audit["archives"],
+                "report": {"path": "docs/DRMHB.pdf", "size_bytes": report.stat().st_size, "sha256": digest(report)},
+                "inputs": [{"path": str(p.relative_to(ROOT)), "sha256": digest(p)} for p in sorted(inputs)],
+                "build_command": "python tools/build_report.py", "scope": "Multi-file LaTeX project, built with latexmk"}
+    (ROOT / "latexdocs/provenance/sources.json").write_text(json.dumps(evidence, indent=2) + "\n")
+    print(f"Built {report}")
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--refresh-assets", action="store_true")
+    args = parser.parse_args()
+    if args.refresh_assets:
+        refresh_assets()
+    build()