Run outputs, 2026-10-30

Everything a full run of the pipeline produced, so a clone can open the site, ask the analyst and run the whole test suite without the three-hour run.

• Run: started 2026-10-21T10:18:51+00:00, finished 2026-10-30T13:56:20+00:00; 47/47 steps OK, 319/319 tests passed
• Built from commit dbe1160fbcee
• darklytics-outputs-2026-10-30.zip: 80.6 MB, sha256 83951bdbdf9b8b5dc7e0f3ee0a221287e81e613cf73baea00da5cd0cb1cf4612

Use it: python scripts/fetch_outputs.py (checks the sha256, writes data/ml/ and outputs/).

Inside: the ML feature layer (data/ml) and outputs/ (models, analyst, audits, run log, built site). Observed and simulated results keep their tags in every file; see the README.

Run outputs, 2026-10-29

Everything a full run of the pipeline produced, so a clone can open the site, ask the analyst and run the whole test suite without the three-hour run.

• Run: started 2026-10-21T10:18:51+00:00, finished 2026-10-24T07:39:27+00:00; 42/42 steps OK, 286/286 tests passed
• Built from commit a21ca3dbb4ca
• darklytics-outputs-2026-10-29.zip: 75.3 MB, sha256 5713a51183d5af09cbe6e62b52f990d23b4c51c44ca265ad05c31e810e81fb3d

Use it: python scripts/fetch_outputs.py (checks the sha256, writes data/ml/ and outputs/).

Inside: the ML feature layer (data/ml) and outputs/ (models, analyst, audits, run log, built site). Observed and simulated results keep their tags in every file; see the README.

Run outputs, 2026-10-24

Everything a full run of the pipeline produced, so a clone can open the site, ask the analyst and run the whole test suite without the three-hour run.

• Run: started 2026-10-21T10:18:51+00:00, finished 2026-10-24T07:39:27+00:00; 42/42 steps OK, 286/286 tests passed
• Built from commit 3c53708abd03
• shelfpulse-outputs-2026-10-24.zip: 72.0 MB, sha256 34c3bf06fe2f90c9ae5bba1b7364a00cafb244955345cf00a655bd713a1117ac

Use it: python scripts/fetch_outputs.py (checks the sha256, writes data/ml/ and outputs/).

Inside: the ML feature layer (data/ml) and outputs/ (models, analyst, audits, run log, built site). Observed and simulated results keep their tags in every file; see the README.

Run outputs, 2026-10-24

Everything a full run of the pipeline produced, so a clone can open the site, ask the analyst and run the whole test suite without the three-hour run.

• Run: started 2026-10-21T10:18:51+00:00, finished 2026-10-24T07:39:27+00:00; 42/42 steps OK, 286/286 tests passed
• Built from commit 7589266e0ceb
• qcom-outputs-2026-10-24.zip: 72.0 MB, sha256 ebb2dbcfd09abe391c7d20a9b99b5d7c054f5d8940101b08af7c6167e95e2fcd

Use it: python scripts/fetch_outputs.py (checks the sha256, writes data/ml/ and outputs/).

Inside: the ML feature layer (data/ml) and outputs/ (models, analyst, audits, run log, built site). Observed and simulated results keep their tags in every file; see the README.