| 1 | # Recovered raw weekly evidence — W23–W29 (issue #569) |
| 2 | |
| 3 | This directory holds **raw weekly crawl evidence recovered unchanged** from |
| 4 | still-unexpired GitHub Actions `raw-data` artifacts. It is isolated under |
| 5 | `data/archive/` and explicitly provenance-marked; it is **not** a recrawl and |
| 6 | does not regenerate or substitute any data. |
| 7 | |
| 8 | ## Contents |
| 9 | |
| 10 | - `2026-W23/` … `2026-W29/` — the per-week raw payloads, byte-identical to the |
| 11 | files inside their source artifacts. |
| 12 | - `recovered-raw.sha256` — SHA-256 checksums of every recovered file. |
| 13 | - `provenance.json` — full inventory: for each week, the source **artifact ID**, |
| 14 | **source run ID**, artifact `head_sha`, created/expires timestamps, the |
| 15 | **original pipeline path** (`data/raw/…`), file size, and SHA-256. |
| 16 | |
| 17 | ## Provenance summary |
| 18 | |
| 19 | | Week | Artifact ID | Source run ID | Artifact created | Expires | |
| 20 | |------|-------------|---------------|------------------|---------| |
| 21 | | 2026-W23 | 7330965888 | 26753498571 | 2026-06-01 | 2026-08-30 | |
| 22 | | 2026-W24 | 7480423082 | 27137991437 | 2026-06-08 | 2026-09-06 | |
| 23 | | 2026-W25 | 7680608990 | 27650616699 | 2026-06-16 | 2026-09-14 | |
| 24 | | 2026-W26 | 7786268089 | 27937940184 | 2026-06-22 | 2026-09-20 | |
| 25 | | 2026-W27 | 7942942491 | 28349393843 | 2026-06-29 | 2026-09-27 | |
| 26 | | 2026-W28 | 8099693880 | 28767491281 | 2026-07-06 | 2026-10-04 | |
| 27 | | 2026-W29 | 8268630097 | 29222289174 | 2026-07-13 | 2026-10-11 | |
| 28 | |
| 29 | All seven weeks W23–W29 were recoverable; there are no expired/missing gaps. |
| 30 | The W29 artifact (`8268630097`, expires **2026-10-11**) is preserved here |
| 31 | durably before its expiry. |
| 32 | |
| 33 | ## Verify |
| 34 | |
| 35 | ```bash |
| 36 | cd data/archive/recovered-W23-W29 |
| 37 | sha256sum -c recovered-raw.sha256 |
| 38 | ``` |
| 39 | |
| 40 | Each recorded hash was also verified source-bound: a fresh download of every |
| 41 | source artifact was extracted and its target-week files hashed to match the |
| 42 | values recorded here. |