Some checks failed
CI / lint (push) Successful in 48s
CI / notebooks-smoke (push) Successful in 1m57s
Deploy / changes (push) Successful in 15s
CI / test (push) Successful in 3m52s
Infra CI / notebooks (push) Successful in 1m4s
Infra CI / zotero (push) Successful in 19s
Infra CI / api (push) Successful in 1m14s
Infra CI / docs (push) Successful in 1m25s
Infra CI / llm (push) Successful in 54s
Infra CI / mc (push) Successful in 19s
Deploy / notebooks (push) Has been skipped
Deploy / zotero (push) Has been skipped
Infra CI / bcda (push) Successful in 1m39s
Deploy / docs (push) Successful in 36s
Deploy / api (push) Successful in 1m16s
Deploy / report (push) Has been cancelled
Deploy / mc (push) Has been cancelled
Deploy / bcda (push) Has started running
Deploy / llm (push) Has started running
2.1 KiB
2.1 KiB
title, sidebar_position
| title | sidebar_position |
|---|---|
| stack llm restamp | 65 |
stack llm restamp
Usage: stack llm restamp [OPTIONS]
Backfill codes/families/elements onto already-indexed chunks.
Metadata-only: computes `pfs.anchors.anchor_metadata` from each row's
stored document text and merges it into `cmetadata` with a jsonb `||`,
never touching the embedding — no GPU time, no re-chunking. Safe to
re-run: rows already stamped with the current values are skipped.
F6: never rebuilds the HNSW index (`ensure_hnsw` ALTERs the vector
column and takes an ACCESS EXCLUSIVE lock for the life of the
rebuild) — only the cheap metadata GIN indexes, and not even those
under `--dry-run`.
╭─ Options ────────────────────────────────────────────────────────────────────╮
│ --collection <str> Which collection: comments | rules | corpus | │
│ all. │
│ [default: comments] │
│ --batch <int> Rows per transaction. [default: 2000] │
│ --dry-run Scan and report without writing. │
│ --help Show this message and exit. │
╰──────────────────────────────────────────────────────────────────────────────╯