llm: incremental embedding indexer #567

Closed
opened 2026-07-16 19:26:08 +00:00 by kert · 1 comment
Owner

Batch runner: bib store → chunker → Ollama embeddings → pgvector upsert. 164,214 comments ≈ 0.5–1M chunks.

Scope

  • reads comments (doctype:comment) and corpus items from bib
  • content-hash resumable: re-runs embed only new/changed items
  • stack llm index --docket/--collection scoping; progress + throughput logging
  • v1 gate: one docket indexed end-to-end before corpus-wide run

Acceptance

  • kill/restart mid-run loses no work and re-embeds nothing already done
  • one full docket indexed and queryable

Spec: docs/superpowers/specs/2026-07-16-llm-module-design.md

Batch runner: bib store → chunker → Ollama embeddings → pgvector upsert. 164,214 comments ≈ 0.5–1M chunks. **Scope** - [ ] reads comments (`doctype:comment`) and corpus items from bib - [ ] content-hash resumable: re-runs embed only new/changed items - [ ] `stack llm index --docket/--collection` scoping; progress + throughput logging - [ ] v1 gate: one docket indexed end-to-end before corpus-wide run **Acceptance** - kill/restart mid-run loses no work and re-embeds nothing already done - one full docket indexed and queryable Spec: `docs/superpowers/specs/2026-07-16-llm-module-design.md`
kert added this to the P33: LLM Foundation — Ollama, pgvector, embedding index milestone 2026-07-16 19:26:08 +00:00
kert added the datallm labels 2026-07-16 19:26:08 +00:00
Author
Owner

Done in P33 — merged to main as 14d279b (branch llm-p33). Pilot docket CMS-2017-0092 indexed end-to-end with exact search; full test gate green at 99.75%.

Done in P33 — merged to main as 14d279b (branch llm-p33). Pilot docket CMS-2017-0092 indexed end-to-end with exact search; full test gate green at 99.75%.
kert closed this issue 2026-07-17 19:37:21 +00:00
Sign in to join this conversation.