feat(repo)!: retire the summarizer package and LLM symbol-summary feature#284
Open
theagenticguy wants to merge 1 commit into
Open
feat(repo)!: retire the summarizer package and LLM symbol-summary feature#284theagenticguy wants to merge 1 commit into
theagenticguy wants to merge 1 commit into
Conversation
…ture Removes the LLM code-summarization capability end to end. A capable coding agent regenerates task-conditioned summaries better than an index-time generic summary, so OpenCodeHub now hands agents deterministic structure (locations, edges, blast-radius, owners, license facts) and leaves prose to the agent. Removing summaries also drops the one non-deterministic ingestion phase, so every OpenCodeHub output is now re-derivable. BREAKING CHANGE: removes the `@opencodehub/summarizer` package, the ingestion `summarize` phase, the `symbol_summaries` storage table + SymbolSummaryRow type + the four SymbolSummary ITemporalStore methods, MCP + CLI query summary-fusion, the `codehub analyze` flags `--summaries` / `--no-summaries` / `--max-summaries` / `--summary-model` (and `CODEHUB_BEDROCK_SUMMARIES` / `CODEHUB_BEDROCK_DISABLED` env vars), and the `codehub wiki` `--llm` / `--max-llm-calls` / `--llm-model` flags. `codehub wiki` is now deterministic-only. Notes: - Summaries never entered graphHash/packHash, so the determinism + parity contract is unchanged (parity suite stays green). - The embeddings phase gains a direct dep on `confidence-demote` to preserve its canonical last-in-pipeline ordering (previously arrived transitively via summarize). - Drops the now-orphaned `@aws-sdk/client-bedrock-runtime` dependency from `@opencodehub/ingestion` and `@opencodehub/wiki`.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Retires the LLM code-summarization capability end to end. A capable coding agent regenerates task-conditioned summaries better than an index-time generic summary, so OpenCodeHub now hands agents deterministic structure (locations, edges, blast-radius, owners, license facts) and leaves prose to the agent. Removing summaries also drops the one non-deterministic ingestion phase, so every OpenCodeHub output is now re-derivable.
Removed (BREAKING)
@opencodehub/summarizerpackage — deleted entirely.summarizephase — removed from the default phase set.symbol_summariestable + DDL,SymbolSummaryRow, and the fourSymbolSummary*ITemporalStoremethods (bulkLoad/lookup/lookupByNode/count).queryand CLIquerysummary-fusion (the join, thesummary/signatureSummaryfields, thesummaries_joinedprobe).codehub analyze --summaries/--no-summaries/--max-summaries/--summary-model(and theCODEHUB_BEDROCK_SUMMARIES/CODEHUB_BEDROCK_DISABLEDenv vars);codehub wiki --llm/--max-llm-calls/--llm-model.codehub wikiis now deterministic-only.@aws-sdk/client-bedrock-runtimefrom@opencodehub/ingestionand@opencodehub/wiki.summarization-and-fusion.mdarchitecture page, ~15 adjacent doc references, thetsconfigproject refs, and thesummarizercommitlint scope.Determinism is unaffected
Summaries never entered
graphHash/packHash(seestorage/interface.ts— "Summaries never participate in [the hash]"), so the determinism + cross-adapter parity contract is unchanged. The parity suite stays green.Notable
The embeddings phase gains a direct dep on
confidence-demoteto preserve its canonical last-in-pipeline ordering (it must observe the final reconciled graph). That ordering previously arrived transitively via the now-removedsummarizephase; pinning it directly is behavior-preserving and covered by the phase-order regression test.Verification
graphHashparity suite.biome check, typecheck (CI filter, docs excluded perci.yml), andbanned-stringsall pass.Diff: 72 files, +145 / −8266.