fix: ingest stops reading the fixture format's own metadata as knowledge - #217
Merged
Conversation
/kb list showed 11 docs, five titled "doc-meta" — kb load-dir's sidecar files (doc-meta.json / chunks.jsonl) ingested as documents by the generic directory walk (#214). #207 fixed the sample_data_en case by splitting the directories but left the walk trusting layout; the scn_* examples still pair sidecars with source documents in one directory, so the same one command re-poisons a KB today. discover_files now skips the two reserved sidecar names the way it skips dotfiles: they describe the corpus, they aren't part of it. docs/assets/tui.gif is regenerated against a cleaned-up KB (the five leftover doc-meta documents hard-deleted via `opspilot kb delete`, audit-logged) — the /kb list frame now shows 6 real documents. Closes #214. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01KXrF6HMB12V1hC9xFtooDE
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #214.
What
/kb listshowed 11 docs, five titleddoc-meta—kb load-dir's sidecar files (doc-meta.json/chunks.jsonl) ingested as first-class documents byingest's directory walk.#207 fixed this for
examples/sample_data_enby splitting sources (kb/) from fixtures (fixtures/) — deliberately with no product-code change. But the walk still trusts layout: thescn_*examples pair sidecars with source documents in one directory to this day, soopspilot ingest examples/scn_ticket_summary_zh/kb/re-poisons a KB right now. The five rows in the demo gif were leftovers ingested before the #207 split.discover_filesnow skips the two reserved sidecar names exactly the way it skips dotfiles: they describe the corpus, they aren't part of it. New test covers the mixed-directory layout and the explicit-file case.Cleanup + artifact
doc-metadocuments were hard-deleted from the local KB viaopspilot kb delete(audit-logged asdel_22a2fe72…del_c133e42f).docs/assets/tui.gifregenerated: the/kb listframe now shows 6 doc(s), all real SOPs; the/wiki listframe confirms the fix: the TUI tour gif stops racing its own workers #213 race fix still holds.Verification
tests/test_ingestion.py: 23 passed (incl. the new sidecar test); ruff + mypy clean.🤖 Generated with Claude Code
https://claude.ai/code/session_01KXrF6HMB12V1hC9xFtooDE