Feat/auto extracted edges#246
Merged
plind-junior merged 1 commit intoJun 17, 2026
Merged
Conversation
approving a page now files relation proposals for its `[[wiki-links]]` (mentions), `entities` frontmatter (relates_to), and `sources` frontmatter (derived_from), tagged proposed_by: vouch-extractor. extraction calls back into propose_relation -- same review gate, endpoint validation, and audit trail as a hand-filed edge -- so a dangling wiki-link is skipped rather than blocking the page approval that triggered it. adds `vouch reject-extracted [--page <id>]` (+ mcp/jsonl equivalents) to mass-reject pending auto-extracted edges, scoped by proposed_by so it never touches a manually proposed relation. closes vouchdev#224
|
Warning Review limit reached
More reviews will be available in 10 minutes and 40 seconds. Learn how PR review limits work. Your organization has used up its prepaid credits, and credit purchases are no longer available. Enable the review add-on in the billing tab to keep reviews running — you're only billed for reviews past your plan's rate limits ($0.25/file). ⌛ How to resolve this issue?After more reviews become available, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans include higher PR review limits than trial, open-source, and free plans. In all cases, reviews become available again over time. During sustained high-volume PR review activity, CodeRabbit may temporarily slow when the next review becomes available. Please see our Fair Usage Limits Policy for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (144)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
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
2 participants
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 changed
Approving a page now files relation proposals for the edges implied by its content:
mentionsfor every[[wiki-link]]in the body,relates_tofor the page'sentitiesfrontmatter, andderived_fromfor itssourcesfrontmatter. Each is filed via the normalpropose_relationpath withproposed_by: vouch-extractor, so it lands inproposed/and needs a reviewer's approve/reject like any hand-filed edge. Addsvouch reject-extracted [--page <id>](+ MCPkb_reject_extracted/ JSONLkb.reject_extracted) to mass-reject pending auto-extracted edges, scoped byproposed_byso it never touches a manually proposed relation.Why
Closes #224. The
Relationmodel exists but the graph is hand-populated; approved pages already carry wiki-links and structuredentities/sourceslists that imply edges nobody files. gbrain's writeup cites a measurable retrieval win (+31.4 P@5) from doing this automatically with zero LLM calls — this is the same write-time pass, scoped to vouch's existing fields.What might break
RelationTypemembersmentionsandrelates_to(object model addition, additive/non-breaking for existing data).kb.reject_extractedmethod (additive; existingkb.*callers unaffected).[[wiki-link]](target that doesn't resolve to a claim/page/entity/source) is silently skipped rather than erroring — extraction is best-effort and must never block the page approval that triggered it.VEP
None filed yet. This touches the object model (
RelationTypeadditions) and thekb.*method surface (kb.reject_extracted), both of whichproposals/README.mdsays need an accepted VEP before implementation. Flagging for maintainer input — happy to split a VEP out first if that's the right order here.Tests
make checkpasses locally (ruff clean, mypy clean except a pre-existing unrelatedjinja2stub gap behind the[web]extra; full pytest pass except 4 pre-existing Windows-only failures — path-separator and symlink-privilege issues unrelated to this change)tests/test_extractors_edges.py: wiki-link extraction/dedup, mentions/relates_to/derived_from edge filing, dangling-link skip, no-links no-op, bulk reject scoping)CHANGELOG.mdupdated under## [Unreleased]