feat: implement localized relationship predicate extraction for graph edges (#799)#803
Open
suhaniiz wants to merge 1 commit into
Open
feat: implement localized relationship predicate extraction for graph edges (#799)#803suhaniiz wants to merge 1 commit into
suhaniiz wants to merge 1 commit into
Conversation
|
@suhaniiz is attempting to deploy a commit to the param20h's projects Team on Vercel. A member of the Team first needs to authorize it. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
📋 PR Checklist
🔗 Related Issue
Closes #799
📝 What does this PR do?
Implements a direct local relationship extraction feature (predicate labeling) for the knowledge graph building pipeline.
Previously, the GraphRAG implementation built a standard co-occurrence network, incrementing an edge weight but dropping all semantic contextual information explaining why those entities appeared together. This limited the accuracy of downstream LLM RAG queries.
This PR introduces an
extract_relationship_predicate()utility that isolates the textual span between matching entities within a given text chunk. This semantic context description is stored directly within a mutablerelationshipsset on graph edges, which is safely serialized into clean JSON arrays during the graph sanitization phase.🗂️ Type of Change
🧪 How was this tested?
"relationships": [...]list string arrays.📸 Screenshots (if UI change)
N/A (Backend feature)
The predicate window extraction features basic boundary slicing and multi-space scrubbing fallbacks. It runs locally without needing an external LLM, maintaining low ingestion costs and fast building loop benchmarks.
✅ Self-Review Checklist
dev, notmainmainbranch or any HuggingFace deployment config