Skip to content

fix: deduplicate embed_multi_with_metadata by content hash#1405

Open
atharvasingh7007 wants to merge 1 commit intosimonw:mainfrom
atharvasingh7007:fix/embed-multi-content-hash-dedup
Open

fix: deduplicate embed_multi_with_metadata by content hash#1405
atharvasingh7007 wants to merge 1 commit intosimonw:mainfrom
atharvasingh7007:fix/embed-multi-content-hash-dedup

Conversation

@atharvasingh7007
Copy link
Copy Markdown

Summary

  • deduplicate embed_multi_with_metadata() using existing content_hash values instead of existing item IDs
  • align multi-item embedding behavior with the existing single-item embed() path
  • add regression coverage for duplicate content submitted under a new ID

Testing

  • .venv\Scripts\python -m pytest tests/test_embed.py -k embed_multi
  • .venv\Scripts\python -m pytest tests/test_embed.py
  • .venv\Scripts\python -m ruff check .
  • .venv\Scripts\python -m mypy llm

Closes #1397

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

embed_multi_with_metadata() compares item IDs against row IDs instead of content hashes, breaking deduplication

1 participant