Skip to content

refactor(search-core)!: reduce chunk limits and remove embedding aliases#18

Merged
lurkshark merged 1 commit intomainfrom
lowertown
Mar 5, 2026
Merged

refactor(search-core)!: reduce chunk limits and remove embedding aliases#18
lurkshark merged 1 commit intomainfrom
lowertown

Conversation

@lurkshark
Copy link
Owner

Set chunking defaults to TARGET_WORDS=512 and MAX_WORDS=1024 to lower embedding memory pressure on MPS. Remove SearchService alias methods chunk_document and synchronize_document_chunks, and update tests/callers to canonical embedding APIs.

Archive the completed OpenSpec change for chunk-limit tuning and alias removal.

BREAKING CHANGE: SearchService no longer provides chunk_document or synchronize_document_chunks; use embed_document_chunks and synchronize_document_embeddings instead.

Set chunking defaults to TARGET_WORDS=512 and MAX_WORDS=1024 to lower embedding memory pressure on MPS. Remove SearchService alias methods chunk_document and synchronize_document_chunks, and update tests/callers to canonical embedding APIs.

Archive the completed OpenSpec change for chunk-limit tuning and alias removal.

BREAKING CHANGE: SearchService no longer provides chunk_document or synchronize_document_chunks; use embed_document_chunks and synchronize_document_embeddings instead.
@lurkshark lurkshark merged commit 9175456 into main Mar 5, 2026
1 check passed
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.

1 participant