Skip to content

Various updates clean#50

Merged
m1rl0k merged 5 commits intotestfrom
various-updates-clean
Dec 10, 2025
Merged

Various updates clean#50
m1rl0k merged 5 commits intotestfrom
various-updates-clean

Conversation

@m1rl0k
Copy link
Collaborator

@m1rl0k m1rl0k commented Dec 9, 2025

No description provided.

Introduces connection pooling for QdrantClient and uses orjson for faster JSON serialization when available. Adds a shared ThreadPoolExecutor for parallel queries and implements a cache for filter sanitization to improve performance.
Adds 'orjson' to requirements. Improves hybrid_search.py with cached regex compilation, filter object caching, and parallel dense query execution. In ingest_code.py, increases default batch size for indexing. In rerank_local.py, adds background warmup for ONNX reranker session to reduce initial latency.
Introduces list_workspaces to enumerate workspaces and their states by scanning for state.json files, supporting both single and multi-repo modes. Adds _invalidate_router_scratchpad stub to allow router cache invalidation after indexing. Fixes progress reporting in ingest_code.py and minor logging improvements in mcp_indexer_server.py.
batch_ids: list[int] = []
batch_lex: list[list[float]] = []
BATCH_SIZE = int(os.environ.get("INDEX_BATCH_SIZE", "64") or 64)
BATCH_SIZE = int(os.environ.get("INDEX_BATCH_SIZE", "256") or 256)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@m1rl0k any reason you raised defaults?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This was a performance optimization - larger batches = fewer Qdrant upsert roundtrips = faster indexing. Trade-off is higher memory per batch... we can leave at default though

return unchanged, changed


def list_workspaces(search_root: Optional[str] = None) -> List[Dict[str, Any]]:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@m1rl0k Probably won't be working with client -> remote stack semantics

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yes only for local

@m1rl0k m1rl0k merged commit 614a0b7 into test Dec 10, 2025
1 check passed
@m1rl0k m1rl0k deleted the various-updates-clean branch December 13, 2025 15:36
m1rl0k added a commit that referenced this pull request Mar 1, 2026
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.

2 participants