fix(retrieval): return no matches instead of erroring when the vector collection doesn't exist yet - #843
Open
EnjoyBacon7 wants to merge 1 commit into
Open
fix(retrieval): return no matches instead of erroring when the vector collection doesn't exist yet#843EnjoyBacon7 wants to merge 1 commit into
EnjoyBacon7 wants to merge 1 commit into
Conversation
… collection doesn't exist yet The shared Milvus collection is created lazily on the first indexed file system-wide. On a fresh deployment, or any query before that first index, search/multi_query_search raised CollectionNotExists -> VDBSearchError instead of returning an empty result — the same class of bug #508 already fixed for delete_partition.
📝 WalkthroughWalkthroughThe vector-store search methods now check collection existence before embedding or retrieval. Missing collections return empty results. Unit tests cover both search methods and update existing mocks for the collection-present path. ChangesVector collection guard
Possibly related PRs
Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches📝 Generate docstrings
🧪 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 |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@openrag/services/storage/vector_store_searcher.py`:
- Around line 72-74: Update the search flow in VectorStoreSearcher to handle a
collection being dropped between collection_exists() and search(). Prefer an
atomic store-level search operation; otherwise catch the store’s
missing-collection error around search and return [] while preserving
VDBSearchError translation for other failures.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro Plus
Run ID: 74edc257-317a-4801-9533-ae2ba9ffc389
📒 Files selected for processing (2)
openrag/services/storage/vector_store_searcher.pytests/unit/services/storage/test_vector_store_searcher.py
Comment on lines
+72
to
+74
| # Collection is created lazily on first insert (#508); absent means no chunks. | ||
| if not await self._store.collection_exists(self._collection): | ||
| return [] |
There was a problem hiding this comment.
🩺 Stability & Availability | 🟡 Minor | ⚡ Quick win
🧩 Analysis chain
🏁 Script executed:
#!/usr/bin/env bash
set -euo pipefail
# Expect either no concurrent collection-drop path or explicit missing-collection handling.
rg -n -C 8 'drop_collection|release_collection|has_collection|collection_exists|VDBSearchError' openrag testsRepository: linagora/openrag
Length of output: 50372
🏁 Script executed:
#!/usr/bin/env bash
set -euo pipefail
# Inspect the relevant searcher/store implementation and exception shape without running repo code.
sed -n '1,150p' openrag/services/storage/vector_store_searcher.py
printf '\n--- exceptions around VDBSearchError / VDBPartitionNotFound ---\n'
sed -n '35,46p' openrag/core/utils/exceptions.py
sed -n '400,412p' openrag/core/utils/exceptions.py
printf '\n--- Milvus search wrappers and collection drop/resolvers ---\n'
sed -n '800,1105p' openrag/services/storage/milvus_store.py
sed -n '880,980p' openrag/services/storage/milvus_store.pyRepository: linagora/openrag
Length of output: 23633
🏁 Script executed:
#!/usr/bin/env bash
set -euo pipefail
# Find all search_call sites around collection_exists and whether they catch missing-collection errors.
rg -n -C 4 'collection_exists\(.*collection|\.search\(|\.hybrid_search\(' openrag/services/storage/vector_store_searcher.py openrag/services/storage/milvus_store.py tests/integration/repos/test_milvus_store_integration.py tests/integration/repos/test_workspace_scoping_e2e.pyRepository: linagora/openrag
Length of output: 12540
Handle collection absence atomically in search.
collection_exists() and search() are separate calls. If drop_collection() runs concurrently, search() can still translate the missing collection failure to VDBSearchError. Make the store operation atomic, or catch the missing-collection state at VectorStoreSearcher and return an empty result.
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@openrag/services/storage/vector_store_searcher.py` around lines 72 - 74,
Update the search flow in VectorStoreSearcher to handle a collection being
dropped between collection_exists() and search(). Prefer an atomic store-level
search operation; otherwise catch the store’s missing-collection error around
search and return [] while preserving VDBSearchError translation for other
failures.
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.
What
search()/multi_query_search()onVectorStoreSearcherraise instead ofreturning an empty result when the shared Milvus collection doesn't exist yet.
Root cause
The shared collection (
config.vectordb.collection_name) is created lazilyon the first indexed file system-wide (
services/workers/stages/store.py).On a fresh deployment — or any retrieval-needing chat/search query before
that first file is indexed, regardless of partition — Milvus raises
CollectionNotExists, wrapped asVDBSearchError, and propagates as a realerror instead of "no matches."
This is the same class of bug #508 already fixed for
delete_partition(
PartitionService), using the existingVectorStore.collection_exists()port method as the guard. That fix only covered delete; search was still
exposed.
Fix
Mirror #508's pattern in
VectorStoreSearcher.search()andmulti_query_search(): checkcollection_exists()before embedding/searching, and return
[]immediately if the collection is absent — nocollection means no chunks to find, and skipping straight to an empty result
also avoids an unnecessary embedding-API call.
Test
Adds
test_search_returns_empty_when_collection_does_not_existandtest_multi_query_search_returns_empty_when_collection_does_not_exist,asserting the embedder and store are never called when the collection is
absent. Full unit suite: 2433 passed (1 pre-existing unrelated failure on
develop, unaffected by this change), ruff clean, layer-import guard clean.Summary by CodeRabbit
Bug Fixes
Tests