KMS/dify
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix empty results issue in full-text search with Milvus vector database (#14885)

Browse Source 
Co-authored-by: liusurong.lsr <liusurong.lsr@alibaba-inc.com>
This commit is contained in:
llinvokerl 2025-03-05 12:27:01 +08:00 committed by GitHub
parent 9ab4f35b84
commit d04f40c274
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 11 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
api/core/rag/datasource/vdb/milvus/milvus_vector.py
View File

@ -72,8 +72,18 @@ class MilvusVector(BaseVector):
self._client = self._init_client(config) self._client = self._init_client(config)
self._consistency_level = "Session" # Consistency level for Milvus operations self._consistency_level = "Session" # Consistency level for Milvus operations
self._fields: list[str] = [] # List of fields in the collection self._fields: list[str] = [] # List of fields in the collection
if self._client.has_collection(collection_name):
self._load_collection_fields()
self._hybrid_search_enabled = self._check_hybrid_search_support() # Check if hybrid search is supported self._hybrid_search_enabled = self._check_hybrid_search_support() # Check if hybrid search is supported
def _load_collection_fields(self, fields: Optional[list[str]] = None) -> None:
if fields is None:
# Load collection fields from remote server
collection_info = self._client.describe_collection(self._collection_name)
fields = [field["name"] for field in collection_info["fields"]]
# Since primary field is auto-id, no need to track it
self._fields = [f for f in fields if f != Field.PRIMARY_KEY.value]
def _check_hybrid_search_support(self) -> bool: def _check_hybrid_search_support(self) -> bool:
""" """
Check if the current Milvus version supports hybrid search. Check if the current Milvus version supports hybrid search.
@ -306,10 +316,7 @@ class MilvusVector(BaseVector):
) )
schema.add_function(bm25_function) schema.add_function(bm25_function)
for x in schema.fields: self._load_collection_fields([f.name for f in schema.fields])
self._fields.append(x.name)
# Since primary field is auto-id, no need to track it
self._fields.remove(Field.PRIMARY_KEY.value)
# Create Index params for the collection # Create Index params for the collection
index_params_obj = IndexParams() index_params_obj = IndexParams()