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

remove .value (#26633)

Browse Source 
Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
This commit is contained in:
Asuka Minato
2025-10-11 10:08:29 +09:00
committed by GitHub
parent bb6a331490
commit 1bd621f819
138 changed files with 613 additions and 633 deletions
Download Patch File Download Diff File Expand all files Collapse all files

40
api/core/rag/datasource/vdb/milvus/milvus_vector.py
View File

@@ -85,7 +85,7 @@ class MilvusVector(BaseVector):
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]
self._fields = [f for f in fields if f != Field.PRIMARY_KEY]
def _check_hybrid_search_support(self) -> bool:
"""
@@ -130,9 +130,9 @@ class MilvusVector(BaseVector):
insert_dict = {
# Do not need to insert the sparse_vector field separately, as the text_bm25_emb
# function will automatically convert the native text into a sparse vector for us.
Field.CONTENT_KEY.value: documents[i].page_content,
Field.VECTOR.value: embeddings[i],
Field.METADATA_KEY.value: documents[i].metadata,
Field.CONTENT_KEY: documents[i].page_content,
Field.VECTOR: embeddings[i],
Field.METADATA_KEY: documents[i].metadata,
}
insert_dict_list.append(insert_dict)
# Total insert count
@@ -243,15 +243,15 @@ class MilvusVector(BaseVector):
results = self._client.search(
collection_name=self._collection_name,
data=[query_vector],
anns_field=Field.VECTOR.value,
anns_field=Field.VECTOR,
limit=kwargs.get("top_k", 4),
output_fields=[Field.CONTENT_KEY.value, Field.METADATA_KEY.value],
output_fields=[Field.CONTENT_KEY, Field.METADATA_KEY],
filter=filter,
)
return self._process_search_results(
results,
output_fields=[Field.CONTENT_KEY.value, Field.METADATA_KEY.value],
output_fields=[Field.CONTENT_KEY, Field.METADATA_KEY],
score_threshold=float(kwargs.get("score_threshold") or 0.0),
)
@@ -264,7 +264,7 @@ class MilvusVector(BaseVector):
"Full-text search is disabled: set MILVUS_ENABLE_HYBRID_SEARCH=true (requires Milvus >= 2.5.0)."
)
return []
if not self.field_exists(Field.SPARSE_VECTOR.value):
if not self.field_exists(Field.SPARSE_VECTOR):
logger.warning(
"Full-text search unavailable: collection missing 'sparse_vector' field; "
"recreate the collection after enabling MILVUS_ENABLE_HYBRID_SEARCH to add BM25 sparse index."
@@ -279,15 +279,15 @@ class MilvusVector(BaseVector):
results = self._client.search(
collection_name=self._collection_name,
data=[query],
anns_field=Field.SPARSE_VECTOR.value,
anns_field=Field.SPARSE_VECTOR,
limit=kwargs.get("top_k", 4),
output_fields=[Field.CONTENT_KEY.value, Field.METADATA_KEY.value],
output_fields=[Field.CONTENT_KEY, Field.METADATA_KEY],
filter=filter,
)
return self._process_search_results(
results,
output_fields=[Field.CONTENT_KEY.value, Field.METADATA_KEY.value],
output_fields=[Field.CONTENT_KEY, Field.METADATA_KEY],
score_threshold=float(kwargs.get("score_threshold") or 0.0),
)
@@ -311,7 +311,7 @@ class MilvusVector(BaseVector):
dim = len(embeddings[0])
fields = []
if metadatas:
fields.append(FieldSchema(Field.METADATA_KEY.value, DataType.JSON, max_length=65_535))
fields.append(FieldSchema(Field.METADATA_KEY, DataType.JSON, max_length=65_535))
# Create the text field, enable_analyzer will be set True to support milvus automatically
# transfer text to sparse_vector, reference: https://milvus.io/docs/full-text-search.md
@@ -326,15 +326,15 @@ class MilvusVector(BaseVector):
):
content_field_kwargs["analyzer_params"] = self._client_config.analyzer_params
fields.append(FieldSchema(Field.CONTENT_KEY.value, DataType.VARCHAR, **content_field_kwargs))
fields.append(FieldSchema(Field.CONTENT_KEY, DataType.VARCHAR, **content_field_kwargs))
# Create the primary key field
fields.append(FieldSchema(Field.PRIMARY_KEY.value, DataType.INT64, is_primary=True, auto_id=True))
fields.append(FieldSchema(Field.PRIMARY_KEY, DataType.INT64, is_primary=True, auto_id=True))
# Create the vector field, supports binary or float vectors
fields.append(FieldSchema(Field.VECTOR.value, infer_dtype_bydata(embeddings[0]), dim=dim))
fields.append(FieldSchema(Field.VECTOR, infer_dtype_bydata(embeddings[0]), dim=dim))
# Create Sparse Vector Index for the collection
if self._hybrid_search_enabled:
fields.append(FieldSchema(Field.SPARSE_VECTOR.value, DataType.SPARSE_FLOAT_VECTOR))
fields.append(FieldSchema(Field.SPARSE_VECTOR, DataType.SPARSE_FLOAT_VECTOR))
schema = CollectionSchema(fields)
@@ -342,8 +342,8 @@ class MilvusVector(BaseVector):
if self._hybrid_search_enabled:
bm25_function = Function(
name="text_bm25_emb",
input_field_names=[Field.CONTENT_KEY.value],
output_field_names=[Field.SPARSE_VECTOR.value],
input_field_names=[Field.CONTENT_KEY],
output_field_names=[Field.SPARSE_VECTOR],
function_type=FunctionType.BM25,
)
schema.add_function(bm25_function)
@@ -352,12 +352,12 @@ class MilvusVector(BaseVector):
# Create Index params for the collection
index_params_obj = IndexParams()
index_params_obj.add_index(field_name=Field.VECTOR.value, **index_params)
index_params_obj.add_index(field_name=Field.VECTOR, **index_params)
# Create Sparse Vector Index for the collection
if self._hybrid_search_enabled:
index_params_obj.add_index(
field_name=Field.SPARSE_VECTOR.value, index_type="AUTOINDEX", metric_type="BM25"
field_name=Field.SPARSE_VECTOR, index_type="AUTOINDEX", metric_type="BM25"
)
# Create the collection