gmakstutis/cognee
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix: Fix chunk naive llm classifier

Browse source 
Fixed chunk naive llm classifier uuid issue, added fix for deletion of data points for LanceDB

Fix #COG-472
This commit is contained in:
Igor Ilic 2024-10-31 00:42:18 +01:00
parent 50f5712f43
commit 3567e0d7e7
2 changed files with 5 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
cognee/infrastructure/databases/vector/lancedb/LanceDBAdapter.py
View file

@ -164,7 +164,10 @@ class LanceDBAdapter(VectorDBInterface):
async def delete_data_points(self, collection_name: str, data_point_ids: list[str]): async def delete_data_points(self, collection_name: str, data_point_ids: list[str]):
connection = await self.get_connection() connection = await self.get_connection()
collection = await connection.open_table(collection_name) collection = await connection.open_table(collection_name)
results = await collection.delete(f"id IN {tuple(data_point_ids)}") if len(data_point_ids) == 1:
results = await collection.delete(f"id = '{data_point_ids[0]}'")
else:
results = await collection.delete(f"id IN {tuple(data_point_ids)}")
return results return results
async def prune(self): async def prune(self):

3
cognee/tasks/chunk_naive_llm_classifier/chunk_naive_llm_classifier.py
View file

@ -21,7 +21,6 @@ async def chunk_naive_llm_classifier(data_chunks: list[DocumentChunk], classific
for chunk_index, chunk in enumerate(data_chunks): for chunk_index, chunk in enumerate(data_chunks):
chunk_classification = chunk_classifications[chunk_index] chunk_classification = chunk_classifications[chunk_index]
classification_data_points.append(uuid5(NAMESPACE_OID, chunk_classification.label.type)) classification_data_points.append(uuid5(NAMESPACE_OID, chunk_classification.label.type))
classification_data_points.append(uuid5(NAMESPACE_OID, chunk_classification.label.type))
for classification_subclass in chunk_classification.label.subclass: for classification_subclass in chunk_classification.label.subclass:
classification_data_points.append(uuid5(NAMESPACE_OID, classification_subclass.value)) classification_data_points.append(uuid5(NAMESPACE_OID, classification_subclass.value))
@ -39,7 +38,7 @@ async def chunk_naive_llm_classifier(data_chunks: list[DocumentChunk], classific
if await vector_engine.has_collection(collection_name): if await vector_engine.has_collection(collection_name):
existing_data_points = await vector_engine.retrieve( existing_data_points = await vector_engine.retrieve(
collection_name, collection_name,
list(set(classification_data_points)), [str(classification_data) for classification_data in list(set(classification_data_points))],
) if len(classification_data_points) > 0 else [] ) if len(classification_data_points) > 0 else []
existing_points_map = {point.id: True for point in existing_data_points} existing_points_map = {point.id: True for point in existing_data_points}