phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-29 03:01:58 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix: sqlite_vec keyword implementation

Browse source 
Signed-off-by: Varsha Prasad Narsing <varshaprasad96@gmail.com>
This commit is contained in:
Varsha Prasad Narsing 2025-05-07 16:05:25 -07:00
parent e2a7022d3c
commit 2060fdba7f
14 changed files with 146 additions and 101 deletions
Download patch file Download diff file Expand all files Collapse all files

17
llama_stack/providers/utils/memory/vector_store.py
View file

@ -177,9 +177,11 @@ class EmbeddingIndex(ABC):
raise NotImplementedError()
@abstractmethod
async def query(
self, embedding: NDArray, query_string: Optional[str], k: int, score_threshold: float, mode: Optional[str]
) -> QueryChunksResponse:
async def query_vector(self, embedding: NDArray, k: int, score_threshold: float) -> QueryChunksResponse:
raise NotImplementedError()
@abstractmethod
async def query_keyword(self, query_string: str | None, k: int, score_threshold: float) -> QueryChunksResponse:
raise NotImplementedError()
@abstractmethod
@ -215,6 +217,9 @@ class VectorDBWithIndex:
mode = params.get("mode")
score_threshold = params.get("score_threshold", 0.0)
query_string = interleaved_content_as_str(query)
embeddings_response = await self.inference_api.embeddings(self.vector_db.embedding_model, [query_string])
query_vector = np.array(embeddings_response.embeddings[0], dtype=np.float32)
return await self.index.query(query_vector, query_string, k, score_threshold, mode)
if mode == "keyword":
return await self.index.query_keyword(query_string, k, score_threshold)
else:
embeddings_response = await self.inference_api.embeddings(self.vector_db.embedding_model, [query_string])
query_vector = np.array(embeddings_response.embeddings[0], dtype=np.float32)
return await self.index.query_vector(query_vector, k, score_threshold)