Searching similar vectors in SQLModel #57
Description
I want to build a function for fetching similar vectors with the distance metrics. I have used the exact command written in the README.md but looks like it's not working for me.
Here's the code:
async def get_similar_vectors(session: AsyncSession, query_vector, k=10, distance = "cosine_distance"):
if distance == "cosine_distance":
query = select(db_models.Vector).order_by(
db_models.Vector.embedding.cosine_distance(query_vector)).limit(k)
elif distance == "l2_distance":
query = select(db_models.Vector).order_by(
db_models.Vector.embedding.l2_distance(query_vector)).limit(k)
else:
query = select(db_models.Vector).order_by(
db_models.Vector.embedding.max_inner_product(query_vector)).limit(k)
execute_query = await session.execute(query)
similar_vectors = execute_query.all()
results = [
{"id": vector.id, "embedding": vector.embedding}
for vector in similar_vectors
]
return resultsHere's the issue:
select(db_models.Vector).order_by(
db_models.Vector.embedding.cosine_distance(query_vector)).limit(k)Cannot access member "cosine_distance" for type "List[float]"
Member "cosine_distance" is unknown
This goes same for all the other distance metrics.
To give more clarity on the Vector model:
class Vector(SQLModel, table=True):
id: Optional[UUID] = Field(default=None, primary_key=True)
embedding: List[float] = Field(sa_column=Column(Vector(embedding_dim)))query_vector type is List[float]
How to fix this issue? Or more specifically how to search for similar embeddings using SQLModel?