Is there any method in langchain PGVector to search for all documents in a collection?

I need to search for all documents in a collection using the langchain PGVector lib

Upvotes: 0

Views: 274

Answers (1)

alchan
alchan

Reputation: 11

No, there is no supported method.

In ChromaDB, there is a get method, but it has not been implemented in PGVector yet.

If necessary, it can be implemented as shown below.

from langchain_openai import OpenAIEmbeddings
from langchain_postgres import PGVector
from sqlalchemy.orm import Session

embeddings = OpenAIEmbeddings(
    model="...",
    openai_api_key="..."
)

vectorstore = PGVector(
    embeddings=embeddings, collection_name="mobile_plan",
    connection="postgresql+psycopg://postgres:mypassword@localhost/mydb", use_jsonb=True)

with Session(vectorstore.session_maker.bind) as session:
    docs = session.query(vectorstore.EmbeddingStore.document).all()

print(docs[0])

print(len(docs))

Upvotes: 0

Related Questions