[SOLVED] Is there a way to load a saved SKLearn VectorStore using langchain?

Is there a way to load a saved SKLearn VectorStore using langchain?

I created and saved a vectorstore using langchain_community.vectorstores SKLearnVectorStore and I can't load it.

I created and saved vectorstore as below:

from langchain_community.vectorstores import SKLearnVectorStore

vectorstore = SKLearnVectorStore.from_texts(
    texts=doc_splits,
    embedding=OllamaEmbeddings(model="Gemma-2:9b"),
    persist_path="assets/vectorstore",
)
vectorstore.persist()

I want to use this vectorstore in another file how do I use it?

Solution

Save the Vectors

vectorstore = SKLearnVectorStore.from_documents(
        documents=doc_splits,
        persist_path=PERSIST_PATH,
        embedding=OllamaEmbeddings(model="Gemma-2:9b"),
        serializer="parquet",
    )

vectorstore.persist()

Load the Saved parquet file

vectorstore = SKLearnVectorStore(
            persist_path=PERSIST_PATH,
            embedding=OllamaEmbeddings(model="Gemma-2:9b"),
            serializer="parquet"
        )
docs = vectorstore.similarity_search(query)

*Note: PERSIST_PATH is the path where you would like to save the file and load it.

Refer: https://python.langchain.com/docs/integrations/vectorstores/sklearn/*