knowledge_base/embeddings.py

from langchain_huggingface import HuggingFaceEmbeddings
from langchain_chroma import Chroma

# 1. Define the custom embedding object
dense_embeddings = HuggingFaceEmbeddings(
    model_name="sentence-transformers/all-mpnet-base-v2"
)

# 2. Initialize the LangChain Chroma vector store, passing the embeddings
vectorstore = Chroma.from_documents(
    documents=["./docs/markdowns"],  # Placeholder for actual documents
    embedding=dense_embeddings,
    collection_name="langchain_mpnet_collection",
    persist_directory="./knowledge_base/chroma_data" 
)

# 3. Save the database (essential for persistence)
vectorstore.persist()
print("LangChain Chroma vector store created with custom embeddings and persisted.")

if __name__ == "__main__":
    pass