embedding_storage.py

from langchain_openai import OpenAIEmbeddings
from langchain_chroma import Chroma
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.docstore.document import Document
import os

from config import PERSIST_DIRECTORY

def process_safety_with_chroma(text):
    """
    Processes and stores the given text into ChromaDB.

    Args:
        text (str): Text to be embedded and stored.

    Returns:
        Chroma: The Chroma vector store object.
    """
    if os.path.exists(PERSIST_DIRECTORY):
        vector_store = Chroma(persist_directory=PERSIST_DIRECTORY, embedding_function=OpenAIEmbeddings())
    else:
        text_splitter = RecursiveCharacterTextSplitter(chunk_size=2000, chunk_overlap=100)
        text_chunks = text_splitter.split_text(text)
        documents = [Document(page_content=chunk, metadata={"source": f"chunk_{i}"}) for i, chunk in enumerate(text_chunks)]
        
        embeddings = OpenAIEmbeddings()
        vector_store = Chroma.from_documents(documents, embeddings, persist_directory=PERSIST_DIRECTORY)

    return vector_store