src/rag_components.py

from langchain.vectorstores import Chroma
from langchain_huggingface import HuggingFaceEmbeddings
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.document_loaders import TextLoader
from langchain_huggingface import HuggingFacePipeline
from langchain.chains import RetrievalQA
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline

def load_documents(file_path: str):
    """Loads documents from a specified file path."""
    loader = TextLoader(file_path)
    return loader.load()

def split_documents(documents, chunk_size=500, chunk_overlap=50):
    """Splits documents into chunks."""
    splitter = RecursiveCharacterTextSplitter(chunk_size=chunk_size, chunk_overlap=chunk_overlap)
    return splitter.split_documents(documents)

def create_embeddings(model_name="sentence-transformers/all-MiniLM-L6-v2"):
    """Creates HuggingFace embeddings."""
    return HuggingFaceEmbeddings(model_name=model_name)

def setup_vector_store(docs, embeddings, persist_directory="./chroma_db"):
    """Sets up and persists the Chroma vector store."""
    db = Chroma.from_documents(docs, embeddings, persist_directory=persist_directory)
    return db.as_retriever()

def create_qa_chain(retriever, model_name="TinyLlama/TinyLlama-1.1B-Chat-v1.0"):
    """Creates the RetrievalQA chain with streaming capabilities."""
    tokenizer = AutoTokenizer.from_pretrained(model_name)
    model = AutoModelForCausalLM.from_pretrained(
        model_name,
        device_map="auto",
    )
    pipe = pipeline(
        "text-generation",
        model=model,
        tokenizer=tokenizer,
        max_new_tokens=512,
        temperature=0.7,
        top_p=0.9
    )
    llm = HuggingFacePipeline(pipeline=pipe)

    qa_chain = RetrievalQA.from_chain_type(
        llm=llm,
        retriever=retriever,
        chain_type="stuff",
        return_source_documents=True # Added to potentially help with streaming or understanding context
    )
    return qa_chain