app.py

import streamlit as st
from dotenv import load_dotenv
import os
from pinecone import Pinecone
from langchain_google_genai import GoogleGenerativeAIEmbeddings
from langchain_pinecone import PineconeVectorStore
from langchain_core.documents import Document
from uuid import uuid4
from langchain.chains import RetrievalQA
from langchain.llms import HuggingFaceHub

# Load environment variables
load_dotenv()

# Get API keys from environment variables
pinecone_api_key = os.getenv("PINECONE_API_KEY")
google_api_key = os.getenv("GOOGLE_API_KEY")
huggingfacehub_api_token = os.getenv("HUGGINGFACEHUB_API_TOKEN")

# Check if API keys are available
if not pinecone_api_key or not google_api_key or not huggingfacehub_api_token:
    st.error("API keys not found. Please set PINECONE_API_KEY, GOOGLE_API_KEY, and HUGGINGFACEHUB_API_TOKEN in your .env file.")
    st.stop()

# Initialize Pinecone
pc = Pinecone(api_key=pinecone_api_key, environment="us-east1-gcp")  # Replace with your environment if needed
index_name = "online-rag"
index = pc.Index(index_name)

# Initialize embeddings
os.environ['GOOGLE_API_KEY'] = google_api_key
embeddings = GoogleGenerativeAIEmbeddings(model="models/text-embedding-004")

# Initialize vector store
vector_store = PineconeVectorStore(index=index, embedding=embeddings)

# Initialize LLaMA 30B model from Hugging Face
llm = HuggingFaceHub(repo_id="meta-llama/Llama-2-30b-chat-hf", huggingfacehub_api_token=huggingfacehub_api_token)

# Streamlit app
st.title("LLaMA 30B RAG Chatbot")

# Upload document
uploaded_file = st.file_uploader("Upload a document", type=["txt", "pdf"])

if uploaded_file is not None:
    # Read the file
    file_details = {"filename": uploaded_file.name, "filetype": uploaded_file.type, "filesize": uploaded_file.size}
    st.write(file_details)
    file_content = uploaded_file.read().decode("utf-8")

    # Create a document
    document = Document(page_content=file_content, metadata={"source": uploaded_file.name})

    # Add document to vector store
    uuids = [str(uuid4()) for _ in range(1)]
    vector_store.add_documents(documents=[document], ids=uuids)
    st.write("Document added to Pinecone.")

# Query the chatbot
query = st.text_input("Enter your query:")

if query:
    try:
        # Perform similarity search
        results = vector_store.similarity_search(query, k=2)
        st.write("Search Results:")
        for res in results:
            st.write(f"* {res.page_content} [{res.metadata}]")

        # Create a RetrievalQA chain
        qa_chain = RetrievalQA.from_chain_type(llm=llm, chain_type="stuff", retriever=vector_store.as_retriever())

        # Get the answer
        answer = qa_chain.run(query)
        st.write("Chatbot Response:")
        st.write(answer)

    except Exception as e:
        st.error(f"An error occurred: {e}")