app.py

import os
import gradio as gr
from langchain_community.vectorstores import FAISS
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_community.document_loaders import PyMuPDFLoader
from langchain_text_splitters import CharacterTextSplitter
from langchain.chains import RetrievalQA
from langchain_community.llms import HuggingFaceEndpoint  # Updated import
from huggingface_hub import login

# 1. Authentication
login(token=os.environ.get('HF_TOKEN'))

# 2. PDF Processing
def create_qa_system():
    if not os.path.exists("file.pdf"):
        raise gr.Error("❗ Upload file.pdf in Files tab")
    
    loader = PyMuPDFLoader("file.pdf")
    documents = loader.load()
    
    text_splitter = CharacterTextSplitter(chunk_size=500, chunk_overlap=50)
    texts = text_splitter.split_documents(documents)
    
    embeddings = HuggingFaceEmbeddings(
        model_name="sentence-transformers/all-MiniLM-L6-v2"
    )
    
    db = FAISS.from_documents(texts, embeddings)
    
    # 3. Updated LLM initialization
    llm = HuggingFaceEndpoint(
        repo_id="google/flan-t5-base",
        max_length=256,
        temperature=0.2,
        huggingfacehub_api_token=os.environ.get('HF_TOKEN')  # Explicit token passing
    )
    
    return RetrievalQA.from_chain_type(
        llm=llm,
        chain_type="stuff",
        retriever=db.as_retriever(search_kwargs={"k": 2})
    )

# 4. Initialize system
qa = create_qa_system()

# 5. Chat interface
def chat_response(message, history):
    response = qa({"query": message})
    return response["result"]

gr.ChatInterface(chat_response).launch()