app.py

import os
import tempfile
import streamlit as st

from langchain_community.document_loaders import PyPDFLoader
from langchain_community.vectorstores import FAISS
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain.chains import RetrievalQA
from langchain.prompts import PromptTemplate
from langchain_groq import GroqLLM

# Set environment variables (You can also use os.environ or Streamlit secrets)
GROQ_API_KEY = os.getenv("GROQ_API_KEY")
HUGGINGFACE_API_KEY = os.getenv("HUGGINGFACE_API_KEY")

# Initialize Groq LLM
llm = GroqLLM(
    api_key=GROQ_API_KEY,
    model="llama3-8b-8192",  # <- correct param
    temperature=0.1
)

# HuggingFace Embeddings
embedding = HuggingFaceEmbeddings()

st.title("📄 RAG Chat with Groq + HuggingFace")

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

user_query = st.text_input("Ask something about the document")
submit_button = st.button("Submit")

if uploaded_file and submit_button:
    # Save PDF temporarily
    with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp_file:
        tmp_file.write(uploaded_file.read())
        tmp_path = tmp_file.name

    # Load and split
    loader = PyPDFLoader(tmp_path)
    pages = loader.load_and_split()

    # Create FAISS vectorstore
    vectorstore = FAISS.from_documents(pages, embedding)
    retriever = vectorstore.as_retriever()

    # Custom prompt (optional)
    prompt_template = PromptTemplate(
        input_variables=["context", "question"],
        template="""
        Use the following context to answer the question. Be concise and accurate.

        Context: {context}

        Question: {question}
        """
    )

    # Create QA chain
    qa_chain = RetrievalQA.from_chain_type(
        llm=llm,
        retriever=retriever,
        return_source_documents=True,
        chain_type_kwargs={"prompt": prompt_template}
    )

    # Run QA
    result = qa_chain({"query": user_query})
    st.markdown("### 💬 Answer")
    st.write(result["result"])

    # Optional: Show sources
    with st.expander("📄 Sources"):
        for doc in result["source_documents"]:
            st.write(doc.metadata["source"])