app.py

import os
import streamlit as st
from glob import glob
from langchain_community.document_loaders import PyPDFLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.vectorstores import FAISS
from langchain.embeddings import HuggingFaceEmbeddings
from langchain.chains import RetrievalQA
from langchain_groq import ChatGroq  # ✅ Correct import

# Set page config
st.set_page_config(page_title="SMEHelpBot 🤖", layout="wide")
st.title("🤖 SMEHelpBot – Your AI Assistant for Small Businesses")

# Load API key
GROQ_API_KEY = st.secrets.get("GROQ_API_KEY") or os.getenv("GROQ_API_KEY")
if not GROQ_API_KEY:
    st.error("❌ Please set your GROQ_API_KEY in environment or .streamlit/secrets.toml")
    st.stop()

os.environ["GROQ_API_KEY"] = GROQ_API_KEY

# Load all PDFs from the 'docs' folder
pdf_paths = glob("docs/*.pdf")

if not pdf_paths:
    st.warning("📁 Please place some PDF files in the `docs/` folder.")
    st.stop()

#st.info(f"📄 Loaded {len(pdf_paths)} document(s) from `docs/`")

# Load and split all PDFs
documents = []
for path in pdf_paths:
    loader = PyPDFLoader(path)
    documents.extend(loader.load())

splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
chunks = splitter.split_documents(documents)

# Create vector store from chunks
embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
vectorstore = FAISS.from_documents(chunks, embeddings)
retriever = vectorstore.as_retriever()

# Set up LLM with Groq
llm = ChatGroq(temperature=0.3, model_name="llama3-8b-8192")

# Build QA chain
qa_chain = RetrievalQA.from_chain_type(
    llm=llm,
    chain_type="stuff",
    retriever=retriever,
    return_source_documents=True
)

# Ask question
user_question = st.text_input("💬 Ask your question about SME documents:")

if user_question:
    with st.spinner("🤔 Thinking..."):
        result = qa_chain({"query": user_question})
        st.success("✅ Answer:")
        st.write(result["result"])

        with st.expander("📄 Source Snippets"):
            for i, doc in enumerate(result["source_documents"]):
                st.markdown(f"**Source {i+1}:**\n{doc.page_content[:300]}...")