Update app.py

app.py

@@ -1,60 +1,58 @@
-import os
 import streamlit as st
 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_community.chat_models import ChatGroq
-
-# Set Streamlit page config
-st.set_page_config(page_title="SMEHelpBot 🤖", layout="wide")
-st.title("🤖 SMEHelpBot – Your AI Assistant for Small Businesses")
+from langchain_community.llms import HuggingFaceEndpoint
+import os
 
-# File uploader
-uploaded_file = st.file_uploader("📄 Upload a PDF (e.g., SME policy, business doc, etc.):", type=["pdf"])
-user_question = st.text_input("💬 Ask a question related to your document or SME operations:")
+# --- UI ---
+st.set_page_config(page_title="SMEHelpBot", layout="wide")
+st.title("🤖 SMEHelpBot – Your AI Assistant for Small Business")
 
-# Set Groq API key securely (use Streamlit secrets or env var)
-GROQ_API_KEY = st.secrets.get("GROQ_API_KEY") or os.getenv("GROQ_API_KEY") or "your_groq_api_key_here"
+uploaded_file = st.file_uploader("📄 Upload an industry-specific PDF (policy, FAQ, etc.):", type=["pdf"])
+user_query = st.text_input("💬 Ask a business-related question:")
 
+# --- Process PDF + RAG ---
 if uploaded_file:
-    # Save uploaded file temporarily
     with open("temp.pdf", "wb") as f:
         f.write(uploaded_file.read())
-
-    # Load PDF and split into chunks
+    
     loader = PyPDFLoader("temp.pdf")
-    documents = loader.load()
+    pages = loader.load()
 
     splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
-    chunks = splitter.split_documents(documents)
+    chunks = splitter.split_documents(pages)
 
-    # Create vector store using MiniLM embeddings
     embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
-    vectorstore = FAISS.from_documents(chunks, embeddings)
-    retriever = vectorstore.as_retriever()
+    db = FAISS.from_documents(chunks, embeddings)
 
-    # Set up LLM using Groq + LLaMA3
-    os.environ["GROQ_API_KEY"] = GROQ_API_KEY
-    llm = ChatGroq(temperature=0.3, model_name="llama3-8b-8192")
+    retriever = db.as_retriever()
+
+    # --- Groq API (LLaMA3 via HuggingFaceEndpoint) ---
+    os.environ["HUGGINGFACEHUB_API_TOKEN"] = st.secrets.get("HF_TOKEN") or "your_api_token_here"
+
+    llm = HuggingFaceEndpoint(
+        repo_id="meta-llama/Meta-Llama-3-8B-Instruct",
+        temperature=0.6,
+        max_new_tokens=512
+    )
 
-    # QA chain with retrieval
-    qa = RetrievalQA.from_chain_type(
+    qa_chain = RetrievalQA.from_chain_type(
         llm=llm,
-        chain_type="stuff",
         retriever=retriever,
         return_source_documents=True
     )
 
-    if user_question:
-        with st.spinner("Generating answer..."):
-            result = qa({"query": user_question})
+    if user_query:
+        with st.spinner("Generating response..."):
+            result = qa_chain({"query": user_query})
             st.success(result["result"])
 
-            with st.expander("📄 Sources"):
+            with st.expander("📚 Sources"):
                 for doc in result["source_documents"]:
-                    st.markdown(f"• {doc.page_content[:300]}...")
+                    st.markdown(f"• Page content: {doc.page_content[:300]}...")
 
 else:
-    st.info("Upload a PDF and enter a question to begin.")
+    st.info("Upload a PDF and type your question to get started.")