Update app.py

app.py

@@ -1,25 +1,21 @@
 import streamlit as st
 import os
 import tempfile
-import fitz  # PyMuPDF for PDFs
+import fitz
 import docx
 import openpyxl
 import faiss
 
+from groq import Groq
 from langchain_community.embeddings import HuggingFaceEmbeddings
-from langchain_community.llms import Groq
 from langchain.vectorstores import FAISS
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain.docstore.document import Document
-from langchain.chains import RetrievalQA
 
-# Load LLM (API key from Hugging Face secrets)
-llm = Groq(
-    model="llama3-8b-8192",
-    api_key=os.getenv("GROQ_API_KEY")
-)
+# Initialize Groq client
+groq_client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
 
-# Embeddings model
+# Embedding model
 embedding_model = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
 
 # File readers
@@ -58,30 +54,52 @@ def process_file(uploaded_file):
     else:
         return "Unsupported file type."
 
+# Prompt builder
+def build_prompt(context, question):
+    return f"""You are a helpful assistant. Answer the question based only on the context provided below.
+
+Context:
+{context}
+
+Question:
+{question}
+
+Answer:"""
+
 # Streamlit App
 st.set_page_config(page_title="DocuQuery AI", layout="centered")
 st.title("📄 DocuQuery AI")
-st.markdown("Upload a document (PDF, Word, or Excel) and ask questions about its content using LLaMA3.")
+st.markdown("Upload a document and ask questions about it using LLaMA-3 from Groq.")
 
 uploaded_file = st.file_uploader("Upload your document", type=["pdf", "docx", "xlsx"])
 
 if uploaded_file:
     st.success("✅ File uploaded successfully.")
-    with st.spinner("Reading and processing file..."):
+    with st.spinner("Processing file..."):
         raw_text = process_file(uploaded_file)
 
     splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
     docs = [Document(page_content=chunk) for chunk in splitter.split_text(raw_text)]
 
-    with st.spinner("Indexing document with FAISS..."):
+    with st.spinner("Embedding & indexing..."):
         db = FAISS.from_documents(docs, embedding_model)
         retriever = db.as_retriever(search_type="similarity", search_kwargs={"k": 4})
-        qa_chain = RetrievalQA.from_chain_type(llm=llm, chain_type="stuff", retriever=retriever)
 
-    st.success("📚 Document indexed. Ask your question below!")
+    st.success("📚 Document indexed. Ask a question!")
 
     user_query = st.text_input("❓ Ask something about the document:")
     if user_query:
-        with st.spinner("Generating answer..."):
-            response = qa_chain.run(user_query)
-            st.markdown(f"**💬 Answer:** {response}")
+        with st.spinner("Generating response..."):
+            retrieved_docs = retriever.get_relevant_documents(user_query)
+            context = "\n".join([doc.page_content for doc in retrieved_docs])
+
+            prompt = build_prompt(context, user_query)
+
+            response = groq_client.chat.completions.create(
+                model="llama3-8b-8192",
+                messages=[
+                    {"role": "user", "content": prompt}
+                ]
+            )
+
+            st.markdown(f"**💬 Answer:** {response.choices[0].message.content}")