Create components/pdf_chat.py

components/pdf_chat.py

@@ -1,27 +1,42 @@
+# components/pdf_chat.py
+
 import streamlit as st
-from utils.session import get_state_value
 from utils.retriever import load_vectorstore
-from langchain_community.chat_models import ChatGroq
+from utils.session import load_from_json
 from langchain.chains import RetrievalQA
+from langchain_core.messages import AIMessage
+from langchain_community.chat_models import ChatGroq
+from langchain_openai import OpenAIEmbeddings
 
 def run_pdf_qa():
-    st.title("💬 Ask Questions from PDF")
-    st.markdown("Chat with your uploaded PDF document using RAG-powered QA system.")
+    st.subheader("💬 Ask Questions From Your PDF")
+
+    pdf_path = load_from_json("pdf_path")
 
-    pdf_path = get_state_value("pdf_path", None)
-    if not pdf_path:
-        st.warning("Please upload a PDF document in the Upload Docs section first.")
+    if not pdf_path or not pdf_path.endswith(".pdf"):
+        st.warning("📄 Please upload a PDF document first in the 'Upload Business Docs' section.")
         return
 
-    retriever = load_vectorstore(pdf_path)
-    qa_chain = RetrievalQA.from_chain_type(
-        llm=ChatGroq(model_name="llama3-8b-8192", temperature=0.3),
-        retriever=retriever
-    )
-
-    query = st.text_input("❓ Enter your question about the document:")
-    if query:
-        with st.spinner("Generating answer..."):
-            result = qa_chain.run(query)
-            st.markdown("### ✅ Answer")
-            st.markdown(result)
+    try:
+        retriever = load_vectorstore(pdf_path)
+        llm = ChatGroq(temperature=0.3, model_name="mixtral-8x7b-32768")
+        qa_chain = RetrievalQA.from_chain_type(
+            llm=llm,
+            retriever=retriever.as_retriever(),
+            return_source_documents=False
+        )
+
+        question = st.text_input("🔍 Ask a question based on your uploaded PDF:")
+
+        if question:
+            with st.spinner("Generating answer..."):
+                result = qa_chain.invoke({"query": question})
+
+                if isinstance(result, dict) and "result" in result:
+                    st.markdown("### ✅ Answer")
+                    st.write(result["result"])
+                else:
+                    st.error("Unexpected response from the model.")
+
+    except Exception as e:
+        st.error(f"❌ Error loading vector store: {str(e)}")