Update app.py

app.py

@@ -178,24 +178,15 @@ def demo():
         .chat-message.user .text { align-self: flex-end; background: #d1e7dd; }
         .chat-message.assistant .text { align-self: flex-start; background: #e2e3e5; }
         .gradio-container { max-width: 1000px !important; margin: auto; font-family: 'Segoe UI', sans-serif; }
-        .gr-button { font-weight: bold; background-color: #004085 !important; color: white; }
+        .gr-button { font-weight: bold; background-color: #0d6efd !important; color: white; }
     """) as app:
 
         db_state = gr.State(None)
         chain_state = gr.State(None)
-        model_selection = gr.Dropdown(
-            choices=[
-                "mistralai/Mistral-7B-Instruct-v0.3",
-                "HuggingFaceH4/zephyr-7b-beta",
-                "tiiuae/falcon-7b-instruct"
-            ],
-            value="mistralai/Mistral-7B-Instruct-v0.3",
-            label="Choose Generative Model"
-        )
 
         gr.Markdown("""
         # 🧠 Customer Review Analyzer
-        Chat with your reviews using **custom LLMs**. Upload a `.txt` file of reviews and get insights instantly.
+        Upload a `.txt` file of reviews and chat with a custom AI system based on your data.
         """)
 
         with gr.Row():
@@ -209,58 +200,11 @@ def demo():
                 user_input = gr.Textbox(placeholder="Ask about your reviews...", show_label=False)
                 submit_btn = gr.Button("Send", variant="secondary")
 
-        # Hook updated LLM init with selected model
-        def process_and_initialize_with_model(file, selected_model):
-            token = get_hf_token()
-            if file is None or not token:
-                return None, None, "Please upload a file and ensure HF token is set."
-            with open(file, 'r', encoding='utf-8') as f:
-                reviews = [line.strip() for line in f if line.strip()]
-            analysis = analyze_reviews(reviews)
-            doc = generate_analysis_document(analysis)
-            db = create_db_from_analysis(doc)
-            rag_chain = initialize_rag_chatbot_with_model(db, selected_model)
-            return db, rag_chain, f"Processed {len(reviews)} reviews with {selected_model}. Ready to chat!"
-
-        def initialize_rag_chatbot_with_model(db, selected_model):
-            token = get_hf_token()
-            if not token:
-                return None
-            memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True, output_key="answer")
-            retriever = db.as_retriever(search_kwargs={"k": 4})
-            llm = FallbackLLM(token=token, model_list=[selected_model])
-            return ConversationalRetrievalChain.from_llm(llm=llm, retriever=retriever, memory=memory, return_source_documents=True)
-
-        # Update interaction to show only messages
-        def chat_display_only(query, qa_chain, chatbot):
-            history = chatbot or []
-            if not query.strip():
-                yield history, ""
-                return
-            if qa_chain is None:
-                history.append((None, "Please upload and process a file first."))
-                yield history, ""
-                return
-            history.append((query, ""))
-            yield history, ""
-            try:
-                response = qa_chain.invoke({"question": query, "chat_history": []})
-                assistant_response = response.get("answer", "No answer generated.")
-                for i in range(len(assistant_response)):
-                    history[-1] = (query, assistant_response[:i+1])
-                    yield history, ""
-                    time.sleep(0.01)
-            except Exception as e:
-                logger.error(f"RAG error: {e}")
-                history[-1] = (query, "An error occurred while answering your question.")
-                yield history, ""
-
-        process_btn.click(process_and_initialize_with_model, inputs=[file_input, model_selection], outputs=[db_state, chain_state, status])
-        submit_btn.click(chat_display_only, inputs=[user_input, chain_state, chatbot], outputs=[chatbot, user_input])
-        user_input.submit(chat_display_only, inputs=[user_input, chain_state, chatbot], outputs=[chatbot, user_input])
+        process_btn.click(process_and_initialize, inputs=[file_input], outputs=[db_state, chain_state, status])
+        submit_btn.click(user_query_with_rag, inputs=[user_input, chain_state, chatbot], outputs=[chatbot, user_input])
+        user_input.submit(user_query_with_rag, inputs=[user_input, chain_state, chatbot], outputs=[chatbot, user_input])
 
     return app
 
-
 if __name__ == "__main__":
     demo().launch(server_name="0.0.0.0", server_port=7860, share=False)