Update app.py

app.py

@@ -1,7 +1,7 @@
 import gradio as gr
 import os
 import pandas as pd
-from transformers import pipeline, Conversation
+from transformers import pipeline
 
 # ----- Config -----
 BASE_DIR = "data"
@@ -65,29 +65,45 @@ def reset_all():
         {}  # data_state (gr.State) reset with raw dict
     )
 
-# ----- Initialize Hugging Face conversational pipeline -----
-chatbot = pipeline(model="facebook/blenderbot-400M-distill", task="conversational")
+# ----- Chatbot setup -----
+# Use a text2text-generation pipeline with a chat-capable model
+chat_model_name = "facebook/blenderbot-400M-distill"  # You can replace this with any chat-capable model
+chat_generator = pipeline("text2text-generation", model=chat_model_name)
 
-# To keep conversation history
-message_list = []
-response_list = []
+# We'll maintain chat history as a list of strings (user and bot messages)
+chat_history = []
 
-def chat_with_bot(message, history):
-    global message_list, response_list
+def chat_with_bot(user_message, history):
+    global chat_history
     if history is None:
         history = []
-    message_list.append(message)
-    conversation = Conversation(text=message, past_user_inputs=message_list[:-1], generated_responses=response_list)
-    conversation = chatbot(conversation)
-    response = conversation.generated_responses[-1]
-    response_list.append(response)
-    history.append((message, response))
+        chat_history = []
+
+    # Append user message to history
+    chat_history.append(f"User: {user_message}")
+
+    # Create prompt by joining conversation history with a Bot prompt
+    prompt = "\n".join(chat_history) + "\nBot:"
+
+    # Generate response (adjust max_length and other params as needed)
+    response = chat_generator(prompt, max_length=100, do_sample=True)[0]['generated_text']
+
+    # Extract bot reply by removing prompt prefix if present
+    bot_reply = response[len(prompt):].strip()
+    if not bot_reply:
+        bot_reply = response.strip()  # fallback if slicing fails
+
+    # Append bot reply to history
+    chat_history.append(f"Bot: {bot_reply}")
+
+    # Update Gradio chat history (list of (user, bot) tuples)
+    history.append((user_message, bot_reply))
+
     return history, history
 
 def clear_chat():
-    global message_list, response_list
-    message_list.clear()
-    response_list.clear()
+    global chat_history
+    chat_history = []
     return [], []
 
 # ----- UI -----