Update app.py

app.py

@@ -1,12 +1,11 @@
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
 import matplotlib.pyplot as plt
-import seaborn as sns
 import pandas as pd
 import torch
 
-# Load model
-model_id = "ibm-granite/granite-3b-code-instruct"  # Replace with actual granite model if different
+# Load your model (adjust if needed)
+model_id = "ibm-granite/granite-3b-code-instruct"
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 model = AutoModelForCausalLM.from_pretrained(
     model_id,
@@ -14,40 +13,40 @@ model = AutoModelForCausalLM.from_pretrained(
     torch_dtype=torch.float16
 )
 
-# Load sentiment analysis model
+# Sentiment analysis pipeline
 sentiment_analyzer = pipeline("sentiment-analysis")
 
-# Simulated citizen profiles
+# In-memory storage for feedback
+submitted_data = []
+
+# Dummy user profiles
 user_profiles = {
     "1001": {"location": "Hyderabad", "issues": ["traffic", "air pollution"]},
     "1002": {"location": "Delhi", "issues": ["waste management", "noise"]},
 }
 
-# Store submitted feedback during session
-submitted_data = []
-
-# Chat Function (ChatGPT-style)
+# Chat function
 def chat_fn(message, history):
-    full_prompt = tokenizer.apply_chat_template(
+    prompt = tokenizer.apply_chat_template(
         [{"role": "user", "content": message}],
         tokenize=False,
         add_generation_prompt=True
     )
-    inputs = tokenizer(full_prompt, return_tensors="pt").to(model.device)
+    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
     outputs = model.generate(**inputs, max_new_tokens=200)
-    reply = tokenizer.decode(outputs[0], skip_special_tokens=True).split("assistant")[-1].strip()
-    return reply
+    response = tokenizer.decode(outputs[0], skip_special_tokens=True).split("assistant")[-1].strip()
+    return response
 
-# Sentiment Analysis
+# Sentiment analysis
 def analyze_sentiment(text):
     result = sentiment_analyzer(text)[0]
     return f"{result['label']} ({result['score']*100:.2f}%)"
 
-# Live Feedback → Dashboard
+# Feedback form + live dashboard
 def collect_and_plot_feedback(comment, category):
     sentiment = sentiment_analyzer(comment)[0]["label"]
     submitted_data.append({"Category": category, "Sentiment": sentiment})
-
+    
     df = pd.DataFrame(submitted_data)
     summary = df.groupby(['Category', 'Sentiment']).size().unstack(fill_value=0)
 
@@ -56,32 +55,31 @@ def collect_and_plot_feedback(comment, category):
     plt.title("Live Citizen Sentiment by Category")
     plt.ylabel("Count")
     plt.tight_layout()
+    
     return f"Recorded sentiment: {sentiment}", fig
 
-# Personalized Contextual Assistant
+# Personalized assistant
 def personalized_response(user_id, query):
     profile = user_profiles.get(user_id)
     if not profile:
         return "User profile not found. Please check your user ID."
+    
     context = f"User from {profile['location']} concerned with: {', '.join(profile['issues'])}. Question: {query}"
-    input_tokens = tokenizer(context, return_tensors="pt").to(model.device)
-    output = model.generate(**input_tokens, max_new_tokens=150)
-    return tokenizer.decode(output[0], skip_special_tokens=True)
+    inputs = tokenizer(context, return_tensors="pt").to(model.device)
+    outputs = model.generate(**inputs, max_new_tokens=150)
+    reply = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    return reply
 
-# Build Gradio App
+# Build app
 with gr.Blocks(title="Citizen AI – Intelligent Citizen Engagement Platform") as demo:
     gr.Markdown("## 🧠 Citizen AI – Intelligent Citizen Engagement Platform")
 
     with gr.Tab("🤖 Chat Assistant"):
-        gr.ChatInterface(
+        chat = gr.ChatInterface(
             fn=chat_fn,
             title="🧠 Ask Citizen AI",
-            theme="soft",
             chatbot=gr.Chatbot(label="Citizen Chat"),
-            textbox=gr.Textbox(placeholder="Type your question here...", show_label=False),
-            retry_btn="🔁 Retry",
-            clear_btn="🗑️ Clear",
-            submit_btn="➤ Send"
+            textbox=gr.Textbox(placeholder="Type your question here...", show_label=False)
         )
 
     with gr.Tab("📊 Sentiment Analysis"):