Update app.py

app.py

@@ -25,6 +25,10 @@ def generate_question_paper(
     num_sets,
     progress=gr.Progress()
 ):
+    # Add timeout protection
+    import time
+    start_time = time.time()
+    
     if not pdf_files or len(pdf_files) == 0:
         return "❌ Please upload at least one PDF file."
     
@@ -139,28 +143,45 @@ Do not output conversational text. Output ONLY the exam paper in a well-formatte
             response = ""
             token_count = 0
             max_tokens = 2500  # Increased for longer papers
+            last_update_time = time.time()
+            
+            try:
+                for message in client.chat_completion(
+                    messages=messages,
+                    model="meta-llama/Llama-3.2-3B-Instruct",
+                    max_tokens=max_tokens,
+                    temperature=0.7,
+                    stream=True,
+                    timeout=120,  # 2 minute timeout per request
+                ):
+                    # Check total timeout
+                    if time.time() - start_time > 300:  # 5 minute total timeout
+                        return f"⏱️ Request timeout. Please try with:\n- Fewer PDF files\n- Fewer questions\n- Fewer sets\n\nPartial output:\n{response}"
+                    
+                    if hasattr(message, 'choices') and len(message.choices) > 0:
+                        if hasattr(message.choices[0], 'delta') and hasattr(message.choices[0].delta, 'content'):
+                            response += message.choices[0].delta.content or ""
+                            token_count += 1
+                            
+                            # Update progress every 50 tokens to reduce overhead
+                            if token_count % 50 == 0 or time.time() - last_update_time > 2:
+                                # Calculate progress within this set (70-95% range divided by number of sets)
+                                set_start = 0.70 + (set_num - 1) * 0.30 / num_sets
+                                set_range = 0.25 / num_sets  # 25% of total progress for generation
+                                generation_progress = min((token_count / max_tokens), 1.0)
+                                current_progress = set_start + (generation_progress * set_range)
+                                percentage = int(generation_progress * 100)
+                                
+                                # Update with dynamic percentage
+                                progress(current_progress, 
+                                        desc=f"✍️ Generating Question Paper Set {set_num}/{num_sets}... {percentage}%")
+                                last_update_time = time.time()
             
-            for message in client.chat_completion(
-                messages=messages,
-                model="meta-llama/Llama-3.2-3B-Instruct",
-                max_tokens=max_tokens,
-                temperature=0.7,
-                stream=True,
-            ):
-                if hasattr(message, 'choices') and len(message.choices) > 0:
-                    if hasattr(message.choices[0], 'delta') and hasattr(message.choices[0].delta, 'content'):
-                        response += message.choices[0].delta.content or ""
-                        token_count += 1
-                        # Calculate progress within this set (70-95% range divided by number of sets)
-                        set_start = 0.70 + (set_num - 1) * 0.30 / num_sets
-                        set_range = 0.25 / num_sets  # 25% of total progress for generation
-                        generation_progress = min((token_count / max_tokens), 1.0)
-                        current_progress = set_start + (generation_progress * set_range)
-                        percentage = int(generation_progress * 100)
-                        
-                        # Update with dynamic percentage
-                        progress(current_progress, 
-                                desc=f"✍️ Generating Question Paper Set {set_num}/{num_sets}... {percentage}%")
+            except Exception as e:
+                if response:
+                    return f"⚠️ Generation interrupted: {str(e)}\n\nPartial output for Set {set_num}:\n{response}"
+                else:
+                    raise e
             
             progress(0.70 + set_num * 0.30 / num_sets, 
                     desc=f"✅ Set {set_num}/{num_sets} generated successfully!")
@@ -247,7 +268,10 @@ with gr.Blocks(title="AI Question Paper Generator") as demo:
             """)
         
         with gr.Column(scale=2):
-            output = gr.Markdown(label="Generated Question Paper(s)")
+            output = gr.Markdown(
+                label="Generated Question Paper(s)",
+                value="👋 Upload PDF files and configure settings to generate question papers..."
+            )
 
     btn.click(
         fn=generate_question_paper,
@@ -258,18 +282,32 @@ with gr.Blocks(title="AI Question Paper Generator") as demo:
             long_difficulty, long_count,
             num_sets
         ],
-        outputs=output
+        outputs=output,
+        show_progress="full"
     )
     
     gr.Markdown("""
-    ---
+    ---  
     **Features:**
     - ✅ Multiple PDF support (up to 5 files)
     - ✅ Separate difficulty control for each question type
     - ✅ Customizable question count per section
     - ✅ Generate 1-3 unique question paper sets
     - ✅ Automatic answer key generation for MCQs
+    - ✅ Queue system for concurrent users
+    
+    **Performance Tips:**
+    - For faster results: Use 1-2 PDFs, fewer questions, single set
+    - If timeout occurs: Reduce number of questions or sets
+    - Queue position will be shown when multiple users are active
     """)
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.queue(
+        max_size=20,  # Maximum queue size
+        default_concurrency_limit=2  # Allow 2 concurrent users
+    )
+    demo.launch(
+        show_error=True,
+        share=False
+    )