Update app.py

app.py

@@ -24,6 +24,65 @@
 # print("STARTING MCQ GENERATOR APP")
 # print("=" * 50)
 
+# # ------------------------------
+# # Chapter Names (Actual Textbook Chapters)
+# # ------------------------------
+# CHAPTER_NAMES = {
+#     "biology": [
+#         "Reproduction in Lower and Higher Plants",
+#         "Reproduction in Lower and Higher Animals",
+#         "Inheritance and Variation",
+#         "Molecular Basis of Inheritance",
+#         "Origin and Evolution of Life",
+#         "Plant Water Relation",
+#         "Plant Growth and Mineral Nutrition",
+#         "Respiration and Circulation",
+#         "Control and Co-ordination",
+#         "Human Health and Diseases",
+#         "Enhancement of Food Production",
+#         "Biotechnology",
+#         "Organisms and Populations",
+#         "Ecosystems and Energy Flow",
+#         "Biodiversity, Conservation and Environmental Issues"
+#     ],
+#     "chemistry": [
+#         "Solid State",
+#         "Solutions",
+#         "Ionic Equilibria",
+#         "Chemical Thermodynamics",
+#         "Electrochemistry",
+#         "Chemical Kinetics",
+#         "Elements of Groups 16, 17 and 18",
+#         "Transition and Inner transition Elements",
+#         "Coordination Compounds",
+#         "Halogen Derivatives",
+#         "Alcohols, Phenols and Ethers",
+#         "Aldehydes, Ketones and Carboxylic acids",
+#         "Amines",
+#         "Biomolecules",
+#         "Introduction to Polymer Chemistry",
+#         "Green Chemistry and Nanochemistry"
+#     ],
+#     "physics": [
+#         "Rotational Dynamics",
+#         "Mechanical Properties of Fluids",
+#         "Kinetic Theory of Gases and Radiation",
+#         "Thermodynamics",
+#         "Oscillations",
+#         "Superposition of Waves",
+#         "Wave Optics",
+#         "Electrostatics",
+#         "Current Electricity",
+#         "Magnetic Fields due to Electric Current",
+#         "Magnetic Materials",
+#         "Electromagnetic induction",
+#         "AC Circuits",
+#         "Dual Nature of Radiation and Matter",
+#         "Structure of Atoms and Nuclei",
+#         "Semiconductor Devices"
+#     ]
+# }
+
 # # ------------------------------
 # # Initialize Groq API Client
 # # ------------------------------
@@ -153,6 +212,107 @@
     
 #     return "\n\n".join(results)
 
+# # ------------------------------
+# # Chapter Detection (Using Actual Chapter Names)
+# # ------------------------------
+# def detect_chapter_from_list(context, topic, subject):
+#     """
+#     Detect chapter using the actual chapter list by matching keywords
+#     """
+#     if subject not in CHAPTER_NAMES:
+#         return "Unknown Chapter"
+    
+#     chapters = CHAPTER_NAMES[subject]
+#     combined_text = (topic + " " + context[:1000]).lower()
+    
+#     # Score each chapter based on keyword matching
+#     scores = {}
+#     for chapter in chapters:
+#         score = 0
+#         chapter_words = chapter.lower().split()
+        
+#         # Check if chapter words appear in the content
+#         for word in chapter_words:
+#             if len(word) > 3:  # Ignore small words like "and", "the"
+#                 if word in combined_text:
+#                     score += 1
+        
+#         # Bonus if topic is similar to chapter name
+#         topic_words = topic.lower().split()
+#         for t_word in topic_words:
+#             if len(t_word) > 3 and t_word in chapter.lower():
+#                 score += 2
+        
+#         if score > 0:
+#             scores[chapter] = score
+    
+#     # Return chapter with highest score
+#     if scores:
+#         best_chapter = max(scores.items(), key=lambda x: x[1])[0]
+#         print(f"✓ Matched chapter: {best_chapter} (score: {scores[best_chapter]})")
+#         return best_chapter
+    
+#     # Fallback: Use LLM to choose from the list
+#     return detect_chapter_with_llm(context, topic, subject, chapters)
+
+# def detect_chapter_with_llm(context, topic, subject, chapters):
+#     """
+#     Use LLM to pick the correct chapter from the provided list
+#     """
+#     if not groq_client:
+#         return "Unknown Chapter"
+    
+#     chapter_list = "\n".join([f"{i+1}. {ch}" for i, ch in enumerate(chapters)])
+    
+#     detection_prompt = f"""Based on the following textbook content and topic, identify which chapter from the list below this content belongs to.
+
+# Topic: {topic}
+
+# Content snippet:
+# {context[:600]}
+
+# Available chapters:
+# {chapter_list}
+
+# Respond with ONLY the chapter number and name exactly as listed (e.g., "5. Origin and Evolution of Life"). Choose the most relevant chapter.
+
+# Chapter:"""
+    
+#     try:
+#         response = groq_client.chat.completions.create(
+#             messages=[
+#                 {
+#                     "role": "system",
+#                     "content": "You are an expert at identifying which chapter textbook content belongs to. Respond with only the chapter number and name from the provided list."
+#                 },
+#                 {
+#                     "role": "user",
+#                     "content": detection_prompt
+#                 }
+#             ],
+#             model="llama-3.3-70b-versatile",
+#             temperature=0.1,
+#             max_tokens=50
+#         )
+        
+#         result = response.choices[0].message.content.strip()
+        
+#         # Extract chapter name from response (remove number prefix if present)
+#         chapter = re.sub(r'^\d+\.\s*', '', result).strip()
+        
+#         # Verify it's in our list
+#         for ch in chapters:
+#             if ch.lower() in chapter.lower() or chapter.lower() in ch.lower():
+#                 print(f"✓ LLM detected chapter: {ch}")
+#                 return ch
+        
+#         print(f"⚠️ LLM response not in list: {result}")
+#         return chapters[0]  # Default to first chapter
+        
+#     except Exception as e:
+#         print(f"⚠️ Chapter detection failed: {e}")
+#         return "Unknown Chapter"
+
 # # ------------------------------
 # # MCQ Generation
 # # ------------------------------
@@ -160,14 +320,12 @@
 #     # Check if Groq is available
 #     if not groq_client:
 #         error_msg = """ERROR: Groq API not initialized!
-
 # Please check:
 # 1. GROQ_API_KEY is set in Space Settings → Repository secrets
 # 2. API key is valid (get one from https://console.groq.com/keys)
 # 3. Space has been restarted after adding the key
-
 # Current status: API key not found or invalid."""
-#         return error_msg
+#         return error_msg, "Unknown"
     
 #     # Check cache
 #     context_hash = hashlib.md5(context.encode()).hexdigest()[:8]
@@ -175,14 +333,16 @@
     
 #     if cache_key in MCQ_CACHE:
 #         print("✓ Using cached MCQs")
-#         return MCQ_CACHE[cache_key]
+#         return MCQ_CACHE[cache_key]["mcqs"], MCQ_CACHE[cache_key]["chapter"]
     
 #     print(f"🤖 Generating MCQs for {subject} - {topic}")
     
+#     # Detect the chapter from our actual chapter list
+#     chapter = detect_chapter_from_list(context, topic, subject)
+    
 #     prompt = f"""You are a Class-12 {subject.title()} teacher creating MCQs.
-
 # Topic: "{topic}"
-
+# Chapter: "{chapter}"
 # Reference material from textbook:
 # {context[:1500]}
 
@@ -234,22 +394,21 @@
 #         result = chat_completion.choices[0].message.content.strip()
 #         result = clean_mcq_output(result)
         
-#         cache_mcq(cache_key, result)
+#         # Cache both MCQs and chapter
+#         cache_mcq(cache_key, {"mcqs": result, "chapter": chapter})
         
 #         print("✓ MCQs generated successfully")
-#         return result
+#         return result, chapter
         
 #     except Exception as e:
 #         error_msg = f"""Error calling Groq API: {str(e)}
-
 # Possible causes:
 # 1. Rate limit exceeded (wait a moment)
 # 2. Invalid API key
 # 3. Network issue
-
 # Please try again in a few seconds."""
 #         print(f"❌ Groq API Error: {e}")
-#         return error_msg
+#         return error_msg, chapter
 
 # def clean_mcq_output(text):
 #     lines = text.split('\n')
@@ -357,6 +516,27 @@
 #             margin-bottom: 20px;
 #             font-size: 1.4em;
 #         }
+#         .chapter-info {
+#             background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
+#             color: white;
+#             padding: 15px 20px;
+#             border-radius: 8px;
+#             margin-bottom: 20px;
+#             display: flex;
+#             align-items: center;
+#             gap: 10px;
+#             font-size: 16px;
+#         }
+#         .chapter-icon {
+#             font-size: 24px;
+#         }
+#         .chapter-text {
+#             flex: 1;
+#         }
+#         .chapter-name {
+#             font-weight: 700;
+#             font-size: 18px;
+#         }
 #         .mcq-content {
 #             background: white;
 #             padding: 25px;
@@ -440,11 +620,20 @@
 #             <div class="loading" id="loading">
 #                 <div class="spinner"></div>
 #                 <p style="color: #666; font-size: 16px;">Generating MCQs with AI...</p>
-#                 <p style="color: #999; font-size: 13px; margin-top: 10px;">⚡ Usually takes 5-10 seconds</p>
+#                 <p style="color: #999; font-size: 13px; margin-top: 10px;">⚡ Detecting chapter from textbook...</p>
 #             </div>
             
 #             <div class="result" id="result">
 #                 <h3>📝 Generated MCQs:</h3>
+                
+#                 <div class="chapter-info" id="chapterInfo" style="display: none;">
+#                     <span class="chapter-icon">📖</span>
+#                     <div class="chapter-text">
+#                         <div style="font-size: 13px; opacity: 0.9;">Chapter:</div>
+#                         <div class="chapter-name" id="chapterName"></div>
+#                     </div>
+#                 </div>
+                
 #                 <div style="background: #d4edda; padding: 12px; border-radius: 6px; margin-bottom: 15px; color: #155724; font-size: 14px;">
 #                     ✓ <strong>High Quality:</strong> Generated by Llama 3.3 70B via Groq API
 #                 </div>
@@ -465,9 +654,11 @@
 #             const loading = document.getElementById('loading');
 #             const result = document.getElementById('result');
 #             const btn = document.querySelector('button');
+#             const chapterInfo = document.getElementById('chapterInfo');
             
 #             loading.classList.add('show');
 #             result.classList.remove('show');
+#             chapterInfo.style.display = 'none';
 #             btn.disabled = true;
 #             btn.textContent = '⏳ Generating...';
             
@@ -485,6 +676,12 @@
 #                     return;
 #                 }
                 
+#                 // Display chapter info
+#                 if (data.chapter && data.chapter !== 'Unknown Chapter') {
+#                     document.getElementById('chapterName').textContent = data.chapter;
+#                     chapterInfo.style.display = 'flex';
+#                 }
+                
 #                 document.getElementById('mcqContent').textContent = data.mcqs;
 #                 result.classList.add('show');
 #             } catch (error) {
@@ -535,9 +732,13 @@
         
 #         print(f"✓ Context found ({len(context)} chars)")
         
-#         mcqs = generate_mcqs(context, topic, subject)
+#         mcqs, chapter = generate_mcqs(context, topic, subject)
         
-#         return jsonify({"mcqs": mcqs, "subject": subject})
+#         return jsonify({
+#             "mcqs": mcqs, 
+#             "subject": subject,
+#             "chapter": chapter
+#         })
     
 #     except Exception as e:
 #         print(f"❌ Error: {e}")
@@ -560,6 +761,7 @@
 #     port = int(os.environ.get("PORT", 7860))
 #     print(f"\n🚀 Starting server on port {port}...\n")
 #     app.run(host="0.0.0.0", port=port, debug=False)
+
 import pickle
 import faiss
 from flask import Flask, request, jsonify, render_template_string
@@ -780,9 +982,10 @@ def rag_search(query, subject, k=5):
 def detect_chapter_from_list(context, topic, subject):
     """
     Detect chapter using the actual chapter list by matching keywords
+    Returns None if topic doesn't match the subject
     """
     if subject not in CHAPTER_NAMES:
-        return "Unknown Chapter"
+        return None
     
     chapters = CHAPTER_NAMES[subject]
     combined_text = (topic + " " + context[:1000]).lower()
@@ -820,32 +1023,35 @@ def detect_chapter_from_list(context, topic, subject):
 def detect_chapter_with_llm(context, topic, subject, chapters):
     """
     Use LLM to pick the correct chapter from the provided list
+    Also verifies if the topic belongs to the subject
     """
     if not groq_client:
-        return "Unknown Chapter"
+        return None
     
     chapter_list = "\n".join([f"{i+1}. {ch}" for i, ch in enumerate(chapters)])
     
-    detection_prompt = f"""Based on the following textbook content and topic, identify which chapter from the list below this content belongs to.
+    detection_prompt = f"""Based on the following textbook content and topic, identify which chapter from the Class 12 {subject.title()} textbook this content belongs to.
 
 Topic: {topic}
 
 Content snippet:
 {context[:600]}
 
-Available chapters:
+Available {subject.title()} chapters:
 {chapter_list}
 
-Respond with ONLY the chapter number and name exactly as listed (e.g., "5. Origin and Evolution of Life"). Choose the most relevant chapter.
+IMPORTANT: If the topic and content do NOT belong to {subject.title()}, respond with "NOT_MATCHING".
 
-Chapter:"""
+If it matches, respond with ONLY the chapter number and name exactly as listed (e.g., "5. Origin and Evolution of Life").
+
+Response:"""
     
     try:
         response = groq_client.chat.completions.create(
             messages=[
                 {
                     "role": "system",
-                    "content": "You are an expert at identifying which chapter textbook content belongs to. Respond with only the chapter number and name from the provided list."
+                    "content": f"You are an expert at identifying which chapter textbook content belongs to. You can recognize when content doesn't match the subject. If the topic is from a different subject than {subject.title()}, respond with 'NOT_MATCHING'."
                 },
                 {
                     "role": "user",
@@ -859,6 +1065,11 @@ Chapter:"""
         
         result = response.choices[0].message.content.strip()
         
+        # Check if topic doesn't match the subject
+        if "NOT_MATCHING" in result.upper() or "NOT MATCHING" in result.upper():
+            print(f"⚠️ Topic '{topic}' doesn't belong to {subject}")
+            return None
+        
         # Extract chapter name from response (remove number prefix if present)
         chapter = re.sub(r'^\d+\.\s*', '', result).strip()
         
@@ -869,11 +1080,11 @@ Chapter:"""
                 return ch
         
         print(f"⚠️ LLM response not in list: {result}")
-        return chapters[0]  # Default to first chapter
+        return None
         
     except Exception as e:
         print(f"⚠️ Chapter detection failed: {e}")
-        return "Unknown Chapter"
+        return None
 
 # ------------------------------
 # MCQ Generation
@@ -887,7 +1098,7 @@ Please check:
 2. API key is valid (get one from https://console.groq.com/keys)
 3. Space has been restarted after adding the key
 Current status: API key not found or invalid."""
-        return error_msg, "Unknown"
+        return error_msg, None
     
     # Check cache
     context_hash = hashlib.md5(context.encode()).hexdigest()[:8]
@@ -902,6 +1113,12 @@ Current status: API key not found or invalid."""
     # Detect the chapter from our actual chapter list
     chapter = detect_chapter_from_list(context, topic, subject)
     
+    # If chapter is None, topic doesn't belong to this subject
+    if chapter is None:
+        error_msg = f"❌ The topic '{topic}' does not belong to {subject.title()}.\n\nPlease enter a topic related to {subject.title()} or select the correct subject."
+        print(f"⚠️ Topic mismatch: '{topic}' not in {subject}")
+        return error_msg, None
+    
     prompt = f"""You are a Class-12 {subject.title()} teacher creating MCQs.
 Topic: "{topic}"
 Chapter: "{chapter}"
@@ -1296,6 +1513,10 @@ def generate():
         
         mcqs, chapter = generate_mcqs(context, topic, subject)
         
+        # Check if there was a subject mismatch
+        if chapter is None:
+            return jsonify({"error": mcqs}), 400
+        
         return jsonify({
             "mcqs": mcqs, 
             "subject": subject,