Update app.py

app.py

@@ -586,6 +586,65 @@ print("=" * 50)
 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
 # ------------------------------
@@ -716,33 +775,77 @@ def rag_search(query, subject, k=5):
     return "\n\n".join(results)
 
 # ------------------------------
-# Chapter Detection
+# Chapter Detection (Using Actual Chapter Names)
 # ------------------------------
-def detect_chapter(context, topic, subject):
-    """Detect which chapter the content is from using LLM"""
+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"
     
-    detection_prompt = f"""Based on the following textbook content from Class 12 {subject.title()}, identify the chapter name/title.
+    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.
 
-Content:
-{context[:800]}
+Topic: {topic}
 
-Topic searched: {topic}
+Content snippet:
+{context[:600]}
 
-Respond with ONLY the chapter name, nothing else. For example:
-- "Human Reproduction"
-- "Chemical Bonding"
-- "Laws of Motion"
+Available chapters:
+{chapter_list}
 
-Chapter name:"""
+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 chapter names from Class 12 NCERT textbooks. Respond with only the chapter name, no additional text."
+                    "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",
@@ -754,12 +857,19 @@ Chapter name:"""
             max_tokens=50
         )
         
-        chapter = response.choices[0].message.content.strip()
-        # Clean up the response
-        chapter = chapter.replace('"', '').replace("'", "").strip()
+        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()
         
-        print(f"✓ Detected chapter: {chapter}")
-        return chapter
+        # 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}")
@@ -789,8 +899,8 @@ Current status: API key not found or invalid."""
     
     print(f"🤖 Generating MCQs for {subject} - {topic}")
     
-    # First, detect the chapter
-    chapter = detect_chapter(context, topic, subject)
+    # 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}"
@@ -860,7 +970,7 @@ Possible causes:
 3. Network issue
 Please try again in a few seconds."""
         print(f"❌ Groq API Error: {e}")
-        return error_msg, "Unknown"
+        return error_msg, chapter
 
 def clean_mcq_output(text):
     lines = text.split('\n')
@@ -1072,7 +1182,7 @@ HTML_TEMPLATE = """
             <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;">⚡ Detecting chapter & creating questions...</p>
+                <p style="color: #999; font-size: 13px; margin-top: 10px;">⚡ Detecting chapter from textbook...</p>
             </div>
             
             <div class="result" id="result">
@@ -1212,4 +1322,4 @@ def health():
 if __name__ == "__main__":
     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)
+    app.run(host="0.0.0.0", port=port, debug=False)