Update app.py

app.py

@@ -15,10 +15,10 @@ summarizer = pipeline(
     device=-1  # CPU only
 )
 
-# New AI advice generator - lightweight text2text model
+# Better AI advice generator - use flan-t5-base for improved quality (still CPU-friendly)
 advice_generator = pipeline(
     "text2text-generation",
-    model="google/flan-t5-small",
+    model="google/flan-t5-base",
     device=-1  # CPU only
 )
 
@@ -56,21 +56,37 @@ def chunk_text(text: str):
 
 def generate_ai_advice(summary: str) -> str:
     """Use AI to generate personalized study advice based on the summary."""
-    # Truncate summary if too long for the small model
-    truncated_summary = summary[:800]  # Safe limit for flan-t5-small
+    # Truncate summary if too long
+    truncated_summary = summary[:1000]  # Slightly longer for base model
+    
+    # Improved prompt for better, more relevant tips
     prompt = (
-        f"Based on this summary: {truncated_summary}\n"
-        "Generate 5 concise study tips for a student to enhance learning and retention."
+        f"Read this summary of a technical paper: '{truncated_summary}'\n\n"
+        "Generate exactly 5 practical study tips for a student to better understand and retain this content. "
+        "Focus on active learning techniques, like practice, visualization, or connections to real-world applications. "
+        "Make each tip start with a verb (e.g., 'Review...', 'Apply...') and keep them concise. "
+        "Output only the 5 tips as bullet points, nothing else."
     )
-    generated = advice_generator(prompt, max_length=200, num_return_sequences=1)[0]["generated_text"]
     
-    # Format as markdown bullets
-    tips = generated.split(". ")  # Simple split assuming sentence-based output
+    generated = advice_generator(
+        prompt,
+        max_length=250,
+        num_return_sequences=1,
+        do_sample=False,  # Greedy for consistency
+        temperature=0.7   # Slight creativity if needed
+    )[0]["generated_text"]
+    
+    # Post-process to ensure bullet format
+    tips = [tip.strip() for tip in generated.split('\n') if tip.strip().startswith('-') or tip.strip()]
+    if not tips:
+        tips = generated.split('. ')  # Fallback split
+    
     advice_md = "\n\n---\n\n### 📚 AI-Generated Study Tips\n\n"
-    for tip in tips[:5]:  # Limit to 5
-        if tip.strip():
-            advice_md += f"- {tip.strip()}.\n"
-    advice_md += "\n**Pro tip**: Apply these tips consistently for better results!"
+    for i, tip in enumerate(tips[:5], 1):
+        clean_tip = tip.lstrip('- ').strip()
+        advice_md += f"- {clean_tip}\n"
+    
+    advice_md += "\n**Pro tip**: Combine these with spaced repetition for long-term mastery!"
     
     return advice_md