Create granite_model.py

granite_model.py

@@ -0,0 +1,124 @@
+# Example of how to integrate the granite_model.py into your main app.py
+
+# At the top of your app.py, add this import:
+# Use a pipeline as a high-level helper
+from transformers import pipeline
+
+pipe = pipeline("text-generation", model="ibm-granite/granite-3.3-8b-instruct")
+messages = [
+    {"role": "user", "content": "Who are you?"},
+]
+pipe(messages)
+try:
+    from granite_model import GraniteModelIntegration
+    GRANITE_AVAILABLE = True
+except ImportError:
+    GRANITE_AVAILABLE = False
+    logger.warning("granite_model.py not found. Granite features will be disabled.")
+
+# In your AdvancedDocumentSummarizer.__init__ method, add:
+def __init__(self):
+    self.summarizer = None
+    self.sentiment_analyzer = None
+    self.granite_integration = None  # Add this line
+    self.cache = {}
+    
+    # Initialize AI models
+    if TRANSFORMERS_AVAILABLE:
+        self._initialize_ai_models()
+    
+    # Initialize Granite integration
+    if GRANITE_AVAILABLE:
+        try:
+            self.granite_integration = GraniteModelIntegration()
+            logger.info(f"Granite integration status: {'Available' if self.granite_integration.is_available() else 'Not Available'}")
+        except Exception as e:
+            logger.warning(f"Failed to initialize Granite integration: {e}")
+    
+    # Initialize sentiment analyzer
+    if NLTK_AVAILABLE:
+        try:
+            self.sentiment_analyzer = SentimentIntensityAnalyzer()
+        except Exception as e:
+            logger.warning(f"Failed to initialize sentiment analyzer: {e}")
+
+# Add these methods to your AdvancedDocumentSummarizer class:
+def granite_enhanced_summary(self, text: str, summary_type: str = "medium") -> str:
+    """Generate enhanced summary using Granite model"""
+    if not (self.granite_integration and self.granite_integration.is_available()):
+        return self.advanced_extractive_summary(text)
+    
+    return self.granite_integration.generate_summary(text, summary_type)
+
+def granite_analyze_document(self, text: str) -> Dict:
+    """Use Granite model for advanced document analysis"""
+    if not (self.granite_integration and self.granite_integration.is_available()):
+        return {'analysis_available': False}
+    
+    result = self.granite_integration.analyze_document(text)
+    return {
+        'granite_analysis': result.get('analysis', 'Analysis failed'),
+        'analysis_available': result.get('success', False),
+        'model_used': result.get('model_used', 'Unknown')
+    }
+
+def granite_generate_questions(self, text: str, num_questions: int = 5) -> list:
+    """Generate comprehension questions using Granite"""
+    if not (self.granite_integration and self.granite_integration.is_available()):
+        return []
+    
+    return self.granite_integration.generate_questions(text, num_questions)
+
+# In your process_document method, update the summary generation part:
+# Generate summary - prioritize Granite if available for AI mode
+if summary_type == "ai":
+    if self.granite_integration and self.granite_integration.is_available():
+        summary = self.granite_enhanced_summary(text, summary_length)
+    elif self.summarizer:
+        summary = self.ai_summary(text, params["max_length"], params["min_length"])
+    else:
+        summary = self.advanced_extractive_summary(text, params["sentences"])
+else:
+    summary = self.advanced_extractive_summary(text, params["sentences"])
+
+# Get Granite analysis and questions if available
+granite_analysis = self.granite_analyze_document(text)
+granite_questions = self.granite_generate_questions(text, 5)
+
+# Add to result dictionary:
+result = {
+    'original_text': text[:2000] + "..." if len(text) > 2000 else text,
+    'full_text_length': len(text),
+    'summary': summary,
+    'key_points': key_points,
+    'outline': outline,
+    'stats': stats,
+    'granite_analysis': granite_analysis,
+    'granite_questions': granite_questions,  # Add this
+    'readability_score': readability_score,
+    'file_name': Path(file_path).name,
+    'file_size': os.path.getsize(file_path),
+    'processing_time': datetime.now().isoformat(),
+    'summary_type': summary_type,
+    'summary_length': summary_length,
+    'model_used': 'Granite 3.2 8B' if (summary_type == "ai" and self.granite_integration and self.granite_integration.is_available()) else ('AI (BART/T5)' if self.summarizer else 'Extractive')
+}
+
+# In your UI section, add the questions display:
+# Add Granite questions if available
+granite_questions_html = ""
+if result.get("granite_questions"):
+    questions_list = "".join([f"<li style='margin-bottom: 10px; padding: 8px; background: rgba(255,255,255,0.1); border-radius: 6px;'>{q}</li>" 
+                             for q in result["granite_questions"]])
+    granite_questions_html = f'''
+    <div style="background: linear-gradient(135deg, #11998e 0%, #38ef7d 100%); color: white; padding: 20px; border-radius: 12px; margin: 15px 0; box-shadow: 0 6px 20px rgba(0,0,0,0.1);">
+        <h3>AI-Generated Questions</h3>
+        <p style="margin-bottom: 15px; opacity: 0.9;">Test your understanding with these Granite-generated questions:</p>
+        <ol style="padding-left: 20px; line-height: 1.6;">
+            {questions_list}
+        </ol>
+    </div>
+    '''
+
+# Update your system status display:
+**Granite 3.2 8B:** {"✅ Available" if (GRANITE_AVAILABLE and hasattr(summarizer, 'granite_integration') and summarizer.granite_integration and summarizer.granite_integration.is_available()) else "❌ Not Available"}