Update app.py

app.py

@@ -56,20 +56,94 @@ def chunk_document(text):
     return chunks
 
 def test_single_model(model_name, chunks, question):
-    # ... existing code until embedding creation
+    """Test embedding with a single model"""
+    print(f"\n🤖 TESTING MODEL: {model_name}")
+    print(f"📊 Input data: {len(chunks)} chunks, question: '{question[:50]}...'")
+    log_memory_usage("Before model loading")
     
-    print(f"🧮 Creating embeddings in batches...")
-    batch_size = 5  # Process 5 chunks at a time
-    chunk_embeddings = []
-    
-    for i in range(0, len(chunks), batch_size):
-        batch = chunks[i:i+batch_size]
-        print(f"📦 Processing batch {i//batch_size + 1}: chunks {i+1}-{min(i+batch_size, len(chunks))}")
-        batch_embeddings = model.encode(batch, show_progress_bar=False)
-        chunk_embeddings.append(batch_embeddings)
-    
-    chunk_embeddings = np.vstack(chunk_embeddings)
-    print(f"📐 All chunk embeddings shape: {chunk_embeddings.shape}")
+    try:
+        start_time = time.time()
+        
+        # Load model FIRST - this defines the 'model' variable
+        print(f"📥 Loading SentenceTransformer model: {model_name}")
+        model = SentenceTransformer(model_name)
+        load_time = time.time() - start_time
+        print(f"✅ Model loaded successfully in {load_time:.2f}s")
+        log_memory_usage("After model loading")
+        
+        # Create embeddings in batches
+        print(f"🧮 Creating embeddings in batches...")
+        embed_start = time.time()
+        
+        batch_size = 5  # Process 5 chunks at a time
+        chunk_embeddings = []
+        
+        for i in range(0, len(chunks), batch_size):
+            batch = chunks[i:i+batch_size]
+            print(f"📦 Processing batch {i//batch_size + 1}: chunks {i+1}-{min(i+batch_size, len(chunks))}")
+            batch_embeddings = model.encode(batch, show_progress_bar=False)
+            chunk_embeddings.append(batch_embeddings)
+            print(f"✅ Batch {i//batch_size + 1} completed, shape: {batch_embeddings.shape}")
+        
+        # Combine all batch embeddings
+        chunk_embeddings = np.vstack(chunk_embeddings)
+        print(f"📐 All chunk embeddings combined shape: {chunk_embeddings.shape}")
+        
+        print(f"❓ Encoding question...")
+        question_embedding = model.encode([question], show_progress_bar=False)
+        print(f"📐 Question embedding shape: {question_embedding.shape}")
+        
+        embed_time = time.time() - embed_start
+        print(f"✅ All embeddings created in {embed_time:.2f}s")
+        log_memory_usage("After embedding creation")
+        
+        # Calculate similarities
+        print(f"📊 Calculating cosine similarities...")
+        similarities = cosine_similarity(question_embedding, chunk_embeddings)[0]
+        print(f"📈 Similarity scores - Min: {similarities.min():.3f}, Max: {similarities.max():.3f}, Mean: {similarities.mean():.3f}")
+        
+        # Get top K results
+        print(f"🔝 Finding top {TOP_K_RESULTS} results...")
+        top_indices = np.argsort(similarities)[-TOP_K_RESULTS:][::-1]
+        print(f"🏆 Top indices: {top_indices.tolist()}")
+        
+        for i, idx in enumerate(top_indices):
+            score = similarities[idx]
+            print(f"📌 Result {i+1}: Chunk #{idx}, Score: {score:.3f}")
+        
+        total_time = time.time() - start_time
+        print(f"⏱️ Total processing time: {total_time:.2f}s")
+        
+        results = {
+            'status': 'success',
+            'total_time': total_time,
+            'load_time': load_time,
+            'embed_time': embed_time,
+            'top_chunks': [
+                {
+                    'index': idx,
+                    'score': similarities[idx],
+                    'text': chunks[idx]
+                }
+                for idx in top_indices
+            ]
+        }
+        
+        print(f"✅ Model {model_name} completed successfully!")
+        return results
+        
+    except Exception as e:
+        print(f"❌ ERROR in model {model_name}:")
+        print(f"🐛 Error type: {type(e).__name__}")
+        print(f"💬 Error message: {str(e)}")
+        print(f"📋 Full traceback:")
+        print(traceback.format_exc())
+        
+        return {
+            'status': 'failed',
+            'error': str(e),
+            'traceback': traceback.format_exc()
+        }
 
 def process_embeddings(document_text, progress=gr.Progress()):
     """Process document with all embedding models"""