Update rag_engine.py

rag_engine.py

@@ -52,12 +52,26 @@ class RAGEngine:
                 print(f"✅ Collection '{self.collection_name}' created!")
                 self._build_index()
             else:
-                count = self.client.count(collection_name=self.collection_name).count
-                if count == 0:
-                    self._build_index()
-                    self.load_from_huggingface() # Auto-load dataset if empty
-                else:
-                    print(f"✅ Qdrant Collection '{self.collection_name}' ready with {count} vectors.")
+                # Check if dataset is already indexed
+                dataset_filter = models.Filter(
+                    must=[
+                        models.FieldCondition(
+                            key="metadata.source",
+                            match=models.MatchValue(value="hf_dataset")
+                        )
+                    ]
+                )
+                dataset_count = self.client.count(
+                    collection_name=self.collection_name,
+                    count_filter=dataset_filter
+                ).count
+                
+                print(f"✅ Qdrant Collection '{self.collection_name}' ready with {count} vectors.")
+                
+                if dataset_count == 0:
+                    print("⚠️ Phishing dataset not found in collection. Loading...")
+                    self.load_from_huggingface()
+                    
         except Exception as e:
             print(f"⚠️ Collection check/creation failed: {e}")
             # Try to build anyway, maybe wrapper handles it
@@ -183,6 +197,7 @@ class RAGEngine:
             try:
                 self.client.delete_collection(self.collection_name)
                 self._build_index()
+                self.load_from_huggingface()
                 return "✅ Knowledge Base Refreshed on Cloud!"
             except Exception as e:
                 return f"❌ Error refreshing: {e}"