banko-ai-assistant 1.0.8__tar.gz → 1.0.9__tar.gz

{banko_ai_assistant-1.0.8/banko_ai_assistant.egg-info → banko_ai_assistant-1.0.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: banko-ai-assistant
-Version: 1.0.8
+Version: 1.0.9
 Summary: AI-powered expense analysis and RAG system with CockroachDB vector search and multi-provider AI support
 Author-email: Virag Tripathi <virag.tripathi@gmail.com>
 License-Expression: MIT

{banko_ai_assistant-1.0.8 → banko_ai_assistant-1.0.9}/banko_ai/vector_search/generator.py

@@ -224,56 +224,118 @@ class EnhancedExpenseGenerator:
         return expenses
     
     def save_expenses_to_database(self, expenses: List[Dict[str, Any]]) -> int:
-        """Save expenses to the database."""
-        try:
-            import pandas as pd
-            with self.engine.connect() as conn:
-                # Prepare data for insertion
-                data_to_insert = []
-                for expense in expenses:
-                    data_to_insert.append({
-                        'expense_id': expense['expense_id'],
-                        'user_id': expense['user_id'],
-                        'expense_date': expense['expense_date'],
-                        'expense_amount': expense['expense_amount'],
-                        'shopping_type': expense['shopping_type'],
-                        'description': expense['description'],
-                        'merchant': expense['merchant'],
-                        'payment_method': expense['payment_method'],
-                        'recurring': expense['recurring'],
-                        'tags': expense['tags'],
-                        'embedding': expense['embedding']
-                    })
-                
-                # Insert in batches
-                batch_size = 100
-                total_inserted = 0
-                
-                for i in range(0, len(data_to_insert), batch_size):
-                    batch = data_to_insert[i:i + batch_size]
-                    
-                    # Use pandas to insert the batch
-                    df = pd.DataFrame(batch)
-                    df.to_sql('expenses', conn, if_exists='append', index=False, method='multi')
-                    total_inserted += len(batch)
-                
-                return total_inserted
-                
-        except Exception as e:
-            print(f"Error saving expenses to database: {e}")
-            return 0
+        """Save expenses to the database with retry logic for CockroachDB."""
+        import pandas as pd
+        import time
+        import random
+        from sqlalchemy.exc import OperationalError
+        
+        # Prepare data for insertion
+        data_to_insert = []
+        for expense in expenses:
+            data_to_insert.append({
+                'expense_id': expense['expense_id'],
+                'user_id': expense['user_id'],
+                'expense_date': expense['expense_date'],
+                'expense_amount': expense['expense_amount'],
+                'shopping_type': expense['shopping_type'],
+                'description': expense['description'],
+                'merchant': expense['merchant'],
+                'payment_method': expense['payment_method'],
+                'recurring': expense['recurring'],
+                'tags': expense['tags'],
+                'embedding': expense['embedding']
+            })
+        
+        # Insert in smaller batches to reduce transaction conflicts
+        batch_size = 50  # Reduced from 100 to minimize conflicts
+        total_inserted = 0
+        
+        for i in range(0, len(data_to_insert), batch_size):
+            batch = data_to_insert[i:i + batch_size]
+            
+            # Retry logic for CockroachDB transaction conflicts
+            max_retries = 5
+            retry_count = 0
+            
+            while retry_count < max_retries:
+                try:
+                    with self.engine.connect() as conn:
+                        # Use pandas to insert the batch
+                        df = pd.DataFrame(batch)
+                        df.to_sql('expenses', conn, if_exists='append', index=False, method='multi')
+                        conn.commit()
+                        total_inserted += len(batch)
+                        break  # Success, exit retry loop
+                        
+                except OperationalError as e:
+                    # Check if it's a CockroachDB serialization failure (SQL state 40001)
+                    if "40001" in str(e) or "SerializationFailure" in str(e) or "restart transaction" in str(e).lower():
+                        retry_count += 1
+                        if retry_count < max_retries:
+                            # Exponential backoff with jitter
+                            base_delay = 0.1 * (2 ** retry_count)
+                            jitter = random.uniform(0, 0.1)
+                            delay = base_delay + jitter
+                            print(f"Transaction conflict detected, retrying in {delay:.2f}s (attempt {retry_count}/{max_retries})")
+                            time.sleep(delay)
+                            continue
+                        else:
+                            print(f"Max retries exceeded for batch {i//batch_size + 1}: {e}")
+                            return total_inserted
+                    else:
+                        # Non-retryable error
+                        print(f"Non-retryable database error: {e}")
+                        return total_inserted
+                        
+                except Exception as e:
+                    print(f"Unexpected error saving batch {i//batch_size + 1}: {e}")
+                    return total_inserted
+        
+        return total_inserted
     
     def clear_expenses(self) -> bool:
-        """Clear all expenses from the database."""
-        try:
-            from sqlalchemy import text
-            with self.engine.connect() as conn:
-                conn.execute(text("DELETE FROM expenses"))
-                conn.commit()
-                return True
-        except Exception as e:
-            print(f"Error clearing expenses: {e}")
-            return False
+        """Clear all expenses from the database with retry logic."""
+        import time
+        import random
+        from sqlalchemy import text
+        from sqlalchemy.exc import OperationalError
+        
+        max_retries = 5
+        retry_count = 0
+        
+        while retry_count < max_retries:
+            try:
+                with self.engine.connect() as conn:
+                    conn.execute(text("DELETE FROM expenses"))
+                    conn.commit()
+                    return True
+                    
+            except OperationalError as e:
+                # Check if it's a CockroachDB serialization failure (SQL state 40001)
+                if "40001" in str(e) or "SerializationFailure" in str(e) or "restart transaction" in str(e).lower():
+                    retry_count += 1
+                    if retry_count < max_retries:
+                        # Exponential backoff with jitter
+                        base_delay = 0.1 * (2 ** retry_count)
+                        jitter = random.uniform(0, 0.1)
+                        delay = base_delay + jitter
+                        print(f"Transaction conflict detected while clearing, retrying in {delay:.2f}s (attempt {retry_count}/{max_retries})")
+                        time.sleep(delay)
+                        continue
+                    else:
+                        print(f"Max retries exceeded while clearing expenses: {e}")
+                        return False
+                else:
+                    # Non-retryable error
+                    print(f"Non-retryable database error while clearing: {e}")
+                    return False
+                    
+            except Exception as e:
+                print(f"Unexpected error clearing expenses: {e}")
+                return False
+        
+        return False
     
     def get_expense_count(self) -> int:
         """Get the current number of expenses in the database."""

{banko_ai_assistant-1.0.8 → banko_ai_assistant-1.0.9/banko_ai_assistant.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: banko-ai-assistant
-Version: 1.0.8
+Version: 1.0.9
 Summary: AI-powered expense analysis and RAG system with CockroachDB vector search and multi-provider AI support
 Author-email: Virag Tripathi <virag.tripathi@gmail.com>
 License-Expression: MIT

{banko_ai_assistant-1.0.8 → banko_ai_assistant-1.0.9}/pyproject.toml

@@ -7,7 +7,7 @@ py-modules = []
 
 [project]
 name = "banko-ai-assistant"
-version = "1.0.8"
+version = "1.0.9"
 description = "AI-powered expense analysis and RAG system with CockroachDB vector search and multi-provider AI support"
 authors = [
     { name="Virag Tripathi", email="virag.tripathi@gmail.com" }