adding new bug fixes - changed the way meals are tracked and logged to make them copies not references

alembic/versions/31fdce040eea_snapshot_existing_meals.py

@@ -0,0 +1,101 @@
+"""snapshot_existing_meals
+
+Revision ID: 31fdce040eea
+Revises: 4522e2de4143
+Create Date: 2026-01-10 13:30:49.977264
+
+"""
+from typing import Sequence, Union
+from sqlalchemy import orm, text
+from app.database import Base
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = '31fdce040eea'
+down_revision: Union[str, None] = '4522e2de4143'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+
+def upgrade() -> None:
+    bind = op.get_bind()
+    session = orm.Session(bind=bind)
+
+    # Use reflection or raw SQL to avoid importing app models directly
+    # This ensures the migration remains valid even if app models change later
+    
+    # 1. Get all tracked meals that are NOT already snapshots
+    # We join tracked_meals with meals to check the meal_type
+    sql = text("""
+        SELECT tm.id as tracked_meal_id, tm.meal_id, m.name, m.meal_time
+        FROM tracked_meals tm
+        JOIN meals m ON tm.meal_id = m.id
+        WHERE m.meal_type != 'tracked_snapshot'
+    """)
+    
+    tracked_meals_to_snapshot = session.execute(sql).fetchall()
+    
+    print(f"Found {len(tracked_meals_to_snapshot)} tracked meals to snapshot.")
+    
+    for row in tracked_meals_to_snapshot:
+        tm_id = row.tracked_meal_id
+        original_meal_id = row.meal_id
+        original_name = row.name
+        original_meal_time = row.meal_time
+        
+        # 2. Create a new snapshot meal
+        # We can't easily use ORM since we don't have the classes, so we use raw SQL
+        insert_meal_sql = text("""
+            INSERT INTO meals (name, meal_type, meal_time)
+            VALUES (:name, 'tracked_snapshot', :meal_time)
+        """)
+        
+        # execution_options={"autocommit": True} might be needed for some drivers, 
+        # but session.execute usually handles it. 
+        # For SQLite, we can get the last inserted id via cursor, but SQLAlchemy does this via result.lastrowid
+        
+        result = session.execute(insert_meal_sql, {
+            "name": original_name, 
+            "meal_time": original_meal_time
+        })
+        new_meal_id = result.lastrowid
+        
+        # 3. Copy ingredients from original meal to new snapshot
+        # Get ingredients
+        get_foods_sql = text("""
+            SELECT food_id, quantity 
+            FROM meal_foods 
+            WHERE meal_id = :meal_id
+        """)
+        foods = session.execute(get_foods_sql, {"meal_id": original_meal_id}).fetchall()
+        
+        if foods:
+            insert_food_sql = text("""
+                INSERT INTO meal_foods (meal_id, food_id, quantity)
+                VALUES (:meal_id, :food_id, :quantity)
+            """)
+            
+            for food in foods:
+                session.execute(insert_food_sql, {
+                    "meal_id": new_meal_id,
+                    "food_id": food.food_id,
+                    "quantity": food.quantity
+                })
+        
+        # 4. Update the stored tracked_meal to point to the new snapshot
+        update_tm_sql = text("""
+            UPDATE tracked_meals
+            SET meal_id = :new_meal_id
+            WHERE id = :tm_id
+        """)
+        session.execute(update_tm_sql, {"new_meal_id": new_meal_id, "tm_id": tm_id})
+        
+    session.commit()
+
+
+
+def downgrade() -> None:
+    pass