office_translator/scripts/migrate_to_db.py

"""
Migration script to move data from JSON files to database
Run this once to migrate existing users to the new database system
"""
import json
import os
import sys
from pathlib import Path
from datetime import datetime

# Add parent directory to path
sys.path.insert(0, str(Path(__file__).parent.parent))

from database.connection import init_db, get_db_session
from database.repositories import UserRepository
from database.models import PlanType, SubscriptionStatus


def migrate_users_from_json():
    """Migrate users from JSON file to database"""
    json_path = Path("data/users.json")
    
    if not json_path.exists():
        print("No users.json found, nothing to migrate")
        return 0
    
    # Initialize database
    print("Initializing database tables...")
    init_db()
    
    # Load JSON data
    with open(json_path, 'r') as f:
        users_data = json.load(f)
    
    print(f"Found {len(users_data)} users to migrate")
    
    migrated = 0
    skipped = 0
    errors = 0
    
    with get_db_session() as db:
        repo = UserRepository(db)
        
        for user_id, user_data in users_data.items():
            try:
                # Check if user already exists
                existing = repo.get_by_email(user_data.get('email', ''))
                if existing:
                    print(f"  Skipping {user_data.get('email')} - already exists")
                    skipped += 1
                    continue
                
                # Map plan string to enum
                plan_str = user_data.get('plan', 'free')
                try:
                    plan = PlanType(plan_str)
                except ValueError:
                    plan = PlanType.FREE
                
                # Map subscription status
                status_str = user_data.get('subscription_status', 'active')
                try:
                    status = SubscriptionStatus(status_str)
                except ValueError:
                    status = SubscriptionStatus.ACTIVE
                
                # Create user with original ID
                from database.models import User
                user = User(
                    id=user_id,
                    email=user_data.get('email', '').lower(),
                    name=user_data.get('name', ''),
                    password_hash=user_data.get('password_hash', ''),
                    email_verified=user_data.get('email_verified', False),
                    avatar_url=user_data.get('avatar_url'),
                    plan=plan,
                    subscription_status=status,
                    stripe_customer_id=user_data.get('stripe_customer_id'),
                    stripe_subscription_id=user_data.get('stripe_subscription_id'),
                    docs_translated_this_month=user_data.get('docs_translated_this_month', 0),
                    pages_translated_this_month=user_data.get('pages_translated_this_month', 0),
                    api_calls_this_month=user_data.get('api_calls_this_month', 0),
                    extra_credits=user_data.get('extra_credits', 0),
                )
                
                # Parse dates
                if user_data.get('created_at'):
                    try:
                        user.created_at = datetime.fromisoformat(user_data['created_at'].replace('Z', '+00:00'))
                    except:
                        pass
                
                if user_data.get('updated_at'):
                    try:
                        user.updated_at = datetime.fromisoformat(user_data['updated_at'].replace('Z', '+00:00'))
                    except:
                        pass
                
                db.add(user)
                db.commit()
                
                print(f"  Migrated: {user.email}")
                migrated += 1
                
            except Exception as e:
                print(f"  Error migrating {user_data.get('email', user_id)}: {e}")
                errors += 1
                db.rollback()
    
    print(f"\nMigration complete:")
    print(f"  Migrated: {migrated}")
    print(f"  Skipped: {skipped}")
    print(f"  Errors: {errors}")
    
    # Backup original file
    if migrated > 0:
        backup_path = json_path.with_suffix('.json.bak')
        os.rename(json_path, backup_path)
        print(f"\nOriginal file backed up to: {backup_path}")
    
    return migrated


def verify_migration():
    """Verify the migration was successful"""
    from database.connection import get_db_session
    from database.repositories import UserRepository
    
    with get_db_session() as db:
        repo = UserRepository(db)
        count = repo.count_users()
        print(f"\nDatabase now contains {count} users")
        
        # List first 5 users
        users = repo.get_all_users(limit=5)
        if users:
            print("\nSample users:")
            for user in users:
                print(f"  - {user.email} ({user.plan.value})")


if __name__ == "__main__":
    print("=" * 50)
    print("JSON to Database Migration Script")
    print("=" * 50)
    
    # Check environment
    db_url = os.getenv("DATABASE_URL", "")
    if db_url:
        print(f"Database: PostgreSQL")
    else:
        print(f"Database: SQLite (development)")
    
    print()
    
    # Run migration
    migrate_users_from_json()
    
    # Verify
    verify_migration()