Add Alembic for database schema migrations and centralize database connection logic.

## Changes

### Database Migration (Alembic)
- Add alembic.ini configuration file
  * Script location: alembic/
  * File template: %%%(rev)s_%%%(slug)s
  * Timezone support: UTC

- Add alembic/env.py migration environment
  * Async SQLAlchemy support
  * TODO: Connect target_metadata to models for autogenerate

- Add alembic/script.py.mako migration template
  * Standard Alembic revision template

### Database Connection Module (app/database.py)
- Add get_database_url() utility function
  * Centralized URL construction logic (DRY principle)
  * URL-encodes credentials to handle special characters
  * Supports DATABASE_URL env var or individual components
  * Security: Prevents credential leak in error messages

### Application Updates (app/main.py)
- Refactor to use shared get_database_url() function
- Remove duplicate database URL construction code
- Improve error message clarity without exposing credentials

## Security Improvements
- URL-encode database credentials (urllib.parse.quote_plus)
- Prevent password exposure in error messages
- Secure connection string handling

## Benefits
- Consistent database URL handling across app
- Support for passwords with special characters
- Easy to test and maintain (single source of truth)
- Ready for schema version control

## Usage
```bash
# Create a new migration
alembic revision -m "description"

# Apply migrations
alembic upgrade head

# Rollback migrations
alembic downgrade -1
```

## TODO
- Implement lazy database initialization
- Add environment variable support in alembic.ini
- Connect target_metadata for autogenerate feature

## Files Changed (5 files)
- alembic.ini: Alembic configuration
- alembic/env.py: Migration environment setup
- alembic/script.py.mako: Migration template
- app/database.py: NEW - Database connection utilities
- app/main.py: Refactored to use shared database module

…-migration

- Remove unused logger import in app/database.py
- Add URL encoding for credentials in alembic/env.py
- Convert async DB URL to sync for Alembic migrations

- Normalize PostgreSQL URL schemes in app/database.py and alembic/env.py
  - Handle postgres://, postgresql://, postgresql+asyncpg:// formats
  - Convert all to postgresql+psycopg:// for consistency
- Add URL encoding for credentials in alembic/env.py
- Update alembic.ini comments to reflect actual implementation
- Add backward compatibility for MEMU_STORAGE_DIR env var in app/main.py

- Use quote(..., safe='') instead of quote_plus() for URL userinfo encoding
  (quote_plus encodes spaces as '+' which is incorrect for URL userinfo)
- Add asyncpg to psycopg conversion in app/database.py
  (consistent with alembic/env.py, asyncpg is not a project dependency)

…erate

- Create app/models/base.py with Base class (no DB connection on import)
- Update app/database.py to import Base from app.models.base
- Update alembic/env.py to set target_metadata = Base.metadata
- This enables 'alembic revision --autogenerate' to detect model changes

SQLAlchemy 2.x async_sessionmaker does not support the autocommit argument.
Use explicit transaction scopes instead.

Replace 'in' operator with startswith() to avoid false positives when
checking for asyncpg driver in database URLs. This ensures only URLs
that actually start with postgresql+asyncpg:// are converted.

Remove hard-coded defaults from get_sync_database_url() and add explicit
validation for required environment variables. This ensures consistent
behavior between Alembic migrations and the application, preventing
accidental connections to unintended databases.