Files
TimeTrack/FLASK_MIGRATE_GUIDE.md

223 lines
5.4 KiB
Markdown

# Flask-Migrate Migration Guide
## Overview
TimeTrack has been refactored to use Flask-Migrate (which wraps Alembic) for database migrations instead of manual SQL scripts. This provides automatic migration generation, version control, and rollback capabilities.
**IMPORTANT**: The baseline for Flask-Migrate is set at git commit `4214e88d18fce7a9c75927753b8d4e9222771e14`. All schema changes after this commit need to be recreated as Flask-Migrate migrations.
**For Docker Deployments**: See `DOCKER_MIGRATIONS_GUIDE.md` for Docker-specific instructions (no Git required).
## Migration from Old System
### For Existing Deployments
If you have an existing database with the old migration system:
```bash
# 1. Install new dependencies
pip install -r requirements.txt
# 2. Establish baseline from commit 4214e88
python simple_baseline_4214e88.py
# Note: Use simple_baseline_4214e88.py as it handles the models.py transition correctly
# 3. Mark your database as being at the baseline
flask db stamp head
# 4. Apply any post-baseline migrations
# Review migrations_old/postgres_only_migration.py for changes after 4214e88
# Create new migrations for each feature:
flask db migrate -m "Add company updated_at column"
flask db migrate -m "Add user 2FA columns"
flask db migrate -m "Add company invitation table"
# etc...
```
### For New Deployments
```bash
# 1. Install dependencies
pip install -r requirements.txt
# 2. Initialize and create database
python manage_migrations.py init
python manage_migrations.py apply
```
## Daily Usage
### Creating Migrations
When you modify models (add columns, tables, etc.):
```bash
# Generate migration automatically
flask db migrate -m "Add user avatar field"
# Or use the helper script
python manage_migrations.py create -m "Add user avatar field"
```
**Always review the generated migration** in `migrations/versions/` before applying!
### Applying Migrations
```bash
# Apply all pending migrations
flask db upgrade
# Or use the helper script
python manage_migrations.py apply
```
### Rolling Back
```bash
# Rollback one migration
flask db downgrade
# Or use the helper script
python manage_migrations.py rollback
```
### Viewing Status
```bash
# Current migration version
flask db current
# Migration history
flask db history
# Or use the helper script
python manage_migrations.py history
```
## Important Considerations
### 1. PostgreSQL Enums
Flask-Migrate may not perfectly handle PostgreSQL enum types. When adding new enum values:
```python
# In the migration file, you may need to add:
from alembic import op
import sqlalchemy as sa
def upgrade():
# Add new enum value
op.execute("ALTER TYPE taskstatus ADD VALUE 'NEW_STATUS'")
```
### 2. Data Migrations
For complex data transformations, add custom code to migration files:
```python
def upgrade():
# Schema changes
op.add_column('user', sa.Column('new_field', sa.String()))
# Data migration
connection = op.get_bind()
result = connection.execute('SELECT id, old_field FROM user')
for row in result:
connection.execute(
f"UPDATE user SET new_field = '{process(row.old_field)}' WHERE id = {row.id}"
)
```
### 3. Production Deployments
The startup scripts have been updated to automatically run migrations:
```bash
# startup_postgres.sh now includes:
flask db upgrade
```
### 4. Development Workflow
1. Pull latest code
2. Run `flask db upgrade` to apply any new migrations
3. Make your model changes
4. Run `flask db migrate -m "Description"`
5. Review the generated migration
6. Test with `flask db upgrade`
7. Commit both model changes and migration file
## Troubleshooting
### "Target database is not up to date"
```bash
# Check current version
flask db current
# Force upgrade
flask db stamp head # Mark as latest without running
flask db upgrade # Apply any pending
```
### "Can't locate revision"
Your database revision doesn't match any migration file. This happens when switching branches.
```bash
# See all migrations
flask db history
# Stamp to a specific revision
flask db stamp <revision_id>
```
### Migration Conflicts
When multiple developers create migrations:
1. Merge the migration files carefully
2. Update the `down_revision` in the newer migration
3. Test thoroughly
## Best Practices
1. **One migration per feature** - Don't bundle unrelated changes
2. **Descriptive messages** - Use clear migration messages
3. **Review before applying** - Always check generated SQL
4. **Test rollbacks** - Ensure downgrade() works
5. **Backup before major migrations** - Especially in production
## Migration File Structure
```
migrations/
├── README.md # Quick reference
├── alembic.ini # Alembic configuration
├── env.py # Migration environment
├── script.py.mako # Migration template
└── versions/ # Migration files
├── 001_initial_migration.py
├── 002_add_user_avatars.py
└── ...
```
## Helper Scripts
- `manage_migrations.py` - Simplified migration management
- `migrate_to_alembic.py` - One-time transition from old system
- `init_migrations.py` - Quick initialization script
## Environment Variables
```bash
# Required for migrations
export FLASK_APP=app.py
export DATABASE_URL=postgresql://user:pass@host/db
```
## References
- [Flask-Migrate Documentation](https://flask-migrate.readthedocs.io/)
- [Alembic Documentation](https://alembic.sqlalchemy.org/)
- [SQLAlchemy Documentation](https://docs.sqlalchemy.org/)