Files
TimeTracker/migrations
Dries Peeters 3e7b7ab787 feat: add per-user GitHub, Google Calendar, and Slack connectors
Three new opt-in integration connectors that plug into the existing
`app/integrations/base.py:BaseConnector` pattern and the integrations
settings UI. Each connector subclasses `BaseConnector`, persists its
state inside the existing `Integration.config` JSONB (no new tables),
encrypts every secret at rest via `app/utils/secret_crypto`, and
degrades gracefully when the integration row is missing or
`is_active=False` -- every method returns
`{"ok": false, "error": "Integration not configured"}` without
raising, so the timer, exports, and dashboards keep working when a
connector is disabled or broken.

All third-party HTTP calls go through `requests` with a 10-second
timeout and a `try/except requests.RequestException`. Tokens are
never written to logs in their raw form -- only short
`xoxb-...` / `ghp_...` truncations.

GitHub connector (`app/integrations/github_connector.py`, provider
key `github_connector`):
  - Webhook receiver at `POST /api/integrations/github/webhook`
    verifies `X-Hub-Signature-256` with HMAC-SHA256 against the
    per-integration webhook secret before reading the payload.
  - Handles `issues.opened` (creates a task with
    `external_ref="github_issue_{n}"`, mapped priority and `todo`
    status), `issues.assigned` (optionally starts a timer for the
    linked TimeTracker user when `users.github_username` matches),
    `issues.closed` (marks the existing task `done`), and `ping`.
  - Manual sync (`POST /api/integrations/github/sync`, admin only)
    pulls open issues from
    `GET /repos/{owner}/{repo}/issues?state=open&per_page=50` and
    upserts tasks by `external_ref`. Optional `label_filter`.

Google Calendar connector (`app/integrations/google_calendar_connector.py`,
provider key `google_calendar_connector`):
  - OAuth2 flow at `/integrations/google/{connect,callback,disconnect}`
    using raw `requests` against
    `https://oauth2.googleapis.com/token`. Tokens (`access_token`,
    `refresh_token`, `token_expiry`) are stored encrypted in
    `Integration.config`. `client_id`/`client_secret` come from
    Flask config (`GOOGLE_CLIENT_ID`, `GOOGLE_CLIENT_SECRET`) and are
    never hardcoded.
  - `_refresh_token_if_needed()` refreshes within 5 minutes of expiry
    on every API call.
  - `sync()` supports `import` / `export` / `both`:
    * import: pulls dated events from the configured `calendar_id`
      over the last `sync_days_back` days (clamped 1-30), skips
      all-day events and anything tagged `[TT]` or already linked via
      `gcal:{event_id}` in the notes of an existing `TimeEntry`.
    * export: posts completed entries created since `last_sync_at`
      back to Google as `[TT] {project} -- {task or notes}` events
      with `timeZone: "UTC"`.
  - `revoke()` calls `https://oauth2.googleapis.com/revoke` and
    wipes the stored tokens.
  - APScheduler job `google_calendar_sync` runs every 30 minutes;
    each user is wrapped in `try/except` so one broken token cannot
    block the rest.

Slack connector (`app/integrations/slack_connector.py`, provider key
`slack_connector`):
  - Webhook receiver at `POST /api/integrations/slack/events`
    verifies `X-Slack-Signature` (HMAC-SHA256 of
    `v0:{timestamp}:{body}`) and rejects requests older than 5
    minutes. Replies to Slack's URL verification handshake
    immediately.
  - Slash command `/tt` supports `start [project]` (id or
    case-insensitive partial name match against the user's allowed
    projects), `stop`, `status`, `today` (via
    `notification_service.get_today_summary_for_user`), and an
    in-place help text fallback. Every reply is ephemeral JSON so it
    fits inside Slack's 3-second budget without touching
    `response_url`.
  - `notify_timer_started` / `notify_timer_stopped` post a
    stopwatch/checkmark message to the configured channel. Wired
    into both the page route (`app/routes/timer.py`) and the JSON
    API (`app/routes/api.py`) as a fire-and-forget hook: the import
    + call are wrapped in `try/except` and only log at `DEBUG` on
    failure, so Slack outages can't slow down the timer flow.
  - `post_daily_summary` posts a daily roll-up; APScheduler job
    `slack_daily_summary` runs every 30 minutes and matches each
    user's configured `HH:MM` against the window.

Plumbing and storage:
  - New blueprint `app/routes/integrations_webhooks.py` registers
    the webhook receivers (`csrf.exempt`, signature-verified) plus a
    uniform `config`/`status`/`test` API surface
    (`/api/integrations/{github,google,slack}/{config,status,test}`)
    used by the settings UI. Optional-registered in
    `app/blueprint_registry.py`.
  - Alembic revision `155_add_integration_columns`:
    * `users.github_username` (String(100), nullable) - GitHub login
      join key for the assignment auto-start-timer flow.
    * `tasks.external_ref` (String(200), nullable, indexed) -
      canonical external id for connector-created tasks; the new
      index lets webhook receivers de-duplicate cheaply.
    Both columns are added defensively (inspector-checked) so the
    migration is safe to re-run.
  - New cards in `app/templates/integrations/_connector_cards.html`
    (included by `templates/integrations/list.html`) drive the
    Personal connectors UI -- Tailwind CSS only, vanilla JS, per-card
    status fetch, save, test, and sync actions.

Documentation:
  - `docs/integrations/README.md` indexes all built-in connectors.
  - `docs/integrations/GITHUB_CONNECTOR.md`,
    `docs/integrations/GOOGLE_CALENDAR.md`, and
    `docs/integrations/SLACK.md` cover setup, OAuth/webhook wiring,
    config fields, endpoints, and operational notes for each
    connector.
  - `docs/api/REST_API.md` lists the new endpoints under a new
    "Personal integration connectors" subsection.
  - `CHANGELOG.md` notes the feature under the [Unreleased] section.

`LLMService`, `TimeTrackingService`, `ForecastService`, and the
`Integration` model schema are intentionally untouched -- only
`users` and `tasks` gain columns via migration.
2026-05-15 10:45:30 +02:00
..

Database Migrations with Flask-Migrate

This directory contains the database migration system for TimeTracker, now standardized on Flask-Migrate with proper versioning.

Overview

The migration system has been updated from custom Python scripts to use Flask-Migrate, which provides:

  • Standardized migrations using Alembic
  • Version tracking for all database changes
  • Rollback capabilities to previous versions
  • Automatic schema detection from SQLAlchemy models
  • Cross-database compatibility (PostgreSQL, SQLite)

Quick Start

1. Initialize Migrations (First Time Only)

flask db init

2. Create Your First Migration

flask db migrate -m "Initial database schema"

3. Apply Migrations

flask db upgrade

Migration Commands

Basic Commands

  • flask db init - Initialize migrations directory
  • flask db migrate -m "Description" - Create a new migration
  • flask db upgrade - Apply pending migrations
  • flask db downgrade - Rollback last migration
  • flask db current - Show current migration version
  • flask db history - Show migration history

Advanced Commands

  • flask db show <revision> - Show specific migration details
  • flask db stamp <revision> - Mark database as being at specific revision
  • flask db heads - Show current heads (for branched migrations)

Migration Workflow

1. Make Model Changes

Edit your SQLAlchemy models in app/models/:

# Example: Add a new column
class User(db.Model):
    # ... existing fields ...
    phone_number = db.Column(db.String(20), nullable=True)

2. Generate Migration

flask db migrate -m "Add phone number to users"

3. Review Generated Migration

Check the generated migration file in migrations/versions/:

def upgrade():
    op.add_column('users', sa.Column('phone_number', sa.String(length=20), nullable=True))

def downgrade():
    op.drop_column('users', 'phone_number')

4. Apply Migration

flask db upgrade

5. Verify Changes

Check the migration status:

flask db current

Migration Files Structure

migrations/
├── versions/           # Migration files
│   ├── 001_initial_schema.py
│   ├── 002_add_phone_number.py
│   └── ...
├── env.py             # Migration environment
├── script.py.mako     # Migration template
├── alembic.ini        # Alembic configuration
└── README.md          # This file

Transition from Old System

If you're migrating from the old custom migration system:

1. Backup Your Database

# PostgreSQL
pg_dump --format=custom --dbname="$DATABASE_URL" --file=backup_$(date +%Y%m%d_%H%M%S).dump

# SQLite
cp instance/timetracker.db backup_timetracker_$(date +%Y%m%d_%H%M%S).db

2. Use Migration Management Script

python migrations/manage_migrations.py

3. Or Manual Migration

# Initialize Flask-Migrate
flask db init

# Create initial migration (captures current schema)
flask db migrate -m "Initial schema from existing database"

# Apply migration
flask db upgrade

Best Practices

1. Migration Naming

Use descriptive names for migrations:

flask db migrate -m "Add user profile fields"
flask db migrate -m "Create project categories table"
flask db migrate -m "Add invoice payment tracking"

2. Testing Migrations

Always test migrations on a copy of your production data:

# Test upgrade
flask db upgrade

# Test downgrade
flask db downgrade

# Verify data integrity

3. Backup Before Migrations

# Always backup before major migrations
flask db backup  # Custom command
# or
pg_dump --format=custom --dbname="$DATABASE_URL" --file=pre_migration_backup.dump

4. Review Generated Code

Always review auto-generated migrations before applying:

  • Check the upgrade() function
  • Verify the downgrade() function
  • Ensure data types and constraints are correct

Troubleshooting

Common Issues

1. Migration Already Applied

# Check current status
flask db current

# If migration is already applied, stamp the database
flask db stamp <revision>

2. Migration Conflicts

# Show migration heads
flask db heads

# Merge branches if needed
flask db merge -m "Merge migration branches" <revision1> <revision2>

3. Database Out of Sync

# Check migration history
flask db history

# Reset to specific revision
flask db stamp <revision>

4. Model Import Errors

Ensure all models are imported in your application:

# In app/__init__.py or similar
from app.models import User, Project, TimeEntry, Task, Settings, Invoice, Client

Getting Help

  1. Check the migration status: flask db current
  2. Review migration history: flask db history
  3. Check Alembic logs for detailed error messages
  4. Verify database connection and permissions

Advanced Features

Custom Migration Operations

You can add custom operations in your migrations:

def upgrade():
    # Custom data migration
    op.execute("UPDATE users SET role = 'user' WHERE role IS NULL")
    
    # Custom table operations
    op.create_index('custom_idx', 'table_name', ['column_name'])

Data Migrations

For complex data migrations, use the op.execute() method:

def upgrade():
    # Migrate existing data
    op.execute("""
        INSERT INTO new_table (id, name)
        SELECT id, name FROM old_table
    """)

Conditional Migrations

Handle different database types:

def upgrade():
    # PostgreSQL-specific operations
    if op.get_bind().dialect.name == 'postgresql':
        op.execute('CREATE EXTENSION IF NOT EXISTS "uuid-ossp"')

Environment Variables

Ensure these environment variables are set:

export FLASK_APP=app.py
export DATABASE_URL="postgresql://user:pass@localhost/dbname"
# or
export DATABASE_URL="sqlite:///instance/timetracker.db"

CI/CD Integration

For automated deployments, include migration steps:

# Example GitHub Actions step
- name: Run database migrations
  run: |
    flask db upgrade
  env:
    DATABASE_URL: ${{ secrets.DATABASE_URL }}

Support

For migration-related issues:

  1. Check this README
  2. Review Flask-Migrate documentation: https://flask-migrate.readthedocs.io/
  3. Check Alembic documentation: https://alembic.sqlalchemy.org/
  4. Review generated migration files for errors