PLOW-45 - trckr

← PLOW-45

feature

Created: 2025-12-21 • Updated: 2025-12-21

Relationships Loading...

## Summary

Add `--notify` CLI flag to send ntfy notifications on exit.

## Implementation

### CLI Changes

In `src/trckr_plow/cli.py`, add to both `run` and `tui` commands:

```python
@click.option(
    "--notify",
    is_flag=True,
    help="Send ntfy notification on exit (configure via ~/.config/trckr-plow/config.yml)",
)
```

### Exit Handler

Create helper function to wrap all exits:

```python
def notify_and_exit(
    code: int,
    reason: str,
    engine: WorkflowEngine | None,
    notify_flag: bool,
) -> None:
    """Send notification if enabled, then exit."""
    if notify_flag or config.enabled:
        from trckr_plow.notify import send_exit_notification, QueueStatus, SessionStats
        from trckr_plow.settings import load_notify_config
        
        config = load_notify_config()
        config.enabled = notify_flag or config.enabled
        
        # Get stats from engine state
        if engine:
            queue = QueueStatus(
                triage=engine.state.queue.triage,
                todo=engine.state.queue.todo,
                in_progress=engine.state.queue.in_progress,
                in_review=engine.state.queue.in_review,
                done=engine.state.queue.done,
                blocked=engine.state.queue.blocked,
            )
            session = SessionStats(
                completed=engine.state.session.issues_completed,
                failed=engine.state.session.issues_failed,
            )
            project = engine.state.project_key or "unknown"
        else:
            queue = QueueStatus()
            session = SessionStats()
            project = "unknown"
        
        send_exit_notification(project, reason, session, queue, config)
    
    sys.exit(code)
```

### Exit Points to Update

Update all exit points in `run()` and `tui()` commands:

| Exit Code | Reason String |
|-----------|---------------|
| 0 | "Workflow completed successfully" |
| 0 | "No issues to process - queue empty" |
| 1 | "Unexpected error: {error}" |
| 2 | "Configuration error: {error}" |
| 3 | "trckr command error: {error}" |
| 4 | "QA failure requires manual intervention" |
| 5 | "Workflow error: {error}" |
| 130 | "Interrupted by user" |

## Files to Modify

- `src/trckr_plow/cli.py` - Add --notify flag and exit handler
- `tests/test_cli.py` - Add tests for --notify flag

## Acceptance Criteria

- [ ] `--notify` flag exists on both run and tui commands
- [ ] Flag overrides config file (--notify forces notification)
- [ ] All exit points send notification when enabled
- [ ] Notification includes project, hostname, reason, session stats, queue status
- [ ] Works correctly even if engine is None (early exits)
- [ ] Unit tests verify flag behavior
- [ ] Pure Python implementation (no agents)

Pull Requests (comma-separated URLs)

Attachments

Comments (1)

agent · 2025-12-21

## Implementation Complete Added the `--notify` flag to both `run` and `tui` commands that sends ntfy notifications on workflow exit. ### Changes Made **New Files:** - `src/trckr_plow/notify.py` - Core notification module with QueueStatus, SessionStats, NotifyConfig dataclasses and send_exit_notification() - `src/trckr_plow/settings.py` - load_notify_config() to read config from ~/.config/trckr-plow/config.yml **Modified Files:** - `src/trckr_plow/cli.py` - Added notify_and_exit() helper, --notify flag to run (line 184) and tui (line 541) commands, updated all exit points **Test Files:** - `tests/test_notify.py` - 21 tests (100% coverage) - `tests/test_settings.py` - 14 tests (96% coverage) - `tests/test_cli.py` - Added 8 new tests for --notify flag ### Verification - All 777 tests pass - Lint clean - Merged to master and pushed