Datagrok Logging & Notification API

Help the user choose and implement the correct logging/notification approach in Datagrok packages.

Usage

/datagrok-logging

User-Facing Notifications (Balloons)

Toast notifications shown to the user. Primary methods for communicating from package code.

import * as grok from 'datagrok-api/grok';

// Green balloon — success or informational
grok.shell.info('Operation completed');

// Red balloon — error
grok.shell.error('Something went wrong');

// Yellow balloon — warning
grok.shell.warning('Check your settings');

All three accept string | HTMLElement and an optional BalloonOptions:

interface BalloonOptions {
  oneTimeKey?: string;   // Show only once per key (prevents repeated identical messages)
  copyText?: string;     // Text copied to clipboard on click
  autoHide?: boolean;    // Auto-hide after timeout (default: true)
  timeout?: number;      // Timeout in seconds (default: 5)
}

grok.shell.error('Failed to connect', { timeout: 10 });
grok.shell.info('Copied!', { oneTimeKey: 'copy-hint', autoHide: true });

Server-Side Logging (DG.Logger)

For audit trails, usage tracking, and debug logging recorded on the Datagrok server. These do NOT show UI notifications.

import * as DG from 'datagrok-api/dg';

// Create a logger (optionally with default params attached to every entry)
const logger = DG.Logger.create({ params: { source: 'MyPackage' } });

// Log levels
logger.debug('Detailed diagnostic info', { step: 'init' });
logger.info('Normal operation', { action: 'loaded' });
logger.warning('Potential issue', { config: 'missing' });
logger.error('Something failed', { context: 'upload' }, stackTrace);
logger.audit('User did something', { item: 'report' });
logger.usage('Feature used', { feature: 'export' });

PackageLogger

Automatically tags log entries with the package name:

const logger = new DG.PackageLogger(_package);
logger.error('Connection failed');  // tagged with package name

LOG_LEVEL enum

DG.LOG_LEVEL.DEBUG   // 'debug'
DG.LOG_LEVEL.INFO    // 'info'
DG.LOG_LEVEL.WARNING // 'warning'
DG.LOG_LEVEL.ERROR   // 'error'
DG.LOG_LEVEL.AUDIT   // 'audit'
DG.LOG_LEVEL.USAGE   // 'usage'

Progress Indicator with Logging

For long-running operations with status updates shown in the task bar:

const pi = DG.TaskBarProgressIndicator.create('Processing...', { cancelable: true });
pi.update(50, 'Half done');
pi.log('Step 1 finished');   // Append to progress log
pi.close();

Log Event Stream

Subscribe to all log events in real time:

grok.events.onLog.subscribe((msg) => {
  console.log(`[${msg.level}] ${msg.message}`, msg.params);
});

When to Use What

Scenario	Method
Tell the user something succeeded	grok.shell.info()
Show a user-facing error	grok.shell.error()
Show a user-facing warning	grok.shell.warning()
Log for debugging (server-side)	logger.debug() / logger.info()
Record errors for diagnostics	logger.error(message, params, stackTrace)
Track feature usage	logger.usage()
Audit user actions	logger.audit()
Show progress for long ops	DG.TaskBarProgressIndicator
Internal dev logging (not recorded)	console.log() / console.warn()

Rules

• Never use console.log for production logging — use DG.Logger for server-side or grok.shell.* for user-facing.
• console.warn / console.error are acceptable for development diagnostics but won't be recorded on the server.
• Prefer grok.shell.warning() over grok.shell.error() for non-critical issues (e.g., missing optional config).
• Use grok.shell.error() for failures that block the user's workflow.
• Use oneTimeKey when a notification could fire repeatedly (e.g., in a loop or event handler).