Package Exports

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (@memberjunction/queue) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

@memberjunction/queue

A flexible queue management system for MemberJunction applications that enables background task processing, job scheduling, and asynchronous execution with database persistence.

Overview

The @memberjunction/queue package provides a robust framework for implementing persistent queues in MemberJunction applications. It offers:

Database-backed task persistence
Automatic queue creation and management
Concurrent task processing with configurable limits
Heartbeat monitoring for process health
Type-safe task definitions
Extensible queue implementations

Installation

npm install @memberjunction/queue

Dependencies

This package requires the following MemberJunction packages:

@memberjunction/core - Core functionality and entity management
@memberjunction/global - Global utilities and class registration
@memberjunction/core-entities - Entity type definitions
@memberjunction/ai - AI functionality (for AI-related queues)
@memberjunction/aiengine - AI Engine integration

Additional dependencies:

uuid - For generating unique identifiers

Core Components

TaskBase

The TaskBase class represents an individual task in a queue:

export class TaskBase {
  constructor(
    taskRecord: QueueTaskEntity,
    data: any,
    options: TaskOptions
  )
  
  // Properties
  ID: string              // Unique task identifier
  Status: TaskStatus      // Current task status
  Data: any              // Task payload data
  Options: TaskOptions   // Task configuration
  TaskRecord: QueueTaskEntity // Database entity
}

TaskStatus

Available task statuses:

export const TaskStatus = {
  Pending: 'Pending',
  InProgress: 'InProgress',
  Complete: 'Complete',
  Failed: 'Failed',
  Cancelled: 'Cancelled',
} as const;

QueueBase

The abstract QueueBase class serves as the foundation for all queue implementations:

export abstract class QueueBase {
  constructor(
    QueueRecord: QueueEntity,
    QueueTypeID: string,
    ContextUser: UserInfo
  )
  
  // Public methods
  AddTask(task: TaskBase): boolean
  FindTask(ID: string): TaskBase
  
  // Protected abstract method to implement
  protected abstract ProcessTask(
    task: TaskBase, 
    contextUser: UserInfo
  ): Promise<TaskResult>
}

QueueManager

The QueueManager is a singleton that manages all active queues:

export class QueueManager {
  // Singleton access
  static get Instance(): QueueManager
  
  // Static methods
  static async Config(contextUser: UserInfo): Promise<void>
  static async AddTask(
    QueueType: string,
    data: any,
    options: any,
    contextUser: UserInfo
  ): Promise<TaskBase | undefined>
  
  // Instance methods
  async AddTask(
    QueueTypeID: string,
    data: any,
    options: any,
    contextUser: UserInfo
  ): Promise<TaskBase | undefined>
}

TaskResult

Structure returned by task processing:

export class TaskResult {
  success: boolean      // Whether task completed successfully
  userMessage: string   // User-friendly message
  output: any          // Task output data
  exception: any       // Exception details if failed
}

Usage Examples

Basic Queue Implementation

Create a custom queue by extending QueueBase:

import { QueueBase, TaskBase, TaskResult } from '@memberjunction/queue';
import { RegisterClass } from '@memberjunction/global';
import { UserInfo } from '@memberjunction/core';

// Register your queue with a specific queue type name
@RegisterClass(QueueBase, 'Email Notification')
export class EmailNotificationQueue extends QueueBase {
  protected async ProcessTask(
    task: TaskBase, 
    contextUser: UserInfo
  ): Promise<TaskResult> {
    try {
      // Extract task data
      const { recipient, subject, body } = task.Data;
      
      // Implement your email sending logic here
      console.log(`Sending email to ${recipient}`);
      console.log(`Subject: ${subject}`);
      
      // Simulate email sending
      await this.sendEmail(recipient, subject, body);
      
      // Return success result
      return {
        success: true,
        userMessage: 'Email sent successfully',
        output: { sentAt: new Date() },
        exception: null
      };
    } catch (error) {
      // Return failure result
      return {
        success: false,
        userMessage: `Failed to send email: ${error.message}`,
        output: null,
        exception: error
      };
    }
  }
  
  private async sendEmail(to: string, subject: string, body: string) {
    // Your email service integration here
  }
}

Adding Tasks to Queue

import { QueueManager } from '@memberjunction/queue';
import { UserInfo } from '@memberjunction/core';

// Initialize queue manager (typically done once at app startup)
await QueueManager.Config(contextUser);

// Add a task using queue type name
const task = await QueueManager.AddTask(
  'Email Notification',  // Queue type name
  {                     // Task data
    recipient: 'user@example.com',
    subject: 'Welcome to MemberJunction',
    body: 'Thank you for joining!'
  },
  {                     // Task options
    priority: 1
  },
  contextUser
);

if (task) {
  console.log(`Task created with ID: ${task.ID}`);
}

AI Action Queue Example

The package includes built-in queues for AI operations:

import { AIActionQueue, EntityAIActionQueue } from '@memberjunction/queue';

// These queues are automatically registered and available for use
// Add an AI action task
const aiTask = await QueueManager.AddTask(
  'AI Action',
  {
    actionName: 'GenerateText',
    prompt: 'Write a product description',
    parameters: { maxTokens: 100 }
  },
  {},
  contextUser
);

// Add an entity-specific AI action
const entityAITask = await QueueManager.AddTask(
  'Entity AI Action',
  {
    entityName: 'Products',
    entityID: 123,
    actionName: 'GenerateDescription'
  },
  {},
  contextUser
);

Database Schema

The queue system requires the following database tables:

Queue Types Table (`__mj.QueueType`)

Stores definitions of different queue types (e.g., "Email Notification", "AI Action")

Queues Table (`__mj.Queue`)

Tracks active queue instances with process information:

Queue type reference
Process details (PID, platform, hostname)
Heartbeat timestamp
Network information

Queue Tasks Table (`__mj.QueueTask`)

Stores individual tasks:

Queue reference
Task status
Task data (JSON)
Task options (JSON)
Output and error information

Process Management

The QueueManager automatically captures process information for monitoring:

Process ID (PID)
Platform and version
Working directory
Network interfaces
Operating system details
User information
Heartbeat timestamps

This information helps track queue health and enables failover scenarios.

Configuration

Queue behavior can be configured through the implementation:

export class CustomQueue extends QueueBase {
  private _maxTasks = 5;        // Maximum concurrent tasks
  private _checkInterval = 500;  // Check interval in milliseconds
  
  // Override these values in your constructor
  constructor(QueueRecord: QueueEntity, QueueTypeID: string, ContextUser: UserInfo) {
    super(QueueRecord, QueueTypeID, ContextUser);
    // Customize queue behavior
    this._maxTasks = 10;
    this._checkInterval = 1000;
  }
}

Best Practices

Task Data Structure: Keep task data serializable as JSON
Error Handling: Always return proper TaskResult with error details
Queue Registration: Use @RegisterClass decorator for automatic registration
Idempotency: Design tasks to be safely retryable
Resource Cleanup: Clean up resources in finally blocks
Monitoring: Check heartbeat timestamps for queue health

Integration with MemberJunction

The queue system integrates seamlessly with:

Entity System: Use entities for task data and processing
User Context: All operations respect user permissions
Global Registry: Automatic queue discovery via class registration
AI Engine: Built-in support for AI task processing

Build & Development

# Build the package
npm run build

# Development mode with auto-reload
npm run start

# TypeScript compilation only
npm run build

License

ISC