@anygpt/mock

Mock connector for testing and development with the AnyGPT router system.

Features

• Realistic Mock Responses: Generates contextual responses based on input
• Configurable Delays: Simulate API latency for testing
• Failure Simulation: Test error handling with configurable failure rates
• Custom Responses: Override responses for specific test scenarios
• Multiple Mock Models: Simulates GPT-4, GPT-3.5, and Claude models
• Zero External Dependencies: Perfect for testing without API keys

Installation

npm install @anygpt/mock @anygpt/types

Usage

Basic Usage

import { MockConnectorFactory } from '@anygpt/mock';
import { GenAIRouter } from '@anygpt/router';

const router = new GenAIRouter();
router.registerConnector(new MockConnectorFactory());

// Use the mock connector
const response = await router.chatCompletion({
  provider: 'mock',
  model: 'mock-gpt-4',
  messages: [{ role: 'user', content: 'Hello!' }]
});

Configuration

import { MockConnector } from '@anygpt/mock';

const mockConnector = new MockConnector({
  delay: 500,        // 500ms simulated latency
  failureRate: 0.1,  // 10% chance of failure
  customResponses: {
    'chat_mock-gpt-4': {
      id: 'custom-response',
      // ... custom response object
    }
  }
});

Dynamic Configuration

// In your config file
export default {
  providers: {
    'test-provider': {
      connector: '@anygpt/mock',
      api: { url: 'mock://test' },
      config: {
        delay: 200,
        failureRate: 0.05
      }
    }
  }
};

Testing Utilities

const mockConnector = new MockConnector();

// Set custom responses for testing
mockConnector.setCustomResponse('chat_test-model', {
  id: 'test-123',
  choices: [{ message: { content: 'Expected test response' } }]
});

// Simulate network issues
mockConnector.setFailureRate(0.5); // 50% failure rate
mockConnector.setDelay(2000);      // 2 second delay

Mock Models

The connector provides several mock models:

• mock-gpt-4: High-capability model simulation
• mock-gpt-3.5-turbo: Fast model simulation
• mock-claude-3: Vision-capable model simulation

Configuration Options

• delay: Simulated API response delay in milliseconds (default: 100)
• failureRate: Probability of simulated failures (0-1, default: 0)
• customResponses: Override responses for specific scenarios
• timeout: Request timeout (inherited from base config)
• maxRetries: Maximum retry attempts (inherited from base config)

Use Cases

• Unit Testing: Test your application without external API calls
• Development: Develop against consistent, predictable responses
• Integration Testing: Test error handling and retry logic
• Performance Testing: Simulate various latency scenarios
• Offline Development: Work without internet connectivity

Response Patterns

The mock connector generates contextual responses:

• Hello messages: Returns friendly greetings
• Test queries: Confirms testing functionality
• Error keywords: Simulates error scenarios
• Default: Random contextual responses

This makes testing more realistic while remaining predictable for assertions.