Astrux JavaScript/TypeScript SDK

Official SDK for interacting with the Astrux API in JavaScript and TypeScript.

Installation

npm install astrux

Configuration

To use the SDK, you need an Astrux API key. You can provide it in two ways:

Environment variable (recommended)

export ASTRUX_API_KEY=your_api_key

Direct configuration

import { Astrux } from 'astrux';

const client = new Astrux({
  apiKey: 'your_api_key'
});

Usage

Basic prediction

import { Astrux } from 'astrux';

const client = new Astrux();

try {
  const result = await client.models.predict({
    model: 'model-name',
    input: {
      feature1: 'value1',
      feature2: 42
    }
  });

  console.log('Score:', result.score);
  console.log('Class:', result.class_);
  console.log('Probabilities:', result.proba);
} catch (error) {
  console.error('Error:', error.message);
}

Prediction with specific version

const result = await client.models.predict({
  model: 'model-name',
  input: { /* your data */ },
  version: 2
});

Timeout configuration

The default timeout is 30 seconds. You can customize it:

const client = new Astrux({
  apiKey: 'your_api_key',
  timeout: 60000 // 60 seconds
});

Response

The predict method returns a PredictResponse object with the following properties:

{
  score?: number;          // Prediction score (for regression)
  class_?: string;         // Predicted class (for classification)
  proba?: number[];        // Probabilities for each class (for classification)
  model_id?: string;       // Model identifier
  model_name?: string;     // Model name
  version?: number;        // Model version used
  task_type?: string;      // Task type (classification or regression)
}

Error handling

The SDK provides several specific error types:

import { 
  AstruxError,
  AuthenticationError,
  NotFoundError,
  RateLimitError,
  ValidationError,
  ServerError
} from 'astrux';

try {
  const result = await client.models.predict({ /* ... */ });
} catch (error) {
  if (error instanceof AuthenticationError) {
    console.error('Authentication issue:', error.message);
  } else if (error instanceof ValidationError) {
    console.error('Invalid data:', error.message);
  } else if (error instanceof RateLimitError) {
    console.error('Rate limit exceeded:', error.message);
  } else if (error instanceof NotFoundError) {
    console.error('Model not found:', error.message);
  } else if (error instanceof ServerError) {
    console.error('Server error:', error.message);
  }
}

Error types

• AuthenticationError (401): Missing or invalid API key
• NotFoundError (404): Model not found
• ValidationError (400, 422): Invalid parameters
• RateLimitError (429): Request rate limit exceeded
• ServerError (500+): Server-side error
• AstruxError: Generic error

All errors contain status (HTTP code) and payload (server response) properties when available.

Examples

Classification

const result = await client.models.predict({
  model: 'spam-classifier',
  input: {
    text: 'Congratulations! You won $1,000,000!'
  }
});

console.log(`Predicted class: ${result.class_}`);
console.log(`Confidence: ${result.score}`);
console.log(`Probabilities: ${result.proba}`);

Regression

const result = await client.models.predict({
  model: 'house-price-predictor',
  input: {
    bedrooms: 3,
    bathrooms: 2,
    square_feet: 1500,
    location: 'downtown'
  }
});

console.log(`Predicted price: $${result.score}`);

Support

• Website: https://astrux.io
• Documentation: https://docs.astrux.io
• Issues: https://github.com/astrux-labs/astrux-js/issues

Requirements

• Node.js >= 18

License

MIT © Thomas Bodénan