fetchT

Type-safe Fetch API wrapper with abortable requests, timeout support, progress tracking, automatic retry, and Rust-like Result error handling.

---

中文 | API Documentation

---

Features

• Abortable Requests - Cancel requests anytime via FetchTask.abort()
• Type-safe Responses - Specify return type with responseType parameter (text, json, arraybuffer, bytes, blob, stream)
• Timeout Support - Auto-abort requests after specified milliseconds
• Progress Tracking - Monitor download progress with onProgress callback
• Chunk Streaming - Access raw data chunks via onChunk callback
• Automatic Retry - Configurable retry strategies with retry option
• Result Error Handling - Rust-like Result type for explicit error handling
• Cross-platform - Works with Deno, Node.js, Bun, and browsers

Installation

# npm
npm install @happy-ts/fetch-t

# yarn
yarn add @happy-ts/fetch-t

# pnpm
pnpm add @happy-ts/fetch-t

# JSR (Deno)
deno add @happy-ts/fetch-t

# JSR (Bun)
bunx jsr add @happy-ts/fetch-t

Quick Start

Basic Usage

import { fetchT } from '@happy-ts/fetch-t';

// GET JSON data
const result = await fetchT<{ id: number; title: string }>('https://api.example.com/data', {
    responseType: 'json',
});

result.inspect(data => {
    console.log(data.title);
}).inspectErr(err => {
    console.error('Request failed:', err.message);
});

Abortable Request

const task = fetchT('https://api.example.com/large-file', {
    abortable: true,
    responseType: 'arraybuffer',
});

// Abort after 5 seconds
setTimeout(() => {
    task.abort('User cancelled');
}, 5000);

const result = await task.result;

Automatic Retry

const result = await fetchT('https://api.example.com/data', {
    retry: {
        retries: 3,
        delay: (attempt) => Math.min(1000 * Math.pow(2, attempt - 1), 10000),
        when: [500, 502, 503, 504],
        onRetry: (error, attempt) => console.log(`Retry ${attempt}: ${error.message}`),
    },
    responseType: 'json',
});

Examples

• Basic - Basic fetch requests
• Progress Tracking - Download progress and chunk streaming
• Abortable - Cancel and timeout requests
• Retry - Automatic retry strategies
• Error Handling - Error handling patterns

Error Handling Design

fetchT distinguishes between two types of errors:

Programming Errors (Synchronous)

Invalid parameters throw immediately for fail-fast behavior:

// These throw synchronously - no try/catch around await needed
fetchT('https://example.com', { timeout: -1 });     // Error: timeout must be > 0
fetchT('https://example.com', { timeout: 'bad' });  // TypeError: timeout must be a number
fetchT('not-a-url');                                // TypeError: Invalid URL

This differs from native fetch, which returns rejected Promises for parameter errors. Synchronous throws provide clearer stack traces and catch bugs during development.

Runtime Errors (Result Type)

Network failures and HTTP errors are wrapped in Result type:

const result = await fetchT('https://api.example.com/data', { responseType: 'json' });

if (result.isErr()) {
    const error = result.unwrapErr();
    // FetchError with status code, or network Error
}

License

MIT