@obsidize/tar-browserify

Simple utility for packing and unpacking tar files in the browser.

Highlights:

• Zero dependencies
• No node-based requires/imports (fully compatible in browser)
• Inline extraction tools (examples below)
• Builder pattern for creating tarball files in-memory (examples below)
• Supports PAX header reading/writing

Pairs well with these modules:

• pako for gzip / unzip
• path-browserify to further process raw file names

Installation

npm install -P -E @obsidize/tar-browserify
Copy

Usage

Read a tar file

import { ungzip } from 'pako';
import { Archive } from '@obsidize/tar-browserify';

async function readTarFile() {
	const response = await fetch('url/to/some/file.tar.gz');
	const gzBuffer = await response.arrayBuffer();
	const tarBuffer = ungzip(gzBuffer);

	for await (const entry of Archive.read(tarBuffer)) {
		if (entry.isFile()) {
			console.log(`read tar file: ${entry.fileName} with content length ${entry.content!.byteLength}`);
			console.log(`file text contents: ${entry.text()}`);
			// TODO: do something interesting with the file
		}
	}
}

readTarFile().catch(console.error);
Copy

Write a tar file

import { gzip } from 'pako';
import { Archive } from '@obsidize/tar-browserify';

async function writeTarFile() {
	const tarBuffer = new Archive()
		.addDirectory('MyStuff')
		.addTextFile('MyStuff/todo.txt', 'This is my TODO list')
		.addBinaryFile('MyStuff/some-raw-file.obj', Uint8Array.from([1, 2, 3, 4, 5]))
		.addDirectory('Nested1')
		.addDirectory('Nested1/Nested2')
		.addBinaryFile('Nested1/Nested2/supersecret.bin', Uint8Array.from([6, 7, 8, 9]))
		.toUint8Array();

	const gzBuffer = gzip(tarBuffer);
	const fileToSend = new File([gzBuffer], 'my-awesome-new-file.tar.gz');
	// TODO: send the new file somewhere
}

writeTarFile().catch(console.error);
Copy

Modify an existing tar file

import { gzip, ungzip } from 'pako';
import { Archive } from '@obsidize/tar-browserify';

async function modifyTarFile() {
	const response = await fetch('url/to/some/file.tar.gz');
	const gzBuffer = await response.arrayBuffer();
	const tarBuffer = ungzip(gzBuffer);
	const archive = await Archive.extract(tarBuffer);

	const updatedTarBuffer = archive
		.removeEntriesWhere((entry) => /unwanted\-file\-name\.txt/.test(entry.fileName))
		.cleanAllHeaders() // remove unwanted metadata
		.addTextFile('new text file.txt', 'this was added to the original tar file!')
		.toUint8Array();

	const updatedGzBuffer = gzip(updatedTarBuffer);
	const fileToSend = new File([updatedGzBuffer], 'my-awesome-edited-file.tar.gz');
	// TODO: send the modified file somewhere
}

modifyTarFile().catch(console.error);
Copy

Read a BIG tar file (Advanced)

For very large files (>= 50MB) it is better to use the AsyncUint8ArrayLike interface for loading the file in chunks (either from disk or over a network)

import { Archive, AsyncUint8ArrayLike } from '@obsidize/tar-browserify';

async function readBigTarFile() {
  const customAsyncBuffer: AsyncUint8ArrayLike = {
    byteLength: 12345, // the file length in bytes must be known ahead of time
    read: (offset: number, length: number): Promise<Uint8Array> => /* TODO: return chunk from disk or network request */
  };

  for await (const entry of Archive.read(customAsyncBuffer)) {
    if (!entry.isFile()) {
      continue;
    }

	let totalBytesRead = 0;

	for await (const chunk of entry.getContentChunks()) {
	  totalBytesRead += chunk.byteLength;
      // TODO: do something interesting with the data chunk
	}
  }
}

readBigTarFile().catch(console.error);
Copy

API

Full API docs can be found in the repo GitHub Docs Page

Testing

This module has a full Test Suite to ensure breaking changes are not introduced, and is tested against the output of the node-tar package to ensure stability.

• npm test - run unit tests with live-reload.
• npm run coverage - perform a single-pass of unit tests with code-coverage display.