@flyskywhy/gifuct-js

A Simple to use javascript .GIF decoder for Web and React Native.

We needed to be able to efficiently load and manipulate GIF files for the Ruffle hybrid app (for mobiles). There are a couple of example libraries out there like jsgif & its derivative libgif-js, however these are admittedly inefficient, and a mess. After pulling our hair out trying to understand the ancient, mystic gif format (hence the project name), we decided to just roll our own. This library also removes any specific drawing code, and simply parses, and decompresses gif files so that you can manipulate and display them however you like. We do include imageData of full frame or patch of patch frame construction though to get you most of the way there.

Demo

You can see a demo of this library in action here

Usage

Installation:

npm install @flyskywhy/gifuct-js

Decoding:

This decoder uses js-binary-schema-parser to parse the gif files (you can examine the schema in the source). This means the gif file must firstly be converted into a Uint8Array buffer in order to decode it. Some examples:

• fetch

    import { parseGIF, decompressFrames } from '@flyskywhy/gifuct-js'
  
    var promisedGif = fetch(gifURL)
         .then(resp => resp.arrayBuffer())
         .then(buff => {
             var gif = parseGIF(buff)
             var frames = decompressFrames(gif, true, true)
             return gif;
         });
  

• XMLHttpRequest

    import { parseGIF, decompressFrames } from '@flyskywhy/gifuct-js'
  
    var oReq = new XMLHttpRequest();
    oReq.open("GET", gifURL, true);
    oReq.responseType = "arraybuffer";
  
    oReq.onload = function (oEvent) {
        var arrayBuffer = oReq.response; // Note: not oReq.responseText
        if (arrayBuffer) {
            var gif = parseGIF(arrayBuffer);
            var frames = decompressFrames(gif, true, true);
            // do something with the frame data
        }
    };
  
    oReq.send(null);
  

Result:

The result of the decompressFrames(gif, buildPatch, buildImageData) function returns an array of all the GIF image frames, and their meta data. Here is a an example frame:

{
    // The color table lookup index for each pixel
    pixels: [...],
    // the dimensions of the gif frame (see disposal method)
    dims: {
        top: 0,
        left: 10,
        width: 100,
        height: 50
    },
    // the time in milliseconds that this frame should be shown
    delay: 50,
    // the disposal method (see below)
    disposalType: 1,
    // an array of colors that the pixel data points to
    colorTable: [...],
    // An optional color index that represents transparency (see below)
    transparentIndex: 33,
    // Uint8ClampedArray color converted patch information for drawing
    patch: [...]
    // ImageData for drawing
    imageData: [...]
 }

Automatic Patch Generation:

If the buildPatch param of the decompressFrames() function is true, the parser will not only return the parsed and decompressed gif frames, but will also create canvas ready Uint8ClampedArray arrays of each gif frame image, so that they and disposalType and dims can be drawn using ctx.putImageData() for example. This requirement is common, however it was made optional because it makes assumptions about transparency. The demo makes use of this option.

Note the patch maybe only a smaller patch on previous gif frame image, so the disposalType and dims will be needed to assist patch, ref to https://github.com/matt-way/gifuct-js/issues/35 or generateImageData() in src/index.js.

Automatic ImageData Generation:

If the buildPatch and buildImageData param of the decompressFrames() function are all true, the parser will create canvas ready ImageData of each gif frame image, so that they can easily be drawn using ctx.putImageData() for example.

Note the imageData is a full of each gif frame image, so the disposalType and dims are not needed.

Disposal Method:

The pixel data is stored as a list of indexes for each pixel. These each point to a value in the colorTable array, which contain the color that each pixel should be drawn. Each frame of the gif may not be the full size, but instead a patch that needs to be drawn over a particular location. The disposalType defines how that patch should be drawn over the gif canvas. In most cases, that value will be 1, indicating that the gif frame should be simply drawn over the existing gif canvas without altering any pixels outside the frames patch dimensions. More can be read about this here.

Transparency:

If a transparentIndex is defined for a frame, it means that any pixel within the pixel data that matches this index should not be drawn. When drawing the patch using canvas, this means setting the alpha value for this pixel to 0.

Drawing the GIF

Check out the demo for an example of how to draw/manipulate a gif using this library. We wanted the library to be drawing agnostic to allow users to do what they wish with the raw gif data, rather than impose a method that has to be altered. On this note however, we provide an easy interface for creating commonly used canvas pixel data for drawing ease.

Thanks to

We underestimated the convolutedness of the GIF format, so this library couldn't have been made without the help of:

• Project: What's In A GIF - Bit by Byte - An amazingly detailed blog by Matthew Flickinger
• jsgif
• The [almost correct] LZW decompression from this neat gist

Who are we?

Matt Way & Nick Drewe

Wethrift.com