Sharp: feature request: bmp format image support

[Nomia]

this is a feature request,

would this lovely project support bmp format image manipulation?

[lovell]

Hello, please see #543

[Nomia]

Thanks I used jimp to handle bmp format images.

[Nomia]

here is the link: https://github.com/oliver-moran/jimp

read a bmp file, then export it as a jpeg buffer, then send it to sharp

[lovell]

I believe https://github.com/shaozilee/bmp-js is the underlying library used by jimp.

It might be possible to remove the jpeg compression round-trip by using it directly with sharp, something like (untested):

const bmpDecoded = bmp.decode(bmpBuffer);

sharp(bmpDecoded.data, { raw: {
  width: bmpDecoded.width,
  height: bmpDecoded.height,
  channels: 4
}})...

[Nomia]

Great, thanks!

[mooyoul]

Unfortunately, decoded channel order of bmpjs is BGR order, which is not compatible with sharp (sharp uses RGB order - RGBA if 4 channels defined.)

So we've forked bmp-js to provide RGBA option, with improved bitmap format compatibility. (tested with 1/2/4/8/16/24/32bit bitmap image)

Forked repo: https://github.com/balmbees/bmp-js

Just follow below steps to load bitmap images:

1. install @vingle/bmp-js package

$ npm install @vingle/bmp-js --save

1. Change your code like this:

Old:

import * as sharp from "sharp";

async function transform(input: Buffer): Promise<Buffer> {
  const sharpInstance = sharp(input);

   // do stuff...

  return await sharpInstance.jpeg().toBuffer();
}

New:

import * as bmp from "@vingle/bmp-js";
import * as sharp from "sharp";

const BUF_BMP = Buffer.from([0x42, 0x4d]); // "BM" file signature

function isBitmap(buf: Buffer): boolean {
  return Buffer.compare(BUF_BMP, buf.slice(0, 2)) === 0;
}

async function transform(input: Buffer): Promise<Buffer> {
  const sharpInstance = (() => {
    if (isBitmap(buf)) {
      const bitmap = bmp.decode(buf, true);
      return sharp(bitmap.data, {
        raw: {
          width: bitmap.width,
          height: bitmap.height,
          channels: 4,
        },
      });
    }
    return sharp(buf);
  })();

  // do stuff..

  return await sharpInstance.jpeg().toBuffer();
}

Hope this helps!