gulp-imgconv

A gulp plugin to convert images (format conversion, resizing, cutting in/out, watermarking etc.) for distribution or batch image processing

Installation

npm install gulp-imgconv --save-dev

Change Log

• v0.9.x: refactored pipeline construction to Fluent-API (or method-chainning model)

Usage

• DISTRIBUTION: resize all images under src/, cutin with a round shape, add a watermark, sharpen, grascale, boolean with a 5-pointed star and save them to dst/ in PNG format

const gulp = require('gulp'), 
    gic = require('gulp-imgconv');
 
exports.imgconv = () => {
    gulp.src('src/*')
    .pipe(gic(gic
        .begin()
        .resize({
            width: 640,
            height: 480,
            fit: 'contain',
            background: '#00000000'    
        })
        .cutin(Buffer.from(`<?xml version="1.0" encoding="utf-8"?>
            <svg xmlns="http://www.w3.org/2000/svg" width="360" height="360" viewBox="0 0 480 480">
                <circle r="240" cx="240" cy="240"/>
            </svg>`))
        .watermark('png/watermark.png', {
            left: 400,
            top: 280
        })
        .sharpen()
        .grayscale()
        .boolean(Buffer.from(`<?xml version="1.0" encoding="utf-8"?>
            <svg xmlns="http://www.w3.org/2000/svg" width="255" height="240" viewBox="-20 0 71 48">
                <title>Five Pointed Star</title>
                <path fill="none" stroke="#000" d="m25,1 6,17h18l-14,11 5,17-15-10-15,10 5-17-14-11h18z"/>
            </svg>`),
            'eor'
        )
        .toFormat('png', {
            quality: 80
        })
        .commit()
    ))
    .pipe(gulp.dest('dst/')); 
};

Click below links to see the conversion effect

Original

Converted

• PROCESSING: resize the original images in src/ to width = 800, sharpen, grayscale them and save to dst/ in corresponding formats (SVG, GIF -> PNG, otherwise the original format)

const gulp = require('gulp'), 
gic = require('gulp-imgconv');
 
exports.imgconv = () => {
gulp.src('src/*')
    .pipe(gic(gic
        .begin()
        .resize({
            width: 800, 
            fit: 'contain',
            background: '#00000000'    
        })
        .sharpen()
        .grayscale()
        .commit()
    ))
    .pipe(gulp.dest('dst/')); 
};

API

The only argument is a pipeline in Array design, and there're 2 basic functions:

• resize(widith?: number, height?: number, opts?: {[k: string]: any}) - now a direct mapping to Sharp's resize function
  • Where width and height are size to resize to; when any parameter is ommitted, its value will be determined by fit option. Please refer to resizeOptsHelper object of this package to learn how to make up a correct opts object.
• toFormat(fmt: string, opts?: {[k: string]: any})
  • Where fmt is the format to convert to (right now supports jpeg, png, tiff and webp). There are quite a few option choices for each format, please find more details from sharp official document

And 3 overlaying related featured functions inspired by my past experiences:

• cutin/cutout/watermark(src: Buffer | string, opts?: {[k: string]: any}) - an encapsulation of Sharp's composite function
  • Where src is either the svg/png file name (in string) or the data (in Buffer) to overlay upon the original image, and you can learn how to construct the basic opts argument with the help of compositeOptsHelper, or read sharp official document to comprehensively understand the exact meaning of each option.

To construct this pipeline array, you can use chained method calling started by begin() and ended by commit().

Moreover, almost all other transformation related functions of sharp are supported with the same function prototype, please feel free to use like what I did in the test gulpfile.js file:

• extend, extract, trim
• rotate, flip, flop
• sharpen, median, blur
• flatten, gamma, negate, linear, normalize, convolve, threshould, recomb, modulate
• tint, grayscale, toColorspace
• removeAlpha, ensureAlpha, extractChannel, joinChannel, bandbool
• composite, boolean

Thanks

Special gratitude to sharp and other dependencies

Author

tibetty [e-mail]