node package manager

image-manipulation

Image Manipulation Methods

This library exposes a simple API for retrieving, sending, and manipulating images in the browser.

The demo allows the visitor to select a photo from his or her computer, crop and rotate the image, and upload directly to Google Cloud Storage using the gcs-signed-urls module to handle permissions.

See demo: http://blooming-bastion-8931.herokuapp.com/

Features

  • Hermite sampling is used to resize images rather than canvas's drawImage. This results in a much better quality photo after resizing. See this stackoverflow.
  • The library parses EXIF meta data to always provide the correct orientation. Photos taken with older cameras rely on EXIF metadata and the browser does not take this metadata into account. See this article.

Install

The library is available on bower and npm.

bower install image-manipulation

OR

npm install image-manipulation

You can access via "window.ImageMethods"

Examples

Manipulators

A canvas can be changed using the manipulator methods: rotate, resize and crop. One can use the static methods.

var canvas = document.querySelector("canvas");

var resizedCanvas = ImageMethods.resize(canvas, 100, 100);

var rotatedCanvas = ImageMethods.rotate(resizedCanvas, 90);

document.body.append(rotatedCanvas);

or one can also make a manipulator instance and chain these methods.

var canvas = document.querySelector("canvas");

var manipulator = new ImageMethods(canvas);

manipulator.resize(100, 100).rotate(90);

document.body.append(manipulator.canvas);

Retrieving images to manipulate

Grab an image from the DOM and flip it upside down

var img = document.querySelector("img"),
	canvas = ImageMethods.getCanvasFromImage(img);

img.src = ImageMethods.rotate(canvas, 180).toDataURL();

Grab an image from an input element (<input type="file" accept="image/*">), create a thumbnail at 200px width and add it to the screen.

document.querySelector("input[type=file]").onchange = function(e) {
	ImageMethods.getCanvasFromFile(e.files[0], function(canvas) {
		
		var manipulator = new ImageMethods(canvas);
		
		manipulator.resize(200);
		
		// Add our resized canvas to the screen
		document.body.appendChild(manipulator.canvas)
		
	});
};

Download an image from the server, cut it into 2 pieces, and upload the pieces back to the server via xhr2

ImageMethods.getCanvasFromUrl("/path/to/image.jpg", function(canvas, file) {
	
	var manipulator = new ImageMethods(canvas);
	
	var piece1Canvas = ImageMethods.crop(0, 0, canvas.width/2, canvas.height),
		piece2Canvas = ImageMethods.crop(canvas.width/2, 0, canvas.width/2, canvas.height),
	
	// Put together FormData for submission
	var formData = new FormData();
	formData.append("images[]", ImageMethods.toBlob(piece1Canvas), file.name);
	formData.append("images[]", ImageMethods.toBlob(piece2Canvas), file.name);
	
	// Post to server
	var xhr = new XMLHttpRequest();
	xhr.open("POST", "/my/upload-handler", true);
	
});

Reference

Chainable instance methods

// Create an manipulator instance likeso:
var instance = new ImageMethods(canvas)

// The canvas element can be exposed
instance.canvas

// You can convert your instance into a blob 
var blob = instance.toBlob();
rotate(degrees)
crop(x, y, width, height)
resize(width, height)

Static methods

These methods follow this form except getCanvasFromImage which simply returns a canvas element.

ImageMethods.getCanvasFromFile(file, function(canvas) {
	// Get access to the canvas element here
});
getCanvasFromImage(img)
getOrientationFromFile(file, callback)
getCanvasFromUrl(url, callback)
getCanvasFromFile(file, callback)

Testing

Karma

Karma/Jasmine is used for unit tests.

npm install
karma start
Testling

Testling integration can be tested likeso:

npm install
npm install -g testling
testling -u
// A localhost Url will be given to run tests. Simply paste it in the browser

NOTE: ci.testling.com integration was setup, but testling has had ongoing issues timing out service timeout issues

Protractor/WebdriverJS end-to-end test of the demo in Chrome

First run the app (requires you to setup Google Cloud Storage).

node example/app

No need to start a standalone selenium server, protractor uses the chromedriver installed via npm. Simply:

protractor

This process also creates these screenshots automatically:






Running the demo

Setup a google cloud storage with a service account as explained on the gcs-signed-urls page.

Drop your private key in the example directory.

Create gcs-config.js in the example directory with your Google Cloud Storage information like below

module.exports = {
    "storageBucket":    "storage-bucket",
    "servicesEmail":    "your-services-email@developer.gserviceaccount.com",
    "privateKey":       __dirname +"/google-services-private-key.pem"
};

Now you can run the app by running...

node example/app.js

The example will be visible on http://localhost:3001/

Pushing to Heroku

Remove "gcs-config.js" and "google-services-private-key.pem" from .gitignore

git commit -a
heroku create
git push heroku master

Future

  • Use web workers (with transferable objects) to process resize and add new sharpen feature. This will require async methods. Start using promises so we can continue to chain