PixelStream is a base transform stream class for image pixel data.
It propagates image metadata such as size and color space between piped streams,
and makes working with images of multiple frames (e.g. animated GIFs!) much easier.
npm install pixel-stream
PixelStream is a Node transform stream.
You can write data to it manually, or pipe data to it from another stream. You can pipe more than one
frame of data (as in an animated image), and the PixelStream will handle this properly.
The constructor for a PixelStream accepts three optional arguments: width, height, and an
object for other options. If you are not piping another stream into this one that has these
properties, width and height are required. One additional option handled by the PixelStream
base class is colorSpace, described below, which is set to 'rgb' by default. Other options can
be handled by subclasses.
An object describing characteristics about the image, such as its width, height,
colorSpace (e.g. rgb, rgba, gray, cmyk, etc.), and other properties.
This method adds a frame metadata object describing the characteristics of a frame in an
animated image. This does not include the actual pixel data for the frame, which is written
through the stream in the usual way. It just describes characteristics such as frame size, etc.
This frame object can be used by PixelStream subclasses, such as encoders, and is
passed on to PixelStreams further down the pipes by emitting frame events (described below).
If this event is emitted by a source stream (e.g. image decoder), which is then piped to a
PixelStream, the PixelStream will use this opportunity to learn about the above image
characteristics from the source stream automatically. A format object, as described above,
should be passed as an argument to the event.
fs.createReadStream('in.png')
.pipe(new PNGDecoder)
.pipe(new MyPixelStream)In the above example, the instance of MyPixelStream is not initialized with a width, height,
or colorSpace since it learns those characteristics automatically when the PNGDecoder emits
a 'format' event.
If this event is emitted by a source stream (e.g. image decoder), which is then piped to a
PixelStream, the frame object is added to the frame queue through the addFrame method
described above.
The following methods can be implemented by subclasses to provide useful behavior.
Only _writePixels is required.
This method is called at the start of the stream, before any data has been passed to _writePixels.
You should call the provided callback when you are done.
This method is called at the start of each frame, before any data for this frame has been passed to
_writePixels. It is passed a frame metadata object, which either came from a call to addFrame or
piped from another stream (described above). You should call the provided callback when you are done.
This method is called with the actual pixel data for a frame (not necessarily all at once). You should call the provided callback when you are done. This is the only method that you MUST implement.
This method is called at the end of each frame, after all data for the frame has been passed to _writePixels.
You should call the provided callback when you are done.
This method is called at the end of the entire data stream for all frames. You should call the provided callback when you are done.
PixelStream is an abstract class, which means it doesn't do much by itself. You need to extend it
with your functionality to make it useful.
The following example converts RGB images to grayscale.
var PixelStream = require('pixel-stream');
var inherits = require('util').inherits;
function MyPixelStream() {
PixelStream.apply(this, arguments);
}
inherits(MyPixelStream, PixelStream);
MyPixelStream.prototype._writePixels = function(data, done) {
if (this.colorSpace !== 'rgb')
return done(new Error('Only supports rgb input'));
var res = new Buffer(data.length / 3);
var i = 0, j = 0;
while (i < data.length) {
res[j++] = 0.2126 * data[i++] +
0.7152 * data[i++] +
0.0722 * data[i++];
}
this.push(res);
done();
};MIT