Mathis/sharp
1
0
Fork 0
mirror of https://github.com/lovell/sharp.git synced 2025-07-09 10:30:15 +02:00
Code Issues Packages Projects Releases Wiki Activity

Big documentation clean-up

Browse Source 
Add structure via mkdocs (replaces ever-growing README)

Inline usage examples with the method they demonstrate

Add changelog
This commit is contained in:
Lovell Fuller 2015-07-15 10:17:55 +01:00
parent 022a2b1ade
commit d26f6b3b89
9 changed files with 967 additions and 815 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.npmignore
View File

@ -7,3 +7,4 @@ coverage
test test
.travis.yml .travis.yml
appveyor.yml appveyor.yml
mkdocs.yml

7
CONTRIBUTING.md Executable file → Normal file
View File

@ -34,7 +34,12 @@ To test C++ changes, you can compile the module using `npm install` and then run
## Submit a Pull Request with a new feature ## Submit a Pull Request with a new feature
Please add JavaScript [unit tests](https://github.com/lovell/sharp/tree/master/test/unit) to cover your new feature. A test coverage report for the JavaScript code is generated in the `coverage/lcov-report` directory. Please add JavaScript [unit tests](https://github.com/lovell/sharp/tree/master/test/unit) to cover your new feature.
A test coverage report for the JavaScript code is generated in the `coverage/lcov-report` directory.
Where possible, the functional tests use gradient-based perceptual hashes
based on [dHash](http://www.hackerfactor.com/blog/index.php?/archives/529-Kind-of-Like-That.html)
to compare expected vs actual images.
You deserve to add your details to the [list of contributors](https://github.com/lovell/sharp/blob/master/package.json#L5). You deserve to add your details to the [list of contributors](https://github.com/lovell/sharp/blob/master/package.json#L5).

827
README.md Executable file → Normal file
View File

@ -1,831 +1,30 @@
# sharp # sharp
* [Installation](https://github.com/lovell/sharp#installation) The typical use case for this high speed Node.js module
* [Usage examples](https://github.com/lovell/sharp#usage-examples) is to convert large images of many formats to
* [API](https://github.com/lovell/sharp#api) smaller, web-friendly JPEG, PNG and WebP images of varying dimensions.
* [Contributing](https://github.com/lovell/sharp#contributing)
* [Testing](https://github.com/lovell/sharp#testing)
* [Performance](https://github.com/lovell/sharp#performance)
* [Thanks](https://github.com/lovell/sharp#thanks)
* [Licence](https://github.com/lovell/sharp#licence)
The typical use case for this high speed Node.js module is to convert large images of many formats to smaller, web-friendly JPEG, PNG and WebP images of varying dimensions.
This module supports reading and writing JPEG, PNG and WebP images to and from Streams, Buffer objects and the filesystem.
It also supports reading images of many other formats from the filesystem via libmagick, libgraphicsmagick or [OpenSlide](http://openslide.org/) if present.
Deep Zoom image pyramids can be generated, suitable for use with "slippy map" tile viewers like
[OpenSeadragon](https://github.com/openseadragon/openseadragon) and [Leaflet](https://github.com/turban/Leaflet.Zoomify).
Colour spaces, embedded ICC profiles and alpha transparency channels are all handled correctly.
Only small regions of uncompressed image data are held in memory and processed at a time, taking full advantage of multiple CPU cores and L1/L2/L3 cache. Resizing an image is typically 4x faster than using the quickest ImageMagick and GraphicsMagick settings.
Huffman tables are optimised when generating JPEG output images without having to use separate command line tools like [jpegoptim](https://github.com/tjko/jpegoptim) and [jpegtran](http://jpegclub.org/jpegtran/). PNG filtering can be disabled, which for diagrams and line art often produces the same result as [pngcrush](http://pmt.sourceforge.net/pngcrush/).
Everything remains non-blocking thanks to _libuv_, no child processes are spawned and Promises/A+ are supported.
Anyone who has used the Node.js bindings for [GraphicsMagick](https://github.com/aheckmann/gm) will find the API similarly fluent.
This module is powered by the blazingly fast [libvips](https://github.com/jcupitt/libvips) image processing library, originally created in 1989 at Birkbeck College and currently maintained by [John Cupitt](https://github.com/jcupitt).
## Installation
npm install sharp
### Prerequisites
* Node.js v0.10+ or io.js
* [libvips](https://github.com/jcupitt/libvips) v7.40.0+ (7.42.0+ recommended)
* C++11 compatible compiler such as gcc 4.6+, clang 3.0+ or MSVC 2013
To install the most suitable version of libvips on the following Operating Systems:
* Mac OS
* Homebrew
* MacPorts
* Debian Linux
* Debian 7, 8
* Ubuntu 12.04, 14.04, 14.10, 15.04
* Mint 13, 17
* Red Hat Linux
* RHEL/Centos/Scientific 6, 7
* Fedora 21, 22
* Amazon Linux 2014.09
* OpenSuse Linux
* OpenSuse 13.1, 13.2
run the following as a user with `sudo` access:
curl -s https://raw.githubusercontent.com/lovell/sharp/master/preinstall.sh | sudo bash -
or run the following as `root`:
curl -s https://raw.githubusercontent.com/lovell/sharp/master/preinstall.sh | bash -
The [preinstall.sh](https://github.com/lovell/sharp/blob/master/preinstall.sh) script requires `curl` and `pkg-config`.
Add `--with-openslide` to enable OpenSlide support:
curl -s https://raw.githubusercontent.com/lovell/sharp/master/preinstall.sh | sudo bash -s -- --with-openslide
### Mac OS tips
Manual install via homebrew:
brew install homebrew/science/vips --with-webp --with-graphicsmagick
A missing or incorrectly configured _Xcode Command Line Tools_ installation [can lead](https://github.com/lovell/sharp/issues/80) to a `library not found for -ljpeg` error. If so, please try:
xcode-select --install
The _gettext_ dependency of _libvips_ [can lead](https://github.com/lovell/sharp/issues/9) to a `library not found for -lintl` error. If so, please try:
brew link gettext --force
### Windows
Requires x86 32-bit Node.js or io.js (use `iojs.exe` rather than `node.exe`).
The WebP format is currently unsupported.
1. Ensure the [node-gyp prerequisites](https://github.com/TooTallNate/node-gyp#installation) are met.
2. [Download](http://www.vips.ecs.soton.ac.uk/supported/current/win32/) and unzip `vips-dev.x.y.z.zip`.
3. Set the `VIPS_HOME` environment variable to the full path of the `vips-dev-x.y.z` directory.
4. Add `vips-dev-x.y.z\bin` to `PATH`.
Versions of MSVC more recent than 2013 may require the use of `npm install --arch=ia32 --msvs_version=2013`.
### Heroku
[Alessandro Tagliapietra](https://github.com/alex88) maintains an [Heroku buildpack for libvips](https://github.com/alex88/heroku-buildpack-vips) and its dependencies.
### Docker
[Marc Bachmann](https://github.com/marcbachmann) maintains a [Dockerfile for libvips](https://github.com/marcbachmann/dockerfile-libvips).
docker pull marcbachmann/libvips
### Build tools
* [gulp-responsive](https://www.npmjs.com/package/gulp-responsive)
* [gulp-sharp](https://www.npmjs.com/package/gulp-sharp)
* [grunt-sharp](https://www.npmjs.com/package/grunt-sharp)
## Usage examples
```javascript
var sharp = require('sharp');
```
```javascript
sharp('input.jpg').resize(300, 200).toFile('output.jpg', function(err) {
if (err) {
throw err;
}
// output.jpg is a 300 pixels wide and 200 pixels high image
// containing a scaled and cropped version of input.jpg
});
```
```javascript
var transformer = sharp()
.resize(300, 200)
.crop(sharp.gravity.north)
.on('error', function(err) {
console.log(err);
});
// Read image data from readableStream, resize and write image data to writableStream
readableStream.pipe(transformer).pipe(writableStream);
```
```javascript
var image = sharp(inputJpg);
image.metadata(function(err, metadata) {
image.resize(Math.floor(metadata.width / 2)).webp().toBuffer(function(err, outputBuffer, info) {
// outputBuffer contains a WebP image half the width and height of the original JPEG
});
});
```
```javascript
var pipeline = sharp()
.rotate()
.resize(null, 200)
.progressive()
.toBuffer(function(err, outputBuffer, info) {
if (err) {
throw err;
}
// outputBuffer contains 200px high progressive JPEG image data,
// auto-rotated using EXIF Orientation tag
// info.width and info.height contain the dimensions of the resized image
});
readableStream.pipe(pipeline);
```
```javascript
var pipeline = sharp().rotate();
pipeline.clone().resize(800, 600).pipe(firstWritableStream);
pipeline.clone().extract(20, 20, 100, 100).pipe(secondWritableStream);
readableStream.pipe(pipeline);
// firstWritableStream receives auto-rotated, resized readableStream
// secondWritableStream receives auto-rotated, extracted region of readableStream
```
```javascript
sharp('input.png')
.rotate(180)
.resize(300)
.flatten()
.background('#ff6600')
.overlayWith('overlay.png')
.sharpen()
.withMetadata()
.quality(90)
.webp()
.toBuffer()
.then(function(outputBuffer) {
// outputBuffer contains upside down, 300px wide, alpha channel flattened
// onto orange background, composited with `overlay.png`, sharpened,
// with metadata, 90% quality WebP image data
});
```
```javascript
http.createServer(function(request, response) {
response.writeHead(200, {'Content-Type': 'image/webp'});
sharp('input.jpg').rotate().resize(200).webp().pipe(response);
}).listen(8000);
// Create HTTP server that always returns auto-rotated 'input.jpg',
// resized to 200 pixels wide, in WebP format
```
```javascript
sharp(input)
.extract(top, left, width, height)
.toFile(output);
// Extract a region of the input image, saving in the same format.
```
```javascript
sharp(input)
.extract(topOffsetPre, leftOffsetPre, widthPre, heightPre)
.resize(width, height)
.extract(topOffsetPost, leftOffsetPost, widthPost, heightPost)
.toFile(output);
// Extract a region, resize, then extract from the resized image
```
```javascript
sharp(inputBuffer)
.resize(200, 300)
.interpolateWith(sharp.interpolator.nohalo)
.background('white')
.embed()
.toFile('output.tiff')
.then(function() {
// output.tiff is a 200 pixels wide and 300 pixels high image
// containing a nohalo scaled version, embedded on a white canvas,
// of the image data in inputBuffer
});
```
```javascript
sharp('input.gif')
.resize(200, 300)
.background({r: 0, g: 0, b: 0, a: 0})
.embed()
.toFormat(sharp.format.webp)
.toBuffer(function(err, outputBuffer) {
if (err) {
throw err;
}
// outputBuffer contains WebP image data of a 200 pixels wide and 300 pixels high
// containing a scaled version, embedded on a transparent canvas, of input.gif
});
```
```javascript
sharp(inputBuffer)
.resize(200, 200)
.max()
.toFormat('jpeg')
.toBuffer().then(function(outputBuffer) {
// outputBuffer contains JPEG image data no wider than 200 pixels and no higher
// than 200 pixels regardless of the inputBuffer image dimensions
});
```
```javascript
sharp('input.tiff').tile(256).toFile('output.dzi', function(err, info) {
// The output.dzi file is the XML format Deep Zoom definition
// The output_files directory contains 256x256 pixel tiles grouped by zoom level
});
```
```javascript
// Runtime discovery of available formats
console.dir(sharp.format);
```
## API
### Attributes
#### format
An Object containing nested boolean values
representing the available input and output formats/methods,
for example:
```javascript
{ jpeg: { id: 'jpeg',
input: { file: true, buffer: true, stream: true },
output: { file: true, buffer: true, stream: true } },
png: { id: 'png',
input: { file: true, buffer: true, stream: true },
output: { file: true, buffer: true, stream: true } },
webp: { id: 'webp',
input: { file: true, buffer: true, stream: true },
output: { file: true, buffer: true, stream: true } },
tiff: { id: 'tiff',
input: { file: true, buffer: true, stream: true },
output: { file: true, buffer: false, stream: false } },
magick: { id: 'magick',
input: { file: true, buffer: true, stream: true },
output: { file: false, buffer: false, stream: false } },
raw: { id: 'raw',
input: { file: false, buffer: false, stream: false },
output: { file: false, buffer: true, stream: true } } }
```
#### queue
An EventEmitter that emits a `change` event when a task is either:
* queued, waiting for _libuv_ to provide a worker thread
* complete
```javascript
sharp.queue.on('change', function(queueLength) {
console.log('Queue contains ' + queueLength + ' task(s)');
});
```
### Input methods
#### sharp([input])
Constructor to which further methods are chained. `input`, if present, can be one of:
* Buffer containing JPEG, PNG, WebP, GIF* or TIFF image data, or
* String containing the filename of an image, with most major formats supported.
The object returned implements the [stream.Duplex](http://nodejs.org/api/stream.html#stream_class_stream_duplex) class.
JPEG, PNG, WebP, GIF* or TIFF format image data can be streamed into the object when `input` is not provided.
JPEG, PNG or WebP format image data can be streamed out from this object.
\* libvips 8.0.0+ is required for Buffer/Stream input of GIF and other `magick` formats.
#### metadata([callback])
Fast access to image metadata without decoding any compressed image data.
`callback`, if present, gets the arguments `(err, metadata)` where `metadata` has the attributes:
* `format`: Name of decoder to be used to decompress image data e.g. `jpeg`, `png`, `webp` (for file-based input additionally `tiff`, `magick` and `openslide`)
* `width`: Number of pixels wide
* `height`: Number of pixels high
* `space`: Name of colour space interpretation e.g. `srgb`, `rgb`, `scrgb`, `cmyk`, `lab`, `xyz`, `b-w` [...](https://github.com/jcupitt/libvips/blob/master/libvips/iofuncs/enumtypes.c#L522)
* `channels`: Number of bands e.g. `3` for sRGB, `4` for CMYK
* `hasProfile`: Boolean indicating the presence of an embedded ICC profile
* `hasAlpha`: Boolean indicating the presence of an alpha transparency channel
* `orientation`: Number value of the EXIF Orientation header, if present
* `exif`: Buffer containing raw EXIF data, if present
* `icc`: Buffer containing raw [ICC](https://www.npmjs.com/package/icc) profile data, if present
A Promises/A+ promise is returned when `callback` is not provided.
#### sequentialRead()
An advanced setting that switches the libvips access method to `VIPS_ACCESS_SEQUENTIAL`. This will reduce memory usage and can improve performance on some systems.
#### limitInputPixels(pixels)
Do not process input images where the number of pixels (width * height) exceeds this limit.
`pixels` is the integral Number of pixels, with a value between 1 and the default 268402689 (0x3FFF * 0x3FFF).
#### clone()
Takes a "snapshot" of the instance, returning a new instance.
Cloned instances inherit the input of their parent instance.
This allows multiple output Streams
and therefore multiple processing pipelines
to share a single input Stream.
### Image transformation options
#### resize(width, [height])
Scale output to `width` x `height`. By default, the resized image is cropped to the exact size specified.
`width` is the integral Number of pixels wide the resultant image should be, between 1 and 16383 (0x3FFF). Use `null` or `undefined` to auto-scale the width to match the height.
`height` is the integral Number of pixels high the resultant image should be, between 1 and 16383. Use `null` or `undefined` to auto-scale the height to match the width.
#### extract(top, left, width, height)
Extract a region of the image. Can be used with or without a `resize` operation.
`top` and `left` are the offset, in pixels, from the top-left corner.
`width` and `height` are the dimensions of the extracted image.
Use `extract` before `resize` for pre-resize extraction. Use `extract` after `resize` for post-resize extraction. Use `extract` before and after for both.
#### crop([gravity])
Crop the resized image to the exact size specified, the default behaviour.
`gravity`, if present, is an attribute of the `sharp.gravity` Object e.g. `sharp.gravity.north`.
Possible values are `north`, `east`, `south`, `west`, `center` and `centre`. The default gravity is `center`/`centre`.
#### max()
Preserving aspect ratio,
resize the image to be as large as possible
while ensuring its dimensions are less than or equal to
the `width` and `height` specified.
Both `width` and `height` must be provided via `resize` otherwise the behaviour will default to `crop`.
#### min()
Preserving aspect ratio,
resize the image to be as small as possible
while ensuring its dimensions are greater than or equal to
the `width` and `height` specified.
Both `width` and `height` must be provided via `resize` otherwise the behaviour will default to `crop`.
#### ignoreAspectRatio()
Ignoring the aspect ratio of the input, stretch the image to the exact `width` and/or `height` provided via `resize`.
#### background(rgba)
Set the background for the `embed` and `flatten` operations.
`rgba` is parsed by the [color](https://www.npmjs.org/package/color) module to extract values for red, green, blue and alpha.
The alpha value is a float between `0` (transparent) and `1` (opaque).
The default background is `{r: 0, g: 0, b: 0, a: 1}`, black without transparency.
#### embed()
Preserving aspect ratio, resize the image to the maximum `width` or `height` specified then embed on a background of the exact `width` and `height` specified.
If the background contains an alpha value then WebP and PNG format output images will contain an alpha channel, even when the input image does not.
#### flatten()
Merge alpha transparency channel, if any, with `background`.
#### rotate([angle])
Rotate the output image by either an explicit angle or auto-orient based on the EXIF `Orientation` tag.
`angle`, if present, is a Number with a value of `0`, `90`, `180` or `270`.
Use this method without `angle` to determine the angle from EXIF data. Mirroring is supported and may infer the use of a `flip` operation.
Method order is important when both rotating and extracting regions, for example `rotate(x).extract(y)` will produce a different result to `extract(y).rotate(x)`.
The use of `rotate` implies the removal of the EXIF `Orientation` tag, if any.
#### flip()
Flip the image about the vertical Y axis. This always occurs after rotation, if any.
The use of `flip` implies the removal of the EXIF `Orientation` tag, if any.
#### flop()
Flop the image about the horizontal X axis. This always occurs after rotation, if any.
The use of `flop` implies the removal of the EXIF `Orientation` tag, if any.
#### withoutEnlargement()
Do not enlarge the output image if the input image width *or* height are already less than the required dimensions.
This is equivalent to GraphicsMagick's `>` geometry option: "change the dimensions of the image only if its width or height exceeds the geometry specification".
#### blur([sigma])
When used without parameters, performs a fast, mild blur of the output image. This typically reduces performance by 10%.
When a `sigma` is provided, performs a slower, more accurate Gaussian blur. This typically reduces performance by 25%.
* `sigma`, if present, is a Number between 0.3 and 1000 representing the approximate blur radius in pixels.
#### sharpen([radius], [flat], [jagged])
When used without parameters, performs a fast, mild sharpen of the output image. This typically reduces performance by 10%.
When a `radius` is provided, performs a slower, more accurate sharpen of the L channel in the LAB colour space. Separate control over the level of sharpening in "flat" and "jagged" areas is available. This typically reduces performance by 50%.
* `radius`, if present, is an integral Number representing the sharpen mask radius in pixels.
* `flat`, if present, is a Number representing the level of sharpening to apply to "flat" areas, defaulting to a value of 1.0.
* `jagged`, if present, is a Number representing the level of sharpening to apply to "jagged" areas, defaulting to a value of 2.0.
#### interpolateWith(interpolator)
Use the given interpolator for image resizing, where `interpolator` is an attribute of the `sharp.interpolator` Object e.g. `sharp.interpolator.bicubic`.
Possible interpolators, in order of performance, are:
* `nearest`: Use [nearest neighbour interpolation](http://en.wikipedia.org/wiki/Nearest-neighbor_interpolation), suitable for image enlargement only.
* `bilinear`: Use [bilinear interpolation](http://en.wikipedia.org/wiki/Bilinear_interpolation), the default and fastest image reduction interpolation.
* `bicubic`: Use [bicubic interpolation](http://en.wikipedia.org/wiki/Bicubic_interpolation), which typically reduces performance by 5%.
* `vertexSplitQuadraticBasisSpline`: Use [VSQBS interpolation](https://github.com/jcupitt/libvips/blob/master/libvips/resample/vsqbs.cpp#L48), which prevents "staircasing" and typically reduces performance by 5%.
* `locallyBoundedBicubic`: Use [LBB interpolation](https://github.com/jcupitt/libvips/blob/master/libvips/resample/lbb.cpp#L100), which prevents some "[acutance](http://en.wikipedia.org/wiki/Acutance)" and typically reduces performance by a factor of 2.
* `nohalo`: Use [Nohalo interpolation](http://eprints.soton.ac.uk/268086/), which prevents acutance and typically reduces performance by a factor of 3.
#### gamma([gamma])
Apply a gamma correction by reducing the encoding (darken) pre-resize at a factor of `1/gamma` then increasing the encoding (brighten) post-resize at a factor of `gamma`.
`gamma`, if present, is a Number betweem 1 and 3. The default value is `2.2`, a suitable approximation for sRGB images.
This can improve the perceived brightness of a resized image in non-linear colour spaces.
JPEG input images will not take advantage of the shrink-on-load performance optimisation when applying a gamma correction.
#### grayscale() / greyscale()
Convert to 8-bit greyscale; 256 shades of grey.
This is a linear operation. If the input image is in a non-linear colour space such as sRGB, use `gamma()` with `greyscale()` for the best results.
The output image will still be web-friendly sRGB and contain three (identical) channels.
#### normalize() / normalise()
Enhance output image contrast by stretching its luminance to cover the full dynamic range. This typically reduces performance by 30%.
#### overlayWith(filename)
_Experimental_
Alpha composite `filename` over the processed (resized, extracted) image. The dimensions of the two images must match.
* `filename` is a String containing the filename of an image with an alpha channel.
### Output options
#### jpeg()
Use JPEG format for the output image.
#### png()
Use PNG format for the output image.
#### webp()
Use WebP format for the output image.
#### raw()
_Requires libvips 7.42.0+_
Provide raw, uncompressed uint8 (unsigned char) image data for Buffer and Stream based output.
The number of channels depends on the input image and selected options.
* 1 channel for images converted to `greyscale()`, with each byte representing one pixel.
* 3 channels for colour images without alpha transparency, with bytes ordered \[red, green, blue, red, green, blue, etc.\]).
* 4 channels for colour images with alpha transparency, with bytes ordered \[red, green, blue, alpha, red, green, blue, alpha, etc.\].
#### toFormat(format)
Convenience method for the above output format methods, where `format` is either:
* an attribute of the `sharp.format` Object e.g. `sharp.format.jpeg`, or
* a String containing `jpeg`, `png`, `webp` or `raw`.
#### quality(quality)
The output quality to use for lossy JPEG, WebP and TIFF output formats. The default quality is `80`.
`quality` is a Number between 1 and 100.
#### progressive()
Use progressive (interlace) scan for JPEG and PNG output. This typically reduces compression performance by 30% but results in an image that can be rendered sooner when decompressed.
#### withMetadata([metadata])
Include all metadata (EXIF, XMP, IPTC) from the input image in the output image.
This will also convert to and add the latest web-friendly v2 sRGB ICC profile.
The optional `metadata` parameter, if present, is an Object with the attributes to update.
New attributes cannot be inserted, only existing attributes updated.
* `orientation` is an integral Number between 0 and 7, used to update the value of the EXIF `Orientation` tag.
This has no effect if the input image does not have an EXIF `Orientation` tag.
The default behaviour, when `withMetadata` is not used, is to strip all metadata and convert to the device-independent sRGB colour space.
#### tile([size], [overlap])
The size and overlap, in pixels, of square Deep Zoom image pyramid tiles.
* `size` is an integral Number between 1 and 8192. The default value is 256 pixels.
* `overlap` is an integral Number between 0 and 8192. The default value is 0 pixels.
#### withoutChromaSubsampling()
Disable the use of [chroma subsampling](http://en.wikipedia.org/wiki/Chroma_subsampling) with JPEG output (4:4:4).
This can improve colour representation at higher quality settings (90+),
but usually increases output file size and typically reduces performance by 25%.
The default behaviour is to use chroma subsampling (4:2:0).
#### trellisQuantisation() / trellisQuantization()
_Requires libvips 8.0.0+ compiled against mozjpeg 3.0+_
An advanced setting to apply the use of
[trellis quantisation](http://en.wikipedia.org/wiki/Trellis_quantization) with JPEG output.
Reduces file size and slightly increases relative quality at the cost of increased compression time.
#### overshootDeringing()
_Requires libvips 8.0.0+ compiled against mozjpeg 3.0+_
An advanced setting to reduce the effects of
[ringing](http://en.wikipedia.org/wiki/Ringing_%28signal%29) in JPEG output,
in particular where black text appears on a white background (or vice versa).
#### optimiseScans() / optimizeScans()
_Requires libvips 8.0.0+ compiled against mozjpeg 3.0+_
An advanced setting for progressive (interlace) JPEG output.
Calculates which spectrum of DCT coefficients uses the fewest bits.
Usually reduces file size at the cost of increased compression time.
#### compressionLevel(compressionLevel)
An advanced setting for the _zlib_ compression level of the lossless PNG output format. The default level is `6`.
`compressionLevel` is a Number between 0 and 9.
#### withoutAdaptiveFiltering()
_Requires libvips 7.42.0+_
An advanced setting to disable adaptive row filtering for the lossless PNG output format.
### Output methods
#### toFile(filename, [callback])
`filename` is a String containing the filename to write the image data to. The format is inferred from the extension, with JPEG, PNG, WebP, TIFF and DZI supported.
`callback`, if present, is called with two arguments `(err, info)` where:
* `err` contains an error message, if any.
* `info` contains the output image `format`, `size` (bytes), `width` and `height`.
A Promises/A+ promise is returned when `callback` is not provided.
#### toBuffer([callback])
Write image data to a Buffer, the format of which will match the input image by default. JPEG, PNG and WebP are supported.
`callback`, if present, gets three arguments `(err, buffer, info)` where:
* `err` is an error message, if any.
* `buffer` is the output image data.
* `info` contains the output image `format`, `size` (bytes), `width` and `height`.
A Promises/A+ promise is returned when `callback` is not provided.
### Utility methods
#### sharp.cache([memory], [items])
If `memory` or `items` are provided, set the limits of _libvips'_ operation cache.
* `memory` is the maximum memory in MB to use for this cache, with a default value of 100
* `items` is the maximum number of operations to cache, with a default value of 500
This method always returns cache statistics, useful for determining how much working memory is required for a particular task.
```javascript
var stats = sharp.cache(); // { current: 75, high: 99, memory: 100, items: 500 }
sharp.cache(200); // { current: 75, high: 99, memory: 200, items: 500 }
sharp.cache(50, 200); // { current: 49, high: 99, memory: 50, items: 200}
```
#### sharp.concurrency([threads])
`threads`, if provided, is the Number of threads _libvips'_ should create for processing each image. The default value is the number of CPU cores. A value of `0` will reset to this default.
This method always returns the current concurrency.
```javascript
var threads = sharp.concurrency(); // 4
sharp.concurrency(2); // 2
sharp.concurrency(0); // 4
```
The maximum number of images that can be processed in parallel is limited by libuv's `UV_THREADPOOL_SIZE` environment variable.
#### sharp.counters()
Provides access to internal task counters.
* `queue` is the number of tasks this module has queued waiting for _libuv_ to provide a worker thread from its pool.
* `process` is the number of resize tasks currently being processed.
```javascript
var counters = sharp.counters(); // { queue: 2, process: 4 }
```
## Contributing
A [guide for contributors](https://github.com/lovell/sharp/blob/master/CONTRIBUTING.md) covers reporting bugs, requesting features and submitting code changes.
## Testing
### Functional tests
Where possible, the functional tests use gradient-based perceptual hashes
based on [dHash](http://www.hackerfactor.com/blog/index.php?/archives/529-Kind-of-Like-That.html)
to compare expected vs actual images.
#### Coverage
[![Test Coverage](https://coveralls.io/repos/lovell/sharp/badge.png?branch=master)](https://coveralls.io/r/lovell/sharp?branch=master) [![Test Coverage](https://coveralls.io/repos/lovell/sharp/badge.png?branch=master)](https://coveralls.io/r/lovell/sharp?branch=master)
#### Ubuntu 12.04 ### Documentation
[![Ubuntu 12.04 Build Status](https://travis-ci.org/lovell/sharp.png?branch=master)](https://travis-ci.org/lovell/sharp) Visit [sharp.dimens.io](http://sharp.dimens.io/) for
complete installation instructions, API documentation,
benchmark tests and a changelog.
#### Centos 6.5 ### Contributing
[![Centos 6.5 Build Status](https://snap-ci.com/lovell/sharp/branch/master/build_image)](https://snap-ci.com/lovell/sharp/branch/master) A [guide for contributors](https://github.com/lovell/sharp/blob/master/CONTRIBUTING.md)
covers reporting bugs, requesting features and submitting code changes.
#### Windows Server 2012 ### Licence
[![Windows Server 2012 Build Status](https://ci.appveyor.com/api/projects/status/pgtul704nkhhg6sg)](https://ci.appveyor.com/project/lovell/sharp)
#### OS X 10.9.5
[![OS X 10.9.5 Build Status](https://travis-ci.org/lovell/sharp-osx-ci.png?branch=master)](https://travis-ci.org/lovell/sharp-osx-ci)
### Benchmark tests
```
cd sharp/test/bench
npm install
npm test
```
Requires both _ImageMagick_ and _GraphicsMagick_:
```
brew install imagemagick
brew install graphicsmagick
```
```
sudo apt-get install -qq imagemagick graphicsmagick libmagick++-dev
```
```
sudo yum install ImageMagick
sudo yum install -y http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-8.noarch.rpm
sudo yum install -y --enablerepo=epel GraphicsMagick
```
## Performance
### Test environment
* AWS EC2 [c4.xlarge](http://aws.amazon.com/ec2/instance-types/#c4)
* Ubuntu 15.04
* Node.js 0.12.7
* libvips 8.0.2
* liborc 0.4.22
### The contenders
* [lwip](https://www.npmjs.com/package/lwip) 0.0.7 - Wrapper around CImg, compiles dependencies from source
* [imagemagick-native](https://www.npmjs.com/package/imagemagick-native) git@45d4e2e - Supports Buffers only
* [imagemagick](https://www.npmjs.com/package/imagemagick) 0.1.3 - Supports filesystem only and "has been unmaintained for a long time".
* [gm](https://www.npmjs.com/package/gm) 1.18.1 - Fully featured wrapper around GraphicsMagick.
* sharp 0.11.0 - Caching within libvips disabled to ensure a fair comparison.
### The task
Decompress a 2725x2225 JPEG image, resize to 720x480 using bilinear interpolation, then compress to JPEG.
### Results
| Module | Input | Output | Ops/sec | Speed-up |
| :----------------- | :----- | :----- | ------: | -------: |
| lwip | file | file | 1.75 | 1 |
| lwip | buffer | buffer | 2.21 | 1.3 |
| imagemagick-native | buffer | buffer | 7.13 | 4.1 |
| gm | buffer | buffer | 7.27 | 4.2 |
| gm | file | file | 7.33 | 4.2 |
| imagemagick | file | file | 10.04 | 5.7 |
| sharp | stream | stream | 23.12 | 13.2 |
| sharp | file | file | 24.43 | 14.0 |
| sharp | file | buffer | 24.55 | 14.0 |
| sharp | buffer | file | 24.86 | 14.2 |
| sharp | buffer | buffer | 24.92 | 14.2 |
You can expect greater performance with caching enabled (default) and using 8+ core machines.
## Thanks
This module would never have been possible without the help and code contributions of the following people:
* [John Cupitt](https://github.com/jcupitt)
* [Pierre Inglebert](https://github.com/pierreinglebert)
* [Jonathan Ong](https://github.com/jonathanong)
* [Chanon Sajjamanochai](https://github.com/chanon)
* [Juliano Julio](https://github.com/julianojulio)
* [Daniel Gasienica](https://github.com/gasi)
* [Julian Walker](https://github.com/julianwa)
* [Amit Pitaru](https://github.com/apitaru)
* [Brandon Aaron](https://github.com/brandonaaron)
* [Andreas Lind](https://github.com/papandreou)
* [Maurus Cuelenaere](https://github.com/mcuelenaere)
* [Linus Unnebäck](https://github.com/LinusU)
* [Victor Mateevitsi](https://github.com/mvictoras)
* [Alaric Holloway](https://github.com/skedastik)
* [Bernhard K. Weisshuhn](https://github.com/bkw)
Thank you!
## Licence
Copyright 2013, 2014, 2015 Lovell Fuller and contributors. Copyright 2013, 2014, 2015 Lovell Fuller and contributors.
Licensed under the Apache License, Version 2.0 (the "License"); Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License. you may not use this file except in compliance with the License.
You may obtain a copy of the License at [http://www.apache.org/licenses/LICENSE-2.0](http://www.apache.org/licenses/LICENSE-2.0.html) You may obtain a copy of the License at
[http://www.apache.org/licenses/LICENSE-2.0](http://www.apache.org/licenses/LICENSE-2.0.html)
Unless required by applicable law or agreed to in writing, software Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS, distributed under the License is distributed on an "AS IS" BASIS,

597
docs/api.md Normal file
View File

@ -0,0 +1,597 @@
# API
```javascript
var sharp = require('sharp');
```
### Input
#### sharp([input])
Constructor to which further methods are chained. `input`, if present, can be one of:
* Buffer containing JPEG, PNG, WebP, GIF* or TIFF image data, or
* String containing the filename of an image, with most major formats supported.
The object returned implements the
[stream.Duplex](http://nodejs.org/api/stream.html#stream_class_stream_duplex) class.
JPEG, PNG, WebP, GIF* or TIFF format image data
can be streamed into the object when `input` is not provided.
JPEG, PNG or WebP format image data can be streamed out from this object.
\* libvips 8.0.0+ is required for Buffer/Stream input of GIF and other `magick` formats.
```javascript
sharp('input.jpg')
.resize(300, 200)
.toFile('output.jpg', function(err) {
// output.jpg is a 300 pixels wide and 200 pixels high image
// containing a scaled and cropped version of input.jpg
});
```
#### metadata([callback])
Fast access to image metadata without decoding any compressed image data.
`callback`, if present, gets the arguments `(err, metadata)` where `metadata` has the attributes:
* `format`: Name of decoder to be used to decompress image data e.g. `jpeg`, `png`, `webp` (for file-based input additionally `tiff`, `magick` and `openslide`)
* `width`: Number of pixels wide
* `height`: Number of pixels high
* `space`: Name of colour space interpretation e.g. `srgb`, `rgb`, `scrgb`, `cmyk`, `lab`, `xyz`, `b-w` [...](https://github.com/jcupitt/libvips/blob/master/libvips/iofuncs/enumtypes.c#L522)
* `channels`: Number of bands e.g. `3` for sRGB, `4` for CMYK
* `hasProfile`: Boolean indicating the presence of an embedded ICC profile
* `hasAlpha`: Boolean indicating the presence of an alpha transparency channel
* `orientation`: Number value of the EXIF Orientation header, if present
* `exif`: Buffer containing raw EXIF data, if present
* `icc`: Buffer containing raw [ICC](https://www.npmjs.com/package/icc) profile data, if present
A Promises/A+ promise is returned when `callback` is not provided.
```javascript
var image = sharp(inputJpg);
image
.metadata()
.then(function(metadata) {
return image
.resize(Math.round(metadata.width / 2))
.webp()
.toBuffer();
})
.then(function(data) {
// data contains a WebP image half the width and height of the original JPEG
});
```
#### clone()
Takes a "snapshot" of the instance, returning a new instance.
Cloned instances inherit the input of their parent instance.
This allows multiple output Streams
and therefore multiple processing pipelines
to share a single input Stream.
```javascript
var pipeline = sharp().rotate();
pipeline.clone().resize(800, 600).pipe(firstWritableStream);
pipeline.clone().extract(20, 20, 100, 100).pipe(secondWritableStream);
readableStream.pipe(pipeline);
// firstWritableStream receives auto-rotated, resized readableStream
// secondWritableStream receives auto-rotated, extracted region of readableStream
```
#### sequentialRead()
An advanced setting that switches the libvips access method to `VIPS_ACCESS_SEQUENTIAL`.
This will reduce memory usage and can improve performance on some systems.
#### limitInputPixels(pixels)
Do not process input images where the number of pixels (width * height) exceeds this limit.
`pixels` is the integral Number of pixels, with a value between 1 and the default 268402689 (0x3FFF * 0x3FFF).
### Resizing
#### resize([width], [height])
Scale output to `width` x `height`. By default, the resized image is cropped to the exact size specified.
`width` is the integral Number of pixels wide the resultant image should be, between 1 and 16383 (0x3FFF). Use `null` or `undefined` to auto-scale the width to match the height.
`height` is the integral Number of pixels high the resultant image should be, between 1 and 16383. Use `null` or `undefined` to auto-scale the height to match the width.
#### crop([gravity])
Crop the resized image to the exact size specified, the default behaviour.
`gravity`, if present, is an attribute of the `sharp.gravity` Object e.g. `sharp.gravity.north`.
Possible values are `north`, `east`, `south`, `west`, `center` and `centre`. The default gravity is `center`/`centre`.
```javascript
var transformer = sharp()
.resize(300, 200)
.crop(sharp.gravity.north)
.on('error', function(err) {
console.log(err);
});
// Read image data from readableStream, resize and write image data to writableStream
readableStream.pipe(transformer).pipe(writableStream);
```
#### embed()
Preserving aspect ratio, resize the image to the
maximum `width` or `height` specified
then embed on a background of the exact
`width` and `height` specified.
If the background contains an alpha value
then WebP and PNG format output images will
contain an alpha channel,
even when the input image does not.
```javascript
sharp('input.gif')
.resize(200, 300)
.background({r: 0, g: 0, b: 0, a: 0})
.embed()
.toFormat(sharp.format.webp)
.toBuffer(function(err, outputBuffer) {
if (err) {
throw err;
}
// outputBuffer contains WebP image data of a 200 pixels wide and 300 pixels high
// containing a scaled version, embedded on a transparent canvas, of input.gif
});
```
#### max()
Preserving aspect ratio,
resize the image to be as large as possible
while ensuring its dimensions are less than or equal to
the `width` and `height` specified.
Both `width` and `height` must be provided via
`resize` otherwise the behaviour will default to `crop`.
```javascript
sharp(inputBuffer)
.resize(200, 200)
.max()
.toFormat('jpeg')
.toBuffer()
.then(function(outputBuffer) {
// outputBuffer contains JPEG image data no wider than 200 pixels and no higher
// than 200 pixels regardless of the inputBuffer image dimensions
});
```
#### min()
Preserving aspect ratio,
resize the image to be as small as possible
while ensuring its dimensions are greater than or equal to
the `width` and `height` specified.
Both `width` and `height` must be provided via `resize` otherwise the behaviour will default to `crop`.
#### withoutEnlargement()
Do not enlarge the output image
if the input image width *or* height
are already less than the required dimensions.
This is equivalent to GraphicsMagick's `>` geometry option:
"*change the dimensions of the image only
if its width or height exceeds the geometry specification*".
#### ignoreAspectRatio()
Ignoring the aspect ratio of the input, stretch the image to the exact `width` and/or `height` provided via `resize`.
#### interpolateWith(interpolator)
Use the given interpolator for image resizing, where `interpolator` is an attribute of the `sharp.interpolator` Object e.g. `sharp.interpolator.bicubic`.
Possible interpolators, in order of performance, are:
* `nearest`: Use [nearest neighbour interpolation](http://en.wikipedia.org/wiki/Nearest-neighbor_interpolation), suitable for image enlargement only.
* `bilinear`: Use [bilinear interpolation](http://en.wikipedia.org/wiki/Bilinear_interpolation), the default and fastest image reduction interpolation.
* `bicubic`: Use [bicubic interpolation](http://en.wikipedia.org/wiki/Bicubic_interpolation), which typically reduces performance by 5%.
* `vertexSplitQuadraticBasisSpline`: Use [VSQBS interpolation](https://github.com/jcupitt/libvips/blob/master/libvips/resample/vsqbs.cpp#L48), which prevents "staircasing" and typically reduces performance by 5%.
* `locallyBoundedBicubic`: Use [LBB interpolation](https://github.com/jcupitt/libvips/blob/master/libvips/resample/lbb.cpp#L100), which prevents some "[acutance](http://en.wikipedia.org/wiki/Acutance)" and typically reduces performance by a factor of 2.
* `nohalo`: Use [Nohalo interpolation](http://eprints.soton.ac.uk/268086/), which prevents acutance and typically reduces performance by a factor of 3.
```javascript
sharp(inputBuffer)
.resize(200, 300)
.interpolateWith(sharp.interpolator.nohalo)
.background('white')
.embed()
.toFile('output.tiff')
.then(function() {
// output.tiff is a 200 pixels wide and 300 pixels high image
// containing a nohalo scaled version, embedded on a white canvas,
// of the image data in inputBuffer
});
```
### Operations
#### extract(top, left, width, height)
Extract a region of the image. Can be used with or without a `resize` operation.
`top` and `left` are the offset, in pixels, from the top-left corner.
`width` and `height` are the dimensions of the extracted image.
Use `extract` before `resize` for pre-resize extraction. Use `extract` after `resize` for post-resize extraction. Use `extract` before and after for both.
```javascript
sharp(input)
.extract(top, left, width, height)
.toFile(output, function(err) {
// Extract a region of the input image, saving in the same format.
});
```
```javascript
sharp(input)
.extract(topOffsetPre, leftOffsetPre, widthPre, heightPre)
.resize(width, height)
.extract(topOffsetPost, leftOffsetPost, widthPost, heightPost)
.toFile(output, function(err) {
// Extract a region, resize, then extract from the resized image
});
```
#### background(rgba)
Set the background for the `embed` and `flatten` operations.
`rgba` is parsed by the [color](https://www.npmjs.org/package/color) module to extract values for red, green, blue and alpha.
The alpha value is a float between `0` (transparent) and `1` (opaque).
The default background is `{r: 0, g: 0, b: 0, a: 1}`, black without transparency.
#### flatten()
Merge alpha transparency channel, if any, with `background`.
#### rotate([angle])
Rotate the output image by either an explicit angle or auto-orient based on the EXIF `Orientation` tag.
`angle`, if present, is a Number with a value of `0`, `90`, `180` or `270`.
Use this method without `angle` to determine the angle from EXIF data. Mirroring is supported and may infer the use of a `flip` operation.
Method order is important when both rotating and extracting regions, for example `rotate(x).extract(y)` will produce a different result to `extract(y).rotate(x)`.
The use of `rotate` implies the removal of the EXIF `Orientation` tag, if any.
```javascript
var pipeline = sharp()
.rotate()
.resize(null, 200)
.progressive()
.toBuffer(function(err, outputBuffer, info) {
if (err) {
throw err;
}
// outputBuffer contains 200px high progressive JPEG image data,
// auto-rotated using EXIF Orientation tag
// info.width and info.height contain the dimensions of the resized image
});
readableStream.pipe(pipeline);
```
#### flip()
Flip the image about the vertical Y axis. This always occurs after rotation, if any.
The use of `flip` implies the removal of the EXIF `Orientation` tag, if any.
#### flop()
Flop the image about the horizontal X axis. This always occurs after rotation, if any.
The use of `flop` implies the removal of the EXIF `Orientation` tag, if any.
#### blur([sigma])
When used without parameters, performs a fast, mild blur of the output image. This typically reduces performance by 10%.
When a `sigma` is provided, performs a slower, more accurate Gaussian blur. This typically reduces performance by 25%.
* `sigma`, if present, is a Number between 0.3 and 1000 representing the approximate blur radius in pixels.
#### sharpen([radius], [flat], [jagged])
When used without parameters, performs a fast, mild sharpen of the output image. This typically reduces performance by 10%.
When a `radius` is provided, performs a slower, more accurate sharpen of the L channel in the LAB colour space. Separate control over the level of sharpening in "flat" and "jagged" areas is available. This typically reduces performance by 50%.
* `radius`, if present, is an integral Number representing the sharpen mask radius in pixels.
* `flat`, if present, is a Number representing the level of sharpening to apply to "flat" areas, defaulting to a value of 1.0.
* `jagged`, if present, is a Number representing the level of sharpening to apply to "jagged" areas, defaulting to a value of 2.0.
#### gamma([gamma])
Apply a gamma correction by reducing the encoding (darken) pre-resize at a factor of `1/gamma` then increasing the encoding (brighten) post-resize at a factor of `gamma`.
`gamma`, if present, is a Number betweem 1 and 3. The default value is `2.2`, a suitable approximation for sRGB images.
This can improve the perceived brightness of a resized image in non-linear colour spaces.
JPEG input images will not take advantage of the shrink-on-load performance optimisation when applying a gamma correction.
#### grayscale() / greyscale()
Convert to 8-bit greyscale; 256 shades of grey.
This is a linear operation. If the input image is in a non-linear colour space such as sRGB, use `gamma()` with `greyscale()` for the best results.
The output image will still be web-friendly sRGB and contain three (identical) channels.
#### normalize() / normalise()
Enhance output image contrast by stretching its luminance to cover the full dynamic range. This typically reduces performance by 30%.
#### overlayWith(filename)
_Experimental_
Alpha composite `filename` over the processed (resized, extracted) image. The dimensions of the two images must match.
* `filename` is a String containing the filename of an image with an alpha channel.
```javascript
sharp('input.png')
.rotate(180)
.resize(300)
.flatten()
.background('#ff6600')
.overlayWith('overlay.png')
.sharpen()
.withMetadata()
.quality(90)
.webp()
.toBuffer()
.then(function(outputBuffer) {
// outputBuffer contains upside down, 300px wide, alpha channel flattened
// onto orange background, composited with overlay.png, sharpened,
// with metadata, 90% quality WebP image data. Phew!
});
```
### Output
#### toFile(filename, [callback])
`filename` is a String containing the filename to write the image data to. The format is inferred from the extension, with JPEG, PNG, WebP, TIFF and DZI supported.
`callback`, if present, is called with two arguments `(err, info)` where:
* `err` contains an error message, if any.
* `info` contains the output image `format`, `size` (bytes), `width` and `height`.
A Promises/A+ promise is returned when `callback` is not provided.
#### toBuffer([callback])
Write image data to a Buffer, the format of which will match the input image by default. JPEG, PNG and WebP are supported.
`callback`, if present, gets three arguments `(err, buffer, info)` where:
* `err` is an error message, if any.
* `buffer` is the output image data.
* `info` contains the output image `format`, `size` (bytes), `width` and `height`.
A Promises/A+ promise is returned when `callback` is not provided.
#### jpeg()
Use JPEG format for the output image.
#### png()
Use PNG format for the output image.
#### webp()
Use WebP format for the output image.
#### raw()
_Requires libvips 7.42.0+_
Provide raw, uncompressed uint8 (unsigned char) image data for Buffer and Stream based output.
The number of channels depends on the input image and selected options.
* 1 channel for images converted to `greyscale()`, with each byte representing one pixel.
* 3 channels for colour images without alpha transparency, with bytes ordered \[red, green, blue, red, green, blue, etc.\]).
* 4 channels for colour images with alpha transparency, with bytes ordered \[red, green, blue, alpha, red, green, blue, alpha, etc.\].
#### toFormat(format)
Convenience method for the above output format methods, where `format` is either:
* an attribute of the `sharp.format` Object e.g. `sharp.format.jpeg`, or
* a String containing `jpeg`, `png`, `webp` or `raw`.
#### quality(quality)
The output quality to use for lossy JPEG, WebP and TIFF output formats. The default quality is `80`.
`quality` is a Number between 1 and 100.
#### progressive()
Use progressive (interlace) scan for JPEG and PNG output. This typically reduces compression performance by 30% but results in an image that can be rendered sooner when decompressed.
#### withMetadata([metadata])
Include all metadata (EXIF, XMP, IPTC) from the input image in the output image.
This will also convert to and add the latest web-friendly v2 sRGB ICC profile.
The optional `metadata` parameter, if present, is an Object with the attributes to update.
New attributes cannot be inserted, only existing attributes updated.
* `orientation` is an integral Number between 0 and 7, used to update the value of the EXIF `Orientation` tag.
This has no effect if the input image does not have an EXIF `Orientation` tag.
The default behaviour, when `withMetadata` is not used, is to strip all metadata and convert to the device-independent sRGB colour space.
#### tile([size], [overlap])
The size and overlap, in pixels, of square Deep Zoom image pyramid tiles.
* `size` is an integral Number between 1 and 8192. The default value is 256 pixels.
* `overlap` is an integral Number between 0 and 8192. The default value is 0 pixels.
```javascript
sharp('input.tiff').tile(256).toFile('output.dzi', function(err, info) {
// The output.dzi file is the XML format Deep Zoom definition
// The output_files directory contains 256x256 pixel tiles grouped by zoom level
});
```
#### withoutChromaSubsampling()
Disable the use of [chroma subsampling](http://en.wikipedia.org/wiki/Chroma_subsampling) with JPEG output (4:4:4).
This can improve colour representation at higher quality settings (90+),
but usually increases output file size and typically reduces performance by 25%.
The default behaviour is to use chroma subsampling (4:2:0).
#### compressionLevel(compressionLevel)
An advanced setting for the _zlib_ compression level of the lossless PNG output format. The default level is `6`.
`compressionLevel` is a Number between 0 and 9.
#### withoutAdaptiveFiltering()
_Requires libvips 7.42.0+_
An advanced setting to disable adaptive row filtering for the lossless PNG output format.
#### trellisQuantisation() / trellisQuantization()
_Requires libvips 8.0.0+ compiled against mozjpeg 3.0+_
An advanced setting to apply the use of
[trellis quantisation](http://en.wikipedia.org/wiki/Trellis_quantization) with JPEG output.
Reduces file size and slightly increases relative quality at the cost of increased compression time.
#### overshootDeringing()
_Requires libvips 8.0.0+ compiled against mozjpeg 3.0+_
An advanced setting to reduce the effects of
[ringing](http://en.wikipedia.org/wiki/Ringing_%28signal%29) in JPEG output,
in particular where black text appears on a white background (or vice versa).
#### optimiseScans() / optimizeScans()
_Requires libvips 8.0.0+ compiled against mozjpeg 3.0+_
An advanced setting for progressive (interlace) JPEG output.
Calculates which spectrum of DCT coefficients uses the fewest bits.
Usually reduces file size at the cost of increased compression time.
### Attributes
#### format
An Object containing nested boolean values
representing the available input and output formats/methods,
for example:
```javascript
> console.dir(sharp.format);
{ jpeg: { id: 'jpeg',
input: { file: true, buffer: true, stream: true },
output: { file: true, buffer: true, stream: true } },
png: { id: 'png',
input: { file: true, buffer: true, stream: true },
output: { file: true, buffer: true, stream: true } },
webp: { id: 'webp',
input: { file: true, buffer: true, stream: true },
output: { file: true, buffer: true, stream: true } },
tiff: { id: 'tiff',
input: { file: true, buffer: true, stream: true },
output: { file: true, buffer: false, stream: false } },
magick: { id: 'magick',
input: { file: true, buffer: true, stream: true },
output: { file: false, buffer: false, stream: false } },
raw: { id: 'raw',
input: { file: false, buffer: false, stream: false },
output: { file: false, buffer: true, stream: true } } }
```
#### queue
An EventEmitter that emits a `change` event when a task is either:
* queued, waiting for _libuv_ to provide a worker thread
* complete
```javascript
sharp.queue.on('change', function(queueLength) {
console.log('Queue contains ' + queueLength + ' task(s)');
});
```
### Utilities
#### sharp.cache([memory], [items])
If `memory` or `items` are provided, set the limits of _libvips'_ operation cache.
* `memory` is the maximum memory in MB to use for this cache, with a default value of 100
* `items` is the maximum number of operations to cache, with a default value of 500
This method always returns cache statistics, useful for determining how much working memory is required for a particular task.
```javascript
var stats = sharp.cache(); // { current: 75, high: 99, memory: 100, items: 500 }
sharp.cache(200); // { current: 75, high: 99, memory: 200, items: 500 }
sharp.cache(50, 200); // { current: 49, high: 99, memory: 50, items: 200}
```
#### sharp.concurrency([threads])
`threads`, if provided, is the Number of threads _libvips'_ should create for processing each image. The default value is the number of CPU cores. A value of `0` will reset to this default.
This method always returns the current concurrency.
```javascript
var threads = sharp.concurrency(); // 4
sharp.concurrency(2); // 2
sharp.concurrency(0); // 4
```
The maximum number of images that can be processed in parallel is limited by libuv's `UV_THREADPOOL_SIZE` environment variable.
#### sharp.counters()
Provides access to internal task counters.
* `queue` is the number of tasks this module has queued waiting for _libuv_ to provide a worker thread from its pool.
* `process` is the number of resize tasks currently being processed.
```javascript
var counters = sharp.counters(); // { queue: 2, process: 4 }
```

63
docs/changelog.md Normal file
View File

@ -0,0 +1,63 @@
# Changelog
### v0.11 - "*knife*"
#### v0.11.0 - 15<sup>th</sup> July 2015
* Allow alpha transparency compositing via new `overlayWith` method.
[#97](https://github.com/lovell/sharp/issues/97)
[@gasi](https://github.com/gasi)
* Expose raw ICC profile data as a Buffer when using `metadata`.
[#129](https://github.com/lovell/sharp/issues/129)
[@homerjam](https://github.com/homerjam)
* Allow image header updates via a parameter passed to existing `withMetadata` method.
Provide initial support for EXIF `Orientation` tag,
which if present is now removed when using `rotate`, `flip` or `flop`.
[#189](https://github.com/lovell/sharp/issues/189)
[@h2non](https://github.com/h2non)
* Tighten constructor parameter checks.
[#221](https://github.com/lovell/sharp/issues/221)
[@mikemorris](https://github.com/mikemorris)
* Allow one input Stream to be shared with two or more output Streams via new `clone` method.
[#235](https://github.com/lovell/sharp/issues/235)
[@jaubourg](https://github.com/jaubourg)
* Use `round` instead of `floor` when auto-scaling dimensions to avoid floating-point rounding errors.
[#238](https://github.com/lovell/sharp/issues/238)
[@richardadjogah](https://github.com/richardadjogah)
### v0.10 - "*judgment*"
#### v0.10.1 - 1<sup>st</sup> June 2015
* Allow embed of image with alpha transparency onto non-transparent background.
[#204](https://github.com/lovell/sharp/issues/204)
[@mikemliu](https://github.com/mikemliu)
* Include C standard library for `atoi` as Xcode 6.3 appears to no longer do this.
[#228](https://github.com/lovell/sharp/issues/228)
[@doggan](https://github.com/doggan)
#### v0.10.0 - 23<sup>rd</sup> April 2015
* Add support for Windows (x86).
[#19](https://github.com/lovell/sharp/issues/19)
[@DullReferenceException](https://github.com/DullReferenceException)
[@itsananderson](https://github.com/itsananderson)
* Add support for Openslide input and DeepZoom output.
[#146](https://github.com/lovell/sharp/issues/146)
[@mvictoras](https://github.com/mvictoras)
* Allow arbitrary aspect ratios when resizing images via new `ignoreAspectRatio` method.
[#192](https://github.com/lovell/sharp/issues/192)
[@skedastik](https://github.com/skedastik)
* Enhance output image contrast by stretching its luminance to cover the full dynamic range via new `normalize` method.
[#194](https://github.com/lovell/sharp/issues/194)
[@bkw](https://github.com/bkw)
[@codingforce](https://github.com/codingforce)

103
docs/index.md Normal file
View File

@ -0,0 +1,103 @@
# sharp
The typical use case for this high speed Node.js module
is to convert large images of many formats to
smaller, web-friendly JPEG, PNG and WebP images of varying dimensions.
Resizing an image is typically 4x faster than using
the quickest ImageMagick and GraphicsMagick settings.
[![Test Coverage](https://coveralls.io/repos/lovell/sharp/badge.png?branch=master)](https://coveralls.io/r/lovell/sharp?branch=master)
### Formats
This module supports reading JPEG, PNG, WebP, TIFF, OpenSlide,
GIF and other libmagick-supported formats.
Output images can be in JPEG, PNG and WebP formats as well as uncompressed raw pixel data.
Streams, Buffer objects and the filesystem can be used for input and output.
A single input Stream can be split into multiple processing pipelines and output Streams.
Deep Zoom image pyramids can be generated,
suitable for use with "slippy map" tile viewers like
[OpenSeadragon](https://github.com/openseadragon/openseadragon)
and [Leaflet](https://github.com/turban/Leaflet.Zoomify).
### Features
As well as image resizing, operations such as
rotation, extraction, compositing and gamma correction are available.
Colour spaces, embedded ICC profiles and alpha transparency channels
are all handled correctly.
### Fast
This module is powered by the blazingly fast
[libvips](https://github.com/jcupitt/libvips) image processing library,
originally created in 1989 at Birkbeck College
and currently maintained by
[John Cupitt](https://github.com/jcupitt).
Only small regions of uncompressed image data
are held in memory and processed at a time,
taking full advantage of multiple CPU cores and L1/L2/L3 cache.
Everything remains non-blocking thanks to _libuv_,
no child processes are spawned and Promises/A+ are supported.
### Optimal
Huffman tables are optimised when generating JPEG output images
without having to use separate command line tools like
[jpegoptim](https://github.com/tjko/jpegoptim) and
[jpegtran](http://jpegclub.org/jpegtran/).
PNG filtering can be disabled,
which for diagrams and line art often produces the same result
as [pngcrush](http://pmt.sourceforge.net/pngcrush/).
### Contributing
A [guide for contributors](https://github.com/lovell/sharp/blob/master/CONTRIBUTING.md)
covers reporting bugs, requesting features and submitting code changes.
### Credits
This module would never have been possible without
the help and code contributions of the following people:
* [John Cupitt](https://github.com/jcupitt)
* [Pierre Inglebert](https://github.com/pierreinglebert)
* [Jonathan Ong](https://github.com/jonathanong)
* [Chanon Sajjamanochai](https://github.com/chanon)
* [Juliano Julio](https://github.com/julianojulio)
* [Daniel Gasienica](https://github.com/gasi)
* [Julian Walker](https://github.com/julianwa)
* [Amit Pitaru](https://github.com/apitaru)
* [Brandon Aaron](https://github.com/brandonaaron)
* [Andreas Lind](https://github.com/papandreou)
* [Maurus Cuelenaere](https://github.com/mcuelenaere)
* [Linus Unnebäck](https://github.com/LinusU)
* [Victor Mateevitsi](https://github.com/mvictoras)
* [Alaric Holloway](https://github.com/skedastik)
* [Bernhard K. Weisshuhn](https://github.com/bkw)
Thank you!
### Licence
Copyright 2013, 2014, 2015 Lovell Fuller and contributors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
[http://www.apache.org/licenses/LICENSE-2.0](http://www.apache.org/licenses/LICENSE-2.0.html)
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

104
docs/install.md Normal file
View File

@ -0,0 +1,104 @@
# Installation
```sh
npm install sharp
```
### Prerequisites
* Node.js v0.10+ or io.js
* [libvips](https://github.com/jcupitt/libvips) v7.40.0+ (7.42.0+ recommended)
* C++11 compatible compiler such as gcc 4.6+, clang 3.0+ or MSVC 2013
### Linux
[![Ubuntu 12.04 Build Status](https://travis-ci.org/lovell/sharp.png?branch=master)](https://travis-ci.org/lovell/sharp)
[![Centos 6.5 Build Status](https://snap-ci.com/lovell/sharp/branch/master/build_image)](https://snap-ci.com/lovell/sharp/branch/master)
For a system-wide installation of the most suitable version of
libvips and its dependencies on the following Operating Systems:
* Debian 7, 8
* Ubuntu 12.04, 14.04, 14.10, 15.04
* Mint 13, 17
* RHEL/Centos/Scientific 6, 7
* Fedora 21, 22
* Amazon Linux 2014.09, 2015.03
* OpenSuse 13
* Mac OS
run the following as a user with `sudo` access:
```sh
curl -s https://raw.githubusercontent.com/lovell/sharp/master/preinstall.sh | sudo bash -
```
or run the following as `root`:
```sh
curl -s https://raw.githubusercontent.com/lovell/sharp/master/preinstall.sh | bash -
```
The [preinstall.sh](https://github.com/lovell/sharp/blob/master/preinstall.sh) script requires `curl` and `pkg-config`.
Add `--with-openslide` to enable OpenSlide support:
```sh
curl -s https://raw.githubusercontent.com/lovell/sharp/master/preinstall.sh | sudo bash -s -- --with-openslide
```
### Mac OS
[![OS X 10.9.5 Build Status](https://travis-ci.org/lovell/sharp-osx-ci.png?branch=master)](https://travis-ci.org/lovell/sharp-osx-ci)
Manual install via homebrew:
```sh
brew install homebrew/science/vips --with-webp --with-graphicsmagick
```
A missing or incorrectly configured _Xcode Command Line Tools_ installation
[can lead](https://github.com/lovell/sharp/issues/80) to a
`library not found for -ljpeg` error.
If so, please try: `xcode-select --install`.
The _gettext_ dependency of _libvips_
[can lead](https://github.com/lovell/sharp/issues/9)
to a `library not found for -lintl` error.
If so, please try `brew link gettext --force`.
### Windows
[![Windows Server 2012 Build Status](https://ci.appveyor.com/api/projects/status/pgtul704nkhhg6sg)](https://ci.appveyor.com/project/lovell/sharp)
Requires x86 32-bit Node.js or io.js (use `iojs.exe` rather than `node.exe`).
The WebP format is currently unsupported.
1. Ensure the [node-gyp prerequisites](https://github.com/TooTallNate/node-gyp#installation) are met.
2. [Download](http://www.vips.ecs.soton.ac.uk/supported/current/win32/) and unzip `vips-dev.x.y.z.zip`.
3. Set the `VIPS_HOME` environment variable to the full path of the `vips-dev-x.y.z` directory.
4. Add `vips-dev-x.y.z\bin` to `PATH`.
Versions of MSVC more recent than 2013 may require the use of `npm install --arch=ia32 --msvs_version=2013`.
### Heroku
[Alessandro Tagliapietra](https://github.com/alex88) maintains an
[Heroku buildpack for libvips](https://github.com/alex88/heroku-buildpack-vips)
and its dependencies.
### Docker
[Marc Bachmann](https://github.com/marcbachmann) maintains a
[Dockerfile for libvips](https://github.com/marcbachmann/dockerfile-libvips).
```sh
docker pull marcbachmann/libvips
```
### Build tools
* [gulp-responsive](https://www.npmjs.com/package/gulp-responsive)
* [gulp-sharp](https://www.npmjs.com/package/gulp-sharp)
* [grunt-sharp](https://www.npmjs.com/package/grunt-sharp)

63
docs/performance.md Normal file
View File

@ -0,0 +1,63 @@
# Performance
### Test environment
* AWS EC2 [c4.xlarge](http://aws.amazon.com/ec2/instance-types/#c4)
* Ubuntu 15.04
* Node.js 0.12.7
* libvips 8.0.2
* liborc 0.4.22
### The contenders
* [lwip](https://www.npmjs.com/package/lwip) 0.0.7 - Wrapper around CImg, compiles dependencies from source
* [imagemagick-native](https://www.npmjs.com/package/imagemagick-native) git@45d4e2e - Wrapper around libmagick++, supports Buffers only.
* [imagemagick](https://www.npmjs.com/package/imagemagick) 0.1.3 - Supports filesystem only and "*has been unmaintained for a long time*".
* [gm](https://www.npmjs.com/package/gm) 1.18.1 - Fully featured wrapper around GraphicsMagick's `convert` command line utility.
* sharp 0.11.0 - Caching within libvips disabled to ensure a fair comparison.
### The task
Decompress a 2725x2225 JPEG image, resize to 720x480 using bilinear interpolation, then compress to JPEG.
### Results
| Module | Input | Output | Ops/sec | Speed-up |
| :----------------- | :----- | :----- | ------: | -------: |
| lwip | file | file | 1.75 | 1 |
| lwip | buffer | buffer | 2.21 | 1.3 |
| imagemagick-native | buffer | buffer | 7.13 | 4.1 |
| gm | buffer | buffer | 7.27 | 4.2 |
| gm | file | file | 7.33 | 4.2 |
| imagemagick | file | file | 10.04 | 5.7 |
| sharp | stream | stream | 23.12 | 13.2 |
| sharp | file | file | 24.43 | 14.0 |
| sharp | file | buffer | 24.55 | 14.0 |
| sharp | buffer | file | 24.86 | 14.2 |
| sharp | buffer | buffer | 24.92 | 14.2 |
Greater performance can be expected with caching enabled (default) and using 8+ core machines.
The I/O limits of the relevant (de)compression library will generally determine maximum throughput.
### Benchmark test prerequisites
Requires both _ImageMagick_ and _GraphicsMagick_:
```sh
brew install imagemagick
brew install graphicsmagick
```
```sh
sudo apt-get install imagemagick graphicsmagick libmagick++-dev
```
### Running the benchmark test
```sh
git clone https://github.com/lovell/sharp.git
cd sharp/test/bench
npm install
npm test
```

17
mkdocs.yml Normal file
View File

@ -0,0 +1,17 @@
site_name: sharp
site_url: http://sharp.dimens.io/
repo_url: https://github.com/lovell/sharp
site_description: The fastest Node.js module for resizing JPEG, PNG, WebP and TIFF images. Uses the libvips library.
copyright: <a href="https://dimens.io/">dimens.io</a>
google_analytics: ['UA-13034748-12', 'sharp.dimens.io']
theme: flatly
dev_addr: 0.0.0.0:10101
markdown_extensions:
- toc:
permalink: " ♯"
pages:
- Home: index.md
- Installation: install.md
- API: api.md
- Performance: performance.md
- Changelog: changelog.md