The ImageProcessing::MiniMagick
module contains processing methods that use
the MiniMagick gem (which is installed with the image_processing gem).
You will need to install ImageMagick/GraphicsMagick before using this module:
$ brew install imagemagick
# or
$ brew install graphicsmagick
If you're using something other than Homebrew, see the installation instructions for more details.
require "image_processing/mini_magick"
processed = ImageProcessing::MiniMagick
.source(image)
.resize_to_limit(400, 400)
.strip
.call
processed #=> #<Tempfile:/var/folders/.../image_processing20180316-18446-1j247h6.png>
The MiniMagick gem supports GraphicsMagick as well, you just need to specify that you want to use it:
require "image_processing/mini_magick"
MiniMagick.cli = :graphicsmagick
processed = ImageProcessing::MiniMagick
.source(image)
.resize_to_limit(400, 400)
.strip
.call # will use `gm convert` instead of `convert`
processed #=> #<Tempfile:/var/folders/.../image_processing20180316-18446-1j247h6.png>
Tries to recompress the image, and returns true
if no exception was raised,
otherwise returns false
.
ImageProcessing::MiniMagick.valid_image?(normal_image) #=> true
ImageProcessing::MiniMagick.valid_image?(invalid_image) #=> false
Downsizes the image to fit within the specified dimensions while retaining the original aspect ratio. Will only resize the image if it's larger than the specified dimensions. See this article for more details.
pipeline = ImageProcessing::MiniMagick.source(image) # 600x800
result = pipeline.resize_to_limit!(400, 400)
MiniMagick::Image.new(result.path).dimensions #=> [300, 400]
It's possible to omit one dimension, in which case the image will be resized only by the provided dimension.
pipeline.resize_to_limit!(400, nil)
# or
pipeline.resize_to_limit!(nil, 400)
Resizes the image to fit within the specified dimensions while retaining the original aspect ratio. Will downsize the image if it's larger than the specified dimensions or upsize if it's smaller. See this article for more details.
pipeline = ImageProcessing::MiniMagick.source(image) # 600x800
result = pipeline.resize_to_fit!(400, 400)
MiniMagick::Image.new(result.path).dimensions #=> [300, 400]
It's possible to omit one dimension, in which case the image will be resized only by the provided dimension.
pipeline.resize_to_fit!(400, nil)
# or
pipeline.resize_to_fit!(nil, 400)
Resizes the image to fill the specified dimensions while retaining the original aspect ratio. If necessary, will crop the image in the larger dimension. See this article for more details.
pipeline = ImageProcessing::MiniMagick.source(image) # 600x800
result = pipeline.resize_to_fill!(400, 400)
MiniMagick.new(result.path).dimensions #=> [400, 400]
You can specify the direction of the image via the :gravity
option
(defaults to "Center"
)
pipeline.resize_to_fill!(400, 400, gravity: "north-west")
Resizes the image to fit within the specified dimensions while retaining the original aspect ratio. If necessary, will pad the remaining area with the given color. See this article for more details.
pipeline = ImageProcessing::MiniMagick.source(image) # 600x800
result = pipeline.resize_and_pad!(400, 400)
MiniMagick::Image.new(result.path).dimensions #=> [400, 400]
It accepts :background
for specifying the background color that will be
used for padding (defaults to transparent/white).
pipeline.resize_and_pad!(400, 400, background: :transparent) # default
pipeline.resize_and_pad!(400, 400, background: [65, 105, 225]) # RGB value
pipeline.resize_and_pad!(400, 400, background: [65, 105, 225, 1.0]) # RGBA value
pipeline.resize_and_pad!(400, 400, background: "...") # any supported ImageMagick color value
It accepts :gravity
for specifying the [gravity] to apply while cropping
(defaults to "Center"
).
pipeline.resize_and_pad!(400, 400, gravity: "north-west")
Extracts an area from an image. The first two arguments are left & top edges of area to extract, while the last two arguments are the width & height of area to extract:
ImageProcessing::MiniMagick
.crop(20, 50, 300, 300) # extracts 300x300 area with top-left edge 20,50
You can also specify an ImageMagick geometry directly:
ImageProcessing::MiniMagick
.crop("300x300+20+50") # extracts 300x300 area with top-left edge 20,50
Rotates the image by the specified angle. Accepts any value that -rotate
accepts.
ImageProcessing::MiniMagick
.rotate(90)
# ...
For degrees that are not a multiple of 90, you can also specify a background color for the empty triangles in the corners, left over from rotating the image.
rotate(45) # default color
rotate(45, background: :transparent) # transparent
rotate(45, background: [65, 105, 225]) # RGB value
rotate(45, background: [65, 105, 225, 1.0]) # RGBA value
rotate(45, background: "...") # any supported ImageMagick color value
Blends the image with the specified image and an optional mask. One use case for this can be applying a watermark.
composite(overlay)
composite(overlay, mask: mask)
The overlay and mask image can be a String
, Pathname
, or an object that
responds to #path
.
The method of image composition can be specified via the :mode
option (see
Compose Tables for a visual representation of the available methods), and
additional arguments for the compose method can be specified via :args
:
composite(overlay, mode: "src")
composite(overlay, mode: "blend", args: "50,50")
The direction and position of the source or overlay image can be controlled
via :gravity
and :offset
options:
composite(overlay, gravity: "south-east")
composite(overlay, gravity: "north-west", offset: [55, 55])
Any additional options can be specified via a block:
composite(overlay) do |cmd|
cmd.define("compose:outside-overlay=false")
cmd.background("none")
cmd.swap.+
end
See -composite
for more details.
Specifies the output format.
pipeline = ImageProcessing::MiniMagick.source(image)
result = pipeline.convert!("png")
File.extname(result.path)
#=> ".png"
By default the original format is retained when writing the image to a file. If the source file doesn't have a file extension, the format will default to JPEG.
Adds coder/decoder options with -define
from the specified Hash.
ImageProcessing::MiniMagick
.define(png: { compression_level: 8, format: "png8" }) # -define png:compression-level=8 -define png:format=png8
# ...
Any unknown methods will be delegated to MiniMagick::Tool::Convert
. See the
list of all available options by running convert -help
and visiting the
ImageMagick reference.
ImageProcessing::MiniMagick
.quality(100) # -quality 100
.crop("300x300+0+0") # -crop 300x300+0+0
.resample("300x300") # -resample 300x300
.stack { |cmd| ... } # ( ... )
# ...
Yields the intermediary MiniMagick::Tool::Convert
object. If the block return
value is a MiniMagick::Tool::Convert
object it will be used in further
processing, otherwise if nil
is returned the original
MiniMagick::Tool::Convert
object will be used.
ImageProcessing::MiniMagick
.custom { |magick| magick.colorspace("grayscale") if gray? }
# ...
Appends given values directly as arguments to the convert
command.
ImageProcessing::MiniMagick
.append("-quality", 100)
.append("-flip")
# ...
It accepts the following special options:
:loader
-- explicitly set the input file type:page
-- specific page(s) that should be loaded:geometry
-- geometry that should be applied when loading:auto_orient
-- whether the image should be automatically oriented after it's loaded (defaults totrue
):define
-- creates definitions that coders and decoders use for reading and writing image data
ImageProcessing::MiniMagick.loader(loader: "jpg").call(image)
# convert jpg:input.jpg -auto-orient output.jpg
ImageProcessing::MiniMagick.loader(page: 0).convert("png").call(pdf)
# convert input.pdf[0] -auto-orient output.png
ImageProcessing::MiniMagick.loader(geometry: "300x300").call(image)
# convert input.jpg[300x300] -auto-orient output.jpg
ImageProcessing::MiniMagick.loader(auto_orient: false).call(image)
# convert input.jpg output.jpg
ImageProcessing::MiniMagick.loader(define: { jpeg: { size: "300x300" } }).call(image)
# convert -define jpeg:size=300x300 input.jpg -auto-orient output.jpg
All other options given will be interpreted as ImageMagick operations to be
applied before the image is loaded. Operation values can be either a single
argument, an array of arguments, true
/nil
indicating no arguments, or
false
indicating the operator should be a "+" operator. See Reading JPEG
Control Options for some examples.
ImageProcessing::MiniMagick
.loader(strip: true, type: "TrueColorMatte")
.call(image) # convert -strip -type TrueColorMatte input.jpg ... output.jpg
If the #loader
clause is repeated multiple times, the options are merged.
ImageProcessing::MiniMagick
.loader(page: 0)
.loader(geometry: "300x300")
# resolves to
ImageProcessing::MiniMagick
.loader(page: 0, geometry: "300x300")
If you would like to have more control over loading, you can create the
MiniMagick::Tool
object directly, and just pass it as the source file.
magick = MiniMagick::Tool::Convert.new
magick << "..." << "..." << "..."
ImageProcessing::MiniMagick
.source(magick)
# ...
It accepts the following special options:
:define
-- definitions that coders and decoders use for reading and writing image data:allow_splitting
-- allow splitting multi-layer image into multiple single-layer images (defaults tofalse
)
ImageProcessing::MiniMagick.saver(define: { jpeg: { optimize_coding: false } }).call(image)
# convert input.jpg -auto-orient -define jpeg:optimize-coding=false output.jpg
ImageProcessing::MiniMagick.convert("png").call(pdf_document)
# raises ImageProcessing::Error
ImageProcessing::MiniMagick.convert("png").saver(allow_splitting: true).call(pdf_document)
# lets ImageMagick generate a "*-{idx}.png" image for each page
All other options given will be interpreted as ImageMagick operations to be
applied before the image is saved. Operation values can be either a single
argument, an array of arguments, true
/nil
indicating no arguments, or
false
indicating the operator should be a "+" operator. See Writing JPEG
Control Options for some examples. Note that is just syntax sugar over
applying the operations via the chainable API.
ImageProcessing::MiniMagick
.saver(quality: 80, interlace: "Line")
.call(image) # convert input.jpg ... -quality 80 -interlace Line output.jpg
If you would like to have more control over saving, you can call #call(save: false)
to get the MiniMagick::Tool
object, and finish saving yourself.
magick = ImageProcessing::MiniMagick
.resize_to_limit(400, 400)
.call(save: false)
magick #=> #<MiniMagick::Tool::Convert ...>
magick << "output.png"
magick.call
Sets the pixel cache resource limits for the ImageMagick command.
ImageProcessing::MiniMagick
.limits(memory: "50MiB", width: "10MP", time: 30)
.resize_to_limit(400, 400)
.call(image)
# convert -limit memory 50MiB -limit width 10MP -limit time 30 input.jpg ... output.jpg
See the -limit
documentation and the Architecture article for more
details.
This is a convenience method for sending multiple commands to the builder using
a hash. Hash keys can be any method that the builder responds to. Hash values
can be either a single argument, an array of arguments, or true
/nil
indicating no arguments. Instead of a hash you can also use an array if you
want to send multiple commands with the same name.
ImageProcessing::MiniMagick
.apply(
strip: true,
crop: "200x200+0+0",
resize_to_limit: [400, 400],
convert: "jpg",
saver: { quality: 100 },
)
# ...
With #resize_*
operators, you can additional sharpen the thumbnails via the
:sharpen
option, which performs -sharpen
operation (higher sigma means
more sharpening):
ImageProcessing::MiniMagick
.source(image)
.resize_to_limit!(400, 400, sharpen: { radius: 0, sigma: 1 })
# convert input.jpg -resize 400x400> -sharpen 0x1 output.jpg