README- UAnimation
- UBlend
- UChromaticities
- UDispose
- UFormat
- UFrame
- UIcc
- UImage
- UIntent
- ULimits
- UMetadata
- UPixel
- UPixelFormat
- URefused
- USampling
- UText
- Vdecode
- Vdecode_animation
- Vencode
- Vencode_animation
moonbitstack/moonimage/raster does not have a README file
Refused
A refusal to read bytes as an image, or to write an image as bytes.
The variants separate the three questions a caller asks: is this the wrong
kind of file (Unknown), is the file damaged (Unexpected, Truncated,
Checksum), or is the caller asking for more than it allowed (OverLimit).
Unsupported is the fourth: a part of a format this library does not read.
Animation
A sequence of frames over one canvas.
Animation::compose
Every frame as a full canvas, in order, with each rule applied as it goes.
The result is what a viewer shows: the same length as frames, each one the
whole canvas. This is the expensive way to watch an animation and the only
way to see it out of context, which is what a test does.
The canvas is sixteen-bit, so that an eight-bit animation and a sixteen-bit
one come out of the same code without either losing its precision.
Blend
Chromaticities
The white point and the three primaries, as chromaticity coordinates.
Eight numbers rather than a matrix: this is what a file writes, and a matrix
would be a conversion, which this library does not do.
Chromaticities::not_equal
Chromaticities::srgb
The primaries of sRGB, which is what almost every file means when it says
nothing.
Dispose
Format
The formats this library reads and writes.
Format::extension
The extension a file of this format usually has, without the dot.
Format::is_animated
Whether the format can hold more than one picture.
An animated format holds frames over one canvas; TIFF holds pages that are
each a picture of its own. Both come out of decode_animation, because a
caller walking pictures does not care which of the two it was handed.
Format::mime
The media type a server would send for this format.
Format::of
Which format these bytes are, by signature.
Raises Unknown when nothing matches and Truncated when there are too few
bytes to tell: two bytes of a PNG are not a PNG, and saying so is better than
guessing.
Frame
Icc
An embedded ICC profile, kept by name and bytes. Not applied: a caller that
wants colour management hands these bytes to something that does it.
Image
pub(all) struct Image {
width : Int
height : Int
format : PixelFormat
stride : Int
pixels : Bytes
premultiplied : Bool
metadata : Metadata
}An image: its shape, its layout, and the bytes.
Image::blur
The image blurred with a Gaussian of standard deviation radius.
A radius of zero or less is the image itself, so a caller that computes the
radius from a size does not have to check for the case.
Image::brightness
The image with every colour channel multiplied by factor.
One is unchanged, zero is black, and two doubles: this is the linear part of
exposure, done on the encoded samples rather than on light, which is what a
brightness slider in an editor does.
Image::composite
This image with source laid over it, its top-left corner at (x, y).
A source larger than the destination is clipped rather than refused: a
watermark is usually placed by a corner rule, and a corner rule that lands
one pixel outside is not an error.
opacity scales the source's alpha before it is laid down, which is how a
watermark is made faint without being re-encoded.
Image::contrast
The image with its contrast scaled about mid-grey.
One is unchanged, zero is flat grey, and above one pushes the samples away
from the middle: (sample − 0.5) × factor + 0.5, clamped.
Image::convert
The same image, in another sample format.
Colour is preserved as far as the target can hold it: a grey image becomes
grey, an image without alpha becomes opaque, and narrowing rounds to nearest.
Image::crop
The rectangle at (x, y) of the given size, as an image of its own.
Raises when the rectangle does not fit: a crop is a claim about the image,
and a crop that reaches outside it is a bug in the caller.
Image::flip_horizontal
The image reflected left to right.
Image::flip_vertical
The image reflected top to bottom.
Image::hue_rotate
The image with every hue turned by degrees, clockwise.
Saturation and lightness survive, which is what makes this different from
swapping channels; a grey pixel has no hue and comes back as it was.
Image::invert
The image with every sample replaced by one minus itself.
Alpha is left alone, so inverting a transparent picture keeps it
transparent.
Image::new
fn Image::new(width : Int, height : Int, format : PixelFormat, stride? : Int, premultiplied? : Bool, metadata? : Metadata) -> ImageA black image of this shape, with tight rows.
stride is only needed when the rows must be padded; leave it out and the
rows are as wide as the pixels.
Image::of
fn Image::of(width : Int, height : Int, format : PixelFormat, pixels : Bytes, stride? : Int, premultiplied? : Bool, metadata? : Metadata) -> Image raise RefusedAn image over bytes that already exist.
Raises when the bytes cannot hold the shape: a stride narrower than a row, or
fewer bytes than the last row needs.
Image::orient
The image turned the way EXIF orientation code says.
One is the picture as stored and comes back unchanged. A number outside one
to eight is refused rather than guessed at: a caller that got the tag wrong
wants to hear about it.
Image::pixel
The pixel at a coordinate, in the units every operation works in.
Straight alpha, as stored: an image that says it is premultiplied returns
premultiplied samples, because pretending otherwise would be a conversion the
caller did not ask for.
Image::rotate
The image turned a quarter turn clockwise, quarters times.
Negative counts turn anticlockwise, and anything beyond four wraps: this is
a total function because a rotation is.
Image::rotate_by
The image turned clockwise by degrees.
expand grows the canvas to the rectangle the turned picture needs, which
is what a caller rotating a photograph wants; without it the picture is
turned inside the canvas it already had, and the corners are cut off, which
is what a caller animating a rotation wants.
A multiple of ninety degrees is exact and is handed to rotate; nothing is
resampled and no sample changes.
Image::row
One row, without its padding.
Image::scale
The image at a new size.
Raises when either dimension is not positive, and when the result would be
larger than Limits::default() — a scaler is easy to point at a huge size by
accident.
Image::to_premultiplied
The same image with colour multiplied by alpha.
Idempotent: an image that already says it is premultiplied comes back
unchanged, so a caller can ask without checking first.
Image::to_straight
The same image with alpha divided back out.
The inverse of to_premultiplied up to rounding, which is why a round trip
is only exact for images whose colour is already premultiplied.
Image::unsharp_mask
The image sharpened by subtracting a blurred copy of itself.
amount is a percentage, a hundred being the difference added once: an edge
that the blur softened is pushed back out by the same distance it lost.
threshold is how much of a difference, in eight-bit steps, is worth
sharpening at all, so that the grain of a photograph is not sharpened into
noise. Sharpen with radius 1, amount 150 and threshold 3 is the usual
starting point.
Intent
What a column of pixels was meant to be seen in.
Intent::code
The number a file writes for this intent, which is how PNG's sRGB chunk
numbers them.
Intent::of
The intent a number stands for, or None if it is not one of the four.
Limits
pub(all) struct Limits {
width : Int
height : Int
pixels : Int
bytes : Int
}The ceilings a decoder works under.
The defaults are the ones a desktop application can live with: a texture no
larger than a modern GPU takes, and a megabyte of pixels at the very most.
A caller reading files from the network should set them lower; a caller
reading its own assets can leave them alone.
Limits::allows
Whether an image of this shape is inside the limits.
Limits::check
Refuse an image of this shape unless the limits allow it.
The pixel count is multiplied in 64 bits so that a declaration of
65535 × 65535 cannot wrap into a small number and pass.
Limits::check_bytes
Refuse a buffer of this many bytes unless the limits allow it.
Limits::default
The desktop defaults: a million pixels a side, 256 megapixels in all, and a
gigabyte of expansion.
Metadata
Everything the file said about itself besides its pixels.
Metadata::default
No metadata at all, which is what an image built in memory has.
Metadata::is_empty
Whether there is nothing to write.
Pixel
One pixel, in the units every operation in the library works in: four
samples from 0 to 1, alpha straight.
A grey pixel has r, g and b equal, which is why there is no grey
variant: a reader that has to branch on whether a pixel is grey is a reader
that will forget to inflate one of the three channels.
Pixel::of_color
The colour as a pixel, converted to sRGB first.
A colour outside sRGB is brought into it rather than clipped in whatever
space it arrived in, and the conversion keeps the alpha it was given.
Pixel::of_rgba8
A pixel from the eight-bit values an editor shows.
Pixel::rgba16
The pixel as sixteen-bit samples, which is what a PNG stores when its bit
depth is sixteen.
Pixel::rgba8
The pixel as bytes an editor shows: rounded, clamped, alpha last.
Pixel::rotate_hue
The pixel with its hue turned, in degrees.
Done in HSL so that the hue is what moves: lightness and saturation come out
as they went in, which is not true of a hue rotation done by mixing channel
matrices. A grey pixel has no hue to turn and comes back unchanged.
Pixel::to_color
The pixel as a colour in sRGB, alpha included.
PixelFormat
How the samples of an image are laid out.
The names are the usual ones: Gray8 is one byte per pixel, Rgba8 four,
Rgba16 eight with each channel big-endian, RgbaF32 four floats.
impl Show for PixelFormatThe layout spelled the way the type spells it: Rgba8, Gray16.
A log line and an error message both want to say which layout was in play,
and neither wants a sentence about it.
PixelFormat::bits
How many bits one sample takes. For RgbaF32 this is 32, which says how wide
the sample is, not that it is an integer.
PixelFormat::bytes
How many bytes a pixel of this format takes.
PixelFormat::channels
How many samples a pixel of this format has.
PixelFormat::has_alpha
Whether the image carries alpha.
PixelFormat::is_float
Whether the samples are floats rather than integers.
PixelFormat::is_gray
Whether the image carries colour at all.
Sampling
Text
A text note, in the flavour of the chunk it came from, so that writing the
image back produces the same kind of chunk.
float_at
fn float_at(data : BytesView, offset : Int) -> FloatA little-endian float, the layout a GPU buffer wants.
Public because a format whose samples are not held to nought and one — an
exposure, a depth — builds its image from these bytes directly, and the
float in them is a value rather than a colour.
float_to_half
fn float_to_half(value : Float) -> IntA float as a half-precision sample, rounded to nearest.
hex
fn hex(bytes : BytesView, at : Int, count : Int) -> StringA few bytes as hexadecimal, for the got field of a refusal.
put_float
fn put_float(data : FixedArray[Byte], offset : Int, value : Float) -> UnitStore a little-endian float, the inverse of float_at.
Install
Download zipImage codecs for MoonBit: PNG, JPEG, GIF, WebP, BMP, TGA, QOI, TIFF, ICO, PNM, Radiance HDR, OpenEXR, farbfeld and DDS, read and written, with the pixel arithmetic, filters and colour between them.
imageimagesimage-codeccodeccodecspngapngjpegjpgprogressive-jpegexificcgifwebpqoibmptgatifflzwpackbitsbigtifficocuriconcursorpnmpbmpgmppmpamnetpbmhdrradiancergbeexropenexrhalf-floatfloat-imagefarbfeldddsbc1bc3bc5bc6hbc7bptcdxtdecodeencodetranscodeconvertrasterpixelpixelsrgbaresizescaleresamplelanczosbilinearcatmull-romgaussianblursharpenunsharp-maskcroprotateflippremultipliedalphaanimationmetadatathumbnailtexturemoonbit
Apache-2.0
Updated 9 hours ago
47 downloads