A Dart implementation of the "Quite OK Image Format", with a command line interface for console use and a library for use in applications.
Based off of the official C implementation and other implementations.
I'm an under-18 currently living in the UK, and I am in full-time education. I work on this project and all of my others in my spare time, and I currently have no stable income due to my age - although I aspire to work in software/mobile development in the future.
If you have any change to spare, I'd be grateful for any amount, big or small :D. Every donation gives me 'mental fuel' to continue this project, and lets me know that I'm doing a good job. I'll be happy to give you a place on the documentation website's credits, and a shoutout in every release/CHANGELOG.
You can read more about me and what I do on my GitHub Sponsors page, where you can donate as well.
Alternatively, if you prefer not to use GitHub Sponsors, please feel free to use my Ko-fi. Note that the PayPal backend will take a small percentage amount of donations made through this method.
Command Line Interface
An easy to use CLI is provided to get working with QOI!
With Dart (recommended)
If you have Dart installed, you can use the CLI on any operating system!
Just run the command
dart pub global activate dqoi (with administrator/root privileges), then you can use the command
dqoi from anywhere on your system!
If you don't have Dart installed, you can use the pre-compiled executable for Windows.
You can get the .exe by:
- Cloning this repository with Git, then going to 'bin/dqoi.exe'
- Downloading just the executable from the GitHub repo browser: github.com/JaffaKetchup/dqoi/blob/main/bin/dqoi.exe.
Note that Windows Defender or your anti-virus may flag the executable malicious or unwanted, as it is not signed. You'll need to make an exception for the program if this happens.
Add this file to your system path, then you can use the command
dqoi from anywhere on your system!
Other Operating Systems
Unfortunately, I cannot provide executables for other operating systems at this time, as I do not have the appropriate devices.
The best way to get the executable for your OS is to install Dart, then follow the instructions above for setup with Dart.
You can list the available options by just running
dqoi with no arguments or with '--help'. The program will return helpful messages in the event of an error, which should be self-explanatory.
dqoi-test.bat is provided to test the program, comparing the output files with official samples. This is not necessary to run, but may help to verify that the program is working correctly. In the event of a test failure, there will be differences shown in the console.
There are two included libraries, each of which are very similar:
- Flutter applications should use the standard
- Non-Flutter programs should use the
package:dqoi/dqoi_pure.dartimport, which excludes some useful Flutter-only methods.
Note that both export the
Channels enum from the 'image' package, but you can disable this by using
hide Channels on the end of the import statement.
Also note that neither uses 'dart:io', so both are fully compatible with Web applications.
When this documentation refers to the singular "library", it means either library.
QOI class provides access to conversions between binary, PNG, and QOI formats.
When extended by
FlutterQOIExts, it also provides an easy way to render/paint a QOI image in a Flutter app efficiently.
You can find the full API documentation at pub.dev/documentation/dqoi/latest/dqoi/dqoi-library.html.
You can build and install the example application, found in the 'example/' directory. However, below are some quick useful snippets to get you started.
Convert a PNG file to QOI, and then write to another file:
await outputFile.writeAsBytes(QOI.fromPNG(await inputFile.readAsBytes()).toQOI());
Render/paint a QOI file to a widget:
return QOI.fromQOI(await inputFile.readAsBytes()).toImageWidget(loadingWidget: loadingWidget);
Render/paint a QOI asset (bundled) to a widget:
Future<Uint8List> loadAsset(String path) async => (await rootBundle.load(path)).buffer.asUint8List(); return QOI.fromQOI(loadAsset('assetPath.qoi')).toImageWidget(loadingWidget: loadingWidget);
- How do I pronounce the name of this library?
It's up to you, but I like "decoy" best.
- Is this a good implementation?
The outputted QOI files perfectly match the official C implementation's outputs. Any re-encoded PNGs don't always match byte-for-byte, but the pixels are always correct
- Are there any other implementations in different languages?
Sure there are! You can see all the other available ports on the official README.