Check misspellings from any text source with the most popular PHP spellchecker.
PHP-Spellchecker is a spellchecker abstraction library for PHP. By providing a unified interface for many different spellcheckers, you’re able to swap out spellcheckers without extensive rewrites.
Using PHP-Spellchecker can eliminate vendor lock-in, reduce technical debt, and improve the testability of your code.
- 🧐 Supports many popular spellcheckers out of the box: Aspell, Hunspell, Ispell, PHP Pspell, LanguageTools, JamSpell and MultiSpellchecker (add yours!)
- 📄 Supports different text sources: file system file/directory, string, and multi-source (add yours!)
- 🛠 Supports text processors: MarkdownRemover (add yours!)
- 🔁 Supports misspelling handlers: EchoHandler (add yours!)
- ➰ Makes use of generators to reduce memory footprint
- ⚖ Flexible and straightforward design
- 💡 Makes it a breeze to implement your own spellcheckers, text processors and misspellings handlers
- 💪 Runs tests against real spellcheckers to ensure full compatibility
PHP-Spellchecker is a welcoming project for new contributors.
Want to make your first open source contribution? Check the roadmap, pick one task, open an issue and we'll help you go through it 🤓🚀
Via Composer
$ composer require tigitz/php-spellcheckerCheck out the documentation and examples
You can check misspellings directly from a PhpSpellcheck\Spellchecker class and process them on your own.
<?php
// if you made the default aspell installation on your local machine
$aspell = Aspell::create();
// or if you want to use binaries from Docker
$aspell = new Aspell(new CommandLine(['docker','run','--rm', '-i', 'starefossen/aspell']));
$misspellings = $aspell->check('mispell', ['en_US'], ['from_example']);
foreach ($misspellings as $misspelling) {
$misspelling->getWord(); // 'mispell'
$misspelling->getLineNumber(); // '1'
$misspelling->getOffset(); // '0'
$misspelling->getSuggestions(); // ['misspell', ...]
$misspelling->getContext(); // ['from_example']
}You can also use an opinionated MisspellingFinder class to orchestrate your spellchecking flow:
Following the well-known Unix philosophy:
Write programs that do one thing and do it well. Write programs to work together. Write programs to handle text streams, because that is a universal interface.
<?php
// My custom text processor that replaces "_" by " "
$customTextProcessor = new class implements TextProcessorInterface
{
public function process(TextInterface $text): TextInterface
{
$contentProcessed = str_replace('_', ' ', $text->getContent());
return $text->replaceContent($contentProcessed);
}
};
$misspellingFinder = new MisspellingFinder(
Aspell::create(), // Creates aspell spellchecker pointing to "aspell" as it's binary path
new EchoHandler(), // Handles all the misspellings found by echoing their information
$customTextProcessor
);
// using a string
$misspellingFinder->find('It\'s_a_mispelling', ['en_US']);
// word: mispelling | line: 1 | offset: 7 | suggestions: mi spelling,mi-spelling,misspelling | context: []
// using a TextSource
$inMemoryTextProvider = new class implements SourceInterface
{
public function toTexts(array $context): iterable
{
yield new Text('my_mispell', ['from_source_interface']);
// t() is a shortcut for new Text()
yield t('my_other_mispell', ['from_named_constructor']);
}
};
$misspellingFinder->find($inMemoryTextProvider, ['en_US']);
//word: mispell | line: 1 | offset: 3 | suggestions: mi spell,mi-spell,misspell,... | context: ["from_source_interface"]
//word: mispell | line: 1 | offset: 9 | suggestions: mi spell,mi-spell,misspell,... | context: ["from_named_constructor"]The project is still in its initial phase, requiring more real-life usage to stabilize its final 1.0.0 API.
- Add a CLI that could do something like
vendor/bin/php-spellchecker "misspell" Languagetools EchoHandler --lang=en_US - Add asynchronous mechanism to spellcheckers.
- Make some computed misspelling properties optional to improve performance for certain use cases (e.g., lines and offset in
LanguageTools). - Add a language mapper to manage different representations across spellcheckers.
- Evaluate
strtokinstead ofexplodeto parse lines of text, for performance. - Evaluate
MutableMisspellingfor performance comparison. - Wrap
Webmozart/Assertlibrary exceptions to throw PHP-Spellchecker custom exceptions instead. - Improve the
Makefile.
- Make a
SourceInterfaceclass that's able to have an effect on the used spellchecker configuration. -
League/Flysystemsource. -
Symfony/Findersource.
- Markdown - Find a way to keep the original offset and line of words after stripping.
- Add PHPDoc processor.
- Add HTML Processor (inspiration).
- Add XLIFF Processor (inspiration).
- Cache suggestions of already spellchecked words (PSR-6/PSR-16?).
- Pspell - Find way to compute word offset.
- LanguageTools - Evaluate HTTPlug library to make API requests.
- Pspell - find way to list available dictionaries.
- Add JamSpell spellchecker.
- Add NuSpell spellchecker.
- Add SymSpell spellchecker.
- Add Yandex.Speller API spellchecker.
- Add Bing Spell Check API spellchecker.
- MonologHandler
- ChainedHandler
- HTMLReportHandler
- XmlReportHandler
- JSONReportHandler
- ConsoleTableHandler
- Add or improve tests with different text encoding.
- Refactor duplicate Dockerfile content between PHP images.
We follow SemVer v2.0.0.
There still are many design decisions that should be confronted with real-world usage before thinking about a v1.0.0 stable release:
- Are
TextInterfaceandMisspellingInterfacereally useful? - Is using generators the right way to go?
- Should all the contributed spellcheckers be maintained by the package itself?
- How to design an intuitive CLI given the needed flexibility of usage?
- Is the "context" array passed through all the layers the right design to handle data sharing?
Spell checkers come in many different forms, from HTTP API to command line tools. PHP-Spellchecker wants to ensure real-world usage is OK, so it contains integration tests. To run these, spellcheckers need to all be available during tests execution.
The most convenient way to do it is by using Docker and avoid polluting your local machine.
Requires docker and docker-compose to be installed (tested on Linux).
$ make build # build container images
$ make setup # start spellcheckers container
$ make tests-doxYou can also specify PHP version, dependency version target and if you want coverage.
$ PHP_VERSION=8.2 DEPS=LOWEST WITH_COVERAGE="true" make tests-doxRun make help to list all available tasks.
Todo
If spellcheckers execution paths are different than their default values (e.g., docker exec -ti myispell instead of ispell) you can override the path used in tests by redefining environment variables in the PHPUnit config file.
Please see CONTRIBUTING.
- Inspired by php-speller, monolog and flysystem
- Philippe Segatori
- All Contributors
The MIT License (MIT). Please see license file for more information.
Logo: Elements taken for the final rendering are Designed by rawpixel.com / Freepik.