5 releases
0.5.5 | Jun 13, 2024 |
---|---|
0.5.3 | Dec 13, 2023 |
0.5.2 | Oct 30, 2023 |
0.5.1 | Oct 30, 2023 |
0.5.0 | Oct 30, 2023 |
#725 in Text processing
315KB
4K
SLoC
ZSpell
This project is a spellchecker written completely in Rust, that maintains compatibility with the venerable Hunspell dictionary format. It is entirely native and does not rely on any other backends (Enchant, Hunspell, Aspell, etc.). This library also has the goal of being usable via WASM. Full Unicode support is baked in.
The library side has a stabalized checker, but the suggestion API is not yet finalized. The CLI is usable but not yet considered stabalized. See Feature Status for more information on what is available.
Here are some useful quick links:
- Crate info: https://crates.io/crates/zspell
- Crate CLI docs (incomplete): https://pluots.github.io/zspell/
- Crate library docs: https://docs.rs/zspell/
- Python library page: https://pypi.org/project/zspell/
- Crate source: https://github.com/pluots/zspell
Interfaces
This project exposes multiple interfaces to its spellchecker, listed in this section.
Command Line Interface
Just want to use this spellchecker from the command line? Check out the book, located here https://pluots.github.io/zspell/, for a more in-depth explanation of installation and usage.
If you don't want to read further, the easiest way to get started is to download a prebuilt binary from here: https://github.com/pluots/zspell/releases.
Rust Library Interface
This project also aims to create a fully functional spellchecking library, for easy programmatic use. See the documentation for the library side here https://docs.rs/zspell/. This also includes a lot of design methodology discussions, for those who are interested.
Python Interface
There is a python wrapper for this library with prebuilt wheels, available here: https://pypi.org/project/zspell/. Its source is located in the zspell-py crate.
Usage via WASM
The library API should work out of the box. Official WASM bindings will be added at some point.
Feature Status
Feature | Available via Library | Available via CLI | Tracking Issue |
---|---|---|---|
Basic spellcheck functionality | ✓ | ✓ | |
Forbidden word handling | ✓ | ✓ | #17 |
Stemming | ✓ | ✓ | |
Morph analysis | ✓ | ✓ | |
Suggestions | WIP | ✕ | #16 |
Compound word handling | ✕ | ✕ | |
Full Morph/Phone Handling | WIP | ✕ | |
Python Interface | Beta | N/A | #18 |
Prebuilt WASM bindings | ✕ | N/A | #19 |
Performance
This repository has the goal of highly prioritizing the most expected usage,
i.e., that most words to be checked are correct. With optimizations based around
this concept and with the modern computers now able to store entire compiled
word lists in memory (~20 MiB), zspell
tends to outperform other
spellcheckers.
MSRV
This library relies on features from Rust 1.65, so that is our current minimum supported version. Our CI validates this for the library and examples.
The CLI and test runner require newer features and do not keep a specific MSRV.
Test suite
This project keeps a test suite located in zspell/test-suite
(symlinked to
test-suite
). Each file has a simple format that combines a simple affix and
dictionary file. To add a test, just duplicate and edit 0-example.test
.
File names are as follows:
0-*
: meta tests that do not get runb-*
: basic functionality testsh-*
: tests that come from the Hunspell test suitei000-*
: tests that address specific issues
License
See the LICENSE file for license information. The provided license does allow for proprietary use and adaptation; that being said, I kindly suggest that if you come up with an improvement, you submit a pull request and help us all out :)
Test suite license
Some tests are taken from Hunspell's test suite. Hunspell has various licenses, we select MPL and include a SPDX notice on relevant files.
Dictionary data license
The dictionaries provided in this repository for testing purposed have been obtained under license. These files have been sourced from here: https://github.com/wooorm/dictionaries
These dictionaries are licensed under various licenses, different from that of
this project. Please see the applicable .license
file withing the
dictionaries/
directory.
Dependencies
~13–25MB
~387K SLoC