About Nuspell

Nuspell is a fast and safe spelling checker software program. It is designed for languages with rich morphology and complex word compounding. Nuspell is written in modern C++ and it supports Hunspell dictionaries.

Main features of Nuspell spelling checker:

• Provides software library and command-line tool.
• Suggests high-quality spelling corrections.
• Backward compatibility with Hunspell dictionary file format.
• Up to 3.5 times faster than Hunspell.
• Full Unicode support backed by ICU.
• Twofold affix stripping (for agglutinative languages, like Azeri, Basque, Estonian, Finnish, Hungarian, Turkish, etc.).
• Supports complex compounds (for example, Hungarian, German and Dutch).
• Supports advanced features, for example: special casing rules (Turkish dotted i or German sharp s), conditional affixes, circumfixes, fogemorphemes, forbidden words, pseudoroots and homonyms.
• Free and open source software. Licensed under GNU LGPL v3 or later.

Building Nuspell

Dependencies

Build-only dependencies:

• C++ 17 compiler with support for std::filesystem, e.g. GCC >= v9
• CMake >= v3.12
• Catch2 >= v3.1.1 (optional, needed only when building the tests is enabled)
• Getopt (Needed only on Windows + MSVC and only when the CLI tool or the tests are built. It is available in Vcpkg. Other platforms provide it out of the box.)
• Pandoc (optional, needed only when building the man-pages is enabled)
• Doxygen (optional, needed only when building the API docs is enabled)

Run-time (and build-time) dependencies:

• ICU4C

Recommended tools for developers: qtcreator, ninja, clang-format, gdb, vim.

Building on GNU/Linux and Unixes

We first need to download the dependencies. Some may already be preinstalled.

For Ubuntu and Debian:

sudo apt install g++ cmake libicu-dev catch2 pandoc doxygen

Then run the following commands inside the Nuspell directory:

mkdir build
cd build
cmake .. -DBUILD_API_DOCS=ON
make
sudo make install

For faster build process run make -j, or use Ninja instead of Make.

If you are making a Linux distribution package (dep, rpm) you need some additional configurations on the CMake invocation. For example:

cmake .. -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=/usr

Building on OSX and macOS

1. Install Apple's Command-line tools.
2. Install Homebrew package manager.
3. Install dependencies with the next commands.

brew install cmake icu4c catch2 pandoc doxygen
export ICU_ROOT=$(brew --prefix icu4c)

Then run the standard cmake and make. See above. The ICU_ROOT variable is needed because icu4c is keg-only package in Homebrew and CMake can not find it by default. Alternatively, you can use -DICU_ROOT=... on the cmake command line.

If you want to build with GCC instead of Clang, you need to pull GCC with Homebrew and rebuild all the dependencies with it. See Homewbrew manuals.

Building on Windows

Compiling with Visual C++

1. Install Visual Studio 2017 or newer. Alternatively, you can use Visual Studio Build Tools.
2. Install Git for Windows and Cmake.
3. Install Vcpkg in some folder, e.g. in c:\vcpkg.
4. Install Pandoc. You can manually install or use choco install pandoc.
5. Install Doxygen. You can manually install or use choco install doxygen.install.
6. Run the commands bellow. Vcpkg will work in manifest mode and it will automatically install the dependencies.

mkdir build
cd build
cmake .. -DCMAKE_TOOLCHAIN_FILE=c:\vcpkg\scripts\buildsystems\vcpkg.cmake -A x64 -DBUILD_API_DOCS=ON
cmake --build .

Compiling with Mingw64 and MSYS2

Download MSYS2, update everything and install the following packages:

pacman -S base-devel mingw-w64-x86_64-toolchain mingw-w64-x86_64-icu \
          mingw-w64-x86_64-cmake mingw-w64-x86_64-catch mingw-w64-x86_64-doxygen

Then from inside the Nuspell folder run:

mkdir build
cd build
cmake .. -G "Unix Makefiles" -DBUILD_MAN=OFF -DBUILD_API_DOCS=ON
make
make install

Building in Cygwin environment

Download the above mentioned dependencies with Cygwin package manager. Then compile the same way as on Linux. Cygwin builds depend on Cygwin1.dll.

Building on FreeBSD

Install the following required packages

pkg cmake icu catch2 pandoc doxygen

Then run the standard cmake and make as on Linux. See above.

Using the software

Using the command-line tool

The main executable is located in src/nuspell.

After compiling and installing you can run the Nuspell spell checker with a Nuspell, Hunspell or Myspell dictionary:

nuspell -d en_US text.txt

For more details run see the man-page.

Using the Library

Sample program:

#include <iostream>
#include <nuspell/dictionary.hxx>
#include <nuspell/finder.hxx>

using namespace std;

int main()
{
	auto dirs = vector<filesystem::path>();
	nuspell::append_default_dir_paths(dirs);
	auto dict_path = nuspell::search_dirs_for_one_dict(dirs, "en_US");
	if (empty(dict_path))
		return 1; // Return error because we can not find the requested
		          // dictionary.

	auto dict = nuspell::Dictionary();
	try {
		dict.load_aff_dic(dict_path);
	}
	catch (const nuspell::Dictionary_Loading_Error& e) {
		cerr << e.what() << '\n';
		return 1;
	}
	auto word = string();
	auto sugs = vector<string>();
	while (cin >> word) {
		if (dict.spell(word)) {
			cout << "Word \"" << word << "\" is ok.\n";
			continue;
		}

		cout << "Word \"" << word << "\" is incorrect.\n";
		dict.suggest(word, sugs);
		if (sugs.empty())
			continue;
		cout << "  Suggestions are: ";
		for (auto& sug : sugs)
			cout << sug << ' ';
		cout << '\n';
	}
}

On the command line you can link like this:

g++ example.cxx -std=c++17 -lnuspell -licuuc -licudata
# or better, use pkg-config
g++ example.cxx -std=c++17 $(pkg-config --cflags --libs nuspell)

Within Cmake you can use find_package() to link. For example:

find_package(Nuspell)
add_executable(myprogram main.cpp)
target_link_libraries(myprogram Nuspell::nuspell)

Dictionaries

Myspell, Hunspell and Nuspell dictionaries:

https://github.com/nuspell/nuspell/wiki/Dictionaries-and-Contacts

Advanced topics

Debugging Nuspell

First, always install the debugger:

sudo apt install gdb

For debugging we need to create a debug build and then we need to start gdb.

mkdir debug
cd debug
cmake .. -DCMAKE_BUILD_TYPE=Debug
make -j
gdb src/nuspell/nuspell

We recommend debugging to be done with an IDE.

Testing

To run the tests, run the following command after building:

ctest

See also

Full documentation in the wiki.

API Documentation for developers will be generated from the source files by configuring CMake with -DBUILD_API_DOCS=ON and building.

The result can be viewed by opening BUILD_DIR/docs/html/index.html during development, or by opening INSTALL_PREFIX/share/doc/nuspell/html/index.html after installing.