Sonnet

Sonnet, in sonic-frameworks-spell-check (find_package(KF6Sonnet); KF6::SonnetCore and KF6::SonnetUi), solves spell checking for text editors. Its spell checkers are plugins: Hunspell and Aspell on SonicDE, with plugins for HSpell, Voikko and the Windows and macOS system spell checkers in the source as well. It also detects the language of a text automatically, combining several methods.

#Spell checking in your QTextEdit

Adding Sonnet to a QTextEdit is easy:

QTextEdit *textEdit = new QTextEdit;
textEdit->setText(
    QString::fromLatin1("This is a sample buffer. Whih this thingg will "
                        "be checkin for misstakes. Whih, Enviroment, govermant. Whih."));

Sonnet::SpellCheckDecorator *installer = new Sonnet::SpellCheckDecorator(textEdit);
installer->highlighter()->setCurrentLanguage(QStringLiteral("en_US"));

Sonnet::SpellCheckDecorator can be extended to check differently formatted text, such as emails:

class MailSpellCheckDecorator : public Sonnet::SpellCheckDecorator
{
public:
    explicit MailSpellCheckDecorator(QTextEdit *edit)
        : Sonnet::SpellCheckDecorator(edit)
    {
    }

protected:
    bool isSpellCheckingEnabledForBlock(const QString &blockText) const override
    {
        qDebug() << blockText;
        return !blockText.startsWith(QLatin1Char('>'));
    }
};

Use MailSpellCheckDecorator exactly like SpellCheckDecorator; it also skips the quoted parts of an email (lines starting with >).

#Language detection

Sonnet can tell about 70 languages apart in a string. Its method is based on Languid, a Perl script by Maciej Ceglowski, and works in two steps: first it checks which writing systems the text uses, then, for the languages using those writing systems, it compares an n-gram frequency model of each language with one of the text. The most similar language model wins. If no language is found, the result is an empty string.

A simple example with GuessLanguage:

GuessLanguage languageGuesser;
QString lang = languageGuesser.identify("Mon super texte");

identify() returns the language's ISO 639-1 code (a three-letter code only where no two-letter code exists).

#Widgets

Sonnet also has widgets for configuring it in Qt applications. DictionaryComboBox, for example, is a combo box for choosing the dictionary. Connect it to the highlighter of a decorator to switch the spell check language:

Sonnet::DictionaryComboBox *comboBox = new Sonnet::DictionaryComboBox;
QObject::connect(comboBox, &Sonnet::DictionaryComboBox::dictionaryChanged,
                 installer->highlighter(), &Sonnet::Highlighter::setCurrentLanguage);

Sonnet::Dialog checks a whole text interactively:

void TestDialog::check(const QString &buffer)
{
    Sonnet::Dialog *dlg = new Sonnet::Dialog(new BackgroundChecker(this), nullptr);
    connect(dlg, &Dialog::spellCheckDone, this, &TestDialog::doneChecking);
    dlg->setBuffer(buffer);
    dlg->show();
}

ConfigDialog is a fuller settings dialog: the words to ignore, skipping run-together words, and turning automatic language detection on or off.

The examples come from sonic-frameworks-spell-check/examples (LGPL-2.1-or-later).