pyspellchecker Documentation

To use pyspellchecker, import SpellChecker, identify misspelled words using .unknown(), and then retrieve corrections using .correction() or a list of possibilities using .candidates().

from spellchecker import SpellChecker

spell = SpellChecker()

# find those words that may be misspelled
misspelled = spell.unknown(['something', 'is', 'hapenning', 'here'])

for word in misspelled:
    # Get the one `most likely` answer
    print(spell.correction(word))

    # Get a list of `likely` options
    print(spell.candidates(word))

To use the spell checker, import SpellChecker from the spellchecker module and initialize an instance.

from spellchecker import SpellChecker

spell = SpellChecker()

The default Levenshtein distance is 2. For longer words, it is highly recommended to reduce the distance to 1 to improve performance and accuracy. You can set this during initialization or update the property on the instance later.

from spellchecker import SpellChecker

# Set distance at initialization
spell = SpellChecker(distance=1)

# Or set the distance parameter after the fact
spell.distance = 2

When bundling an executable with PyInstaller, you must manually include the required dictionary resource files. These files must be placed in a folder named spellchecker/resources/ within your executable to match the library's expected lookup path.

# Linux/macOS
pyinstaller --add-binary="spellchecker/resources/en.json.gz:spellchecker/resources" my_prog.py

# Windows
pyinstaller --add-binary="spellchecker/resources/en.json.gz;spellchecker/resources" my_prog.py

You can install pyspellchecker using pip. For Python 3 (the recommended version), use the standard install command. If you require legacy support for Python 2.7, you must install version 0.5.6 specifically.

# Standard installation for Python 3
pip install pyspellchecker

# For Python 2.7 support
pip install pyspellchecker==0.5.6

To build a new dictionary from scratch, initialize SpellChecker with language=None. You can then load your corpus and export the result for later use using spell.export(filename, gzipped=True).

from spellchecker import SpellChecker

# Initialize without built-in languages
spell = SpellChecker(language=None, case_sensitive=True)

# Load your custom data
spell.word_frequency.load_dictionary('./my_dict.json')

# Export for later use
spell.export('my_custom_dictionary.gz', gzipped=True)

You can improve the spell checker's accuracy for your specific domain by loading text from a file or manually adding specific words to the frequency list. This ensures that domain-specific terms are not flagged as misspelled.

from spellchecker import SpellChecker

spell = SpellChecker()  # loads default word frequency list

# Load additional text to generate a custom frequency list
spell.word_frequency.load_text_file('./my_free_text_doc.txt')

# Manually add words to ensure they are not flagged as misspelled
spell.word_frequency.load_words(['microsoft', 'apple', 'google'])

# Verify the words are now known
print(spell.known(['microsoft', 'google']))  # returns both

The distance parameter controls the Levenshtein distance used for corrections. The default is 2. For processing long words more quickly, you can set it to 1 during initialization or update the property on the instance.

from spellchecker import SpellChecker

# Set distance at initialization
spell = SpellChecker(distance=1)

# Or update the property later
spell.distance = 2

You can initialize SpellChecker with different languages by passing the language parameter. Supported language codes include:

• English: 'en' (default)
• Spanish: 'es'
• French: 'fr'
• Portuguese: 'pt'
• German: 'de'
• Italian: 'it'
• Russian: 'ru'
• Arabic: 'ar'
• Basque: 'eu'
• Latvian: 'lv'
• Dutch: 'nl'
• Persian: 'fa'

from spellchecker import SpellChecker

english = SpellChecker()           # default is English
spanish = SpellChecker(language='es')
russian = SpellChecker(language='ru')
arabic = SpellChecker(language='ar')

Once a word is identified as misspelled, you can retrieve suggestions:

• spell.correction(word): Returns the single most likely replacement for the word.
• spell.candidates(word): Returns a set of all possible candidate words.

from spellchecker import SpellChecker

spell = SpellChecker()

# Get the single best correction
misspelled = spell.unknown(['hapenning'])  # {'hapenning'}
for word in misspelled:
    print(spell.correction(word))  # 'happening'

# Get all possible candidates
print(spell.candidates('hapenning'))  # {'penning', 'happening', 'henning'}

You can iterate over the SpellChecker instance to get words. To get the frequency of a word, use a lookup on the instance.

from spellchecker import SpellChecker

spell = SpellChecker()

for word in spell:
    print(f"{word}: {spell[word]}")