Blazingly fast cleaning swear words (and their leetspeak) in strings
Inspired from package profanity of Ben Friedland, this library is significantly faster than the original one, by using string comparison instead of regex.
It supports modified spellings (such as p0rn, h4NDjob, handj0b and b*tCh).
This package only works with Python 3.
$ pip install better_profanity
Only Unicode characters from categories Ll, Lu, Mc and Mn are added. More on Unicode categories can be found here.
Not all languages are supported yet, such as Chinese.
Most of the words in the default wordlist are referred from Full List of Bad Words and Top Swear Words Banned by Google.
The wordlist contains a total of 106,992 words, including 317 words from the default profanity_wordlist.txt and their variants by modified spellings.
Its total size in memory is 10.49+MB.
It is highly recommended to call profanity.load_censor_words() at initialization, to reduce the runtime for the first profanity.censor() call.
from better_profanity import profanity
if __name__ == "__main__":
profanity.load_censor_words()
text = "You p1ec3 of sHit."
censored_text = profanity.censor(text)
print(censored_text)
# You **** of ****.
All modified spellings of words in profanity_wordlist.txt will be generated. For example, the word handjob would be loaded into:
'handjob', 'handj*b', 'handj0b', 'handj@b', 'h@ndjob', 'h@ndj*b', 'h@ndj0b', 'h@ndj@b',
'h*ndjob', 'h*ndj*b', 'h*ndj0b', 'h*ndj@b', 'h4ndjob', 'h4ndj*b', 'h4ndj0b', 'h4ndj@b'
The full mapping of the library can be found in profanity.py.
By default, profanity replaces each swear words with 4 asterisks ****.
from better_profanity import profanity
if __name__ == "__main__":
text = "You p1ec3 of sHit."
censored_text = profanity.censor(text)
print(censored_text)
# You **** of ****.
The function .censor() also hide words separated not just by an empty space but also other dividers, such as _, , and .. Except for @, $, *, ", '.
from better_profanity import profanity
if __name__ == "__main__":
text = "...sh1t...hello_cat_fuck,,,,123"
censored_text = profanity.censor(text)
print(censored_text)
# "...****...hello_cat_****,,,,123"
4 instances of the character in second parameter in .censor() will be used to replace the swear words.
from better_profanity import profanity
if __name__ == "__main__":
text = "You p1ec3 of sHit."
censored_text = profanity.censor(text, '-')
print(censored_text)
# You ---- of ----.
Function .contains_profanity() return True if any words in the given string has a word existing in the wordlist.
from better_profanity import profanity
if __name__ == "__main__":
dirty_text = "That l3sbi4n did a very good H4ndjob."
profanity.contains_profanity(dirty_text)
# True
Function .load_censor_words() takes a List of strings as censored words.
The provided list will replace the default wordlist.
from better_profanity import profanity
if __name__ == "__main__":
custom_badwords = ['happy', 'jolly', 'merry']
profanity.load_censor_words(custom_badwords)
print(profanity.contains_profanity("Fuck you!"))
# Fuck you
print(profanity.contains_profanity("Have a merry day! :)"))
# Have a **** day! :)
No extra steps needed!
from better_profanity import profanity
if __name__ == "__main__":
bad_text = "Эффекти́вного противоя́дия от я́да фу́гу не существу́ет до сих пор"
profanity.load_censor_words(["противоя́дия"])
censored_text = profanity.censor(text)
print(censored_text)
# Эффекти́вного **** от я́да фу́гу не существу́ет до сих пор
- As the library compares each word by characters, the censor could easily be bypassed by adding any character(s) to the word:
profanity.censor('I just have sexx')
# returns 'I just have sexx'
profanity.censor('jerkk off')
# returns 'jerkk off'
- Any word in wordlist that have non-space separators cannot be recognised, such as
s & m, and therefore, it won't be filtered out. This problem was raised in #5.
$ python tests.py
- v0.4.0 - Add compatibility to all versions of Python 3.
- v0.3.4 - Add significantly more swear words.
- v0.3.3 - Fix incompatibility with Python 3.5.
- v0.3.2 - Fix a typo in documentation.
- v0.3.1 - Remove unused dependencies.
- v0.3.0 - Add support for Unicode characters (Categories: Ll, Lu, Mc and Mn) #2.
- v0.2.0 - Bug fix + faster censoring
- v0.1.0 - Initial release
Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.
This project is licensed under the MIT License - see the LICENSE.md file for details
- Andrew Grinevich - Add support for Unicode characters.
- Ben Friedland - For the inspiring package profanity.