Alt Profanity Check

1.8.0 · active · verified Thu Apr 16

Alt Profanity Check is a fast, robust Python library designed to detect offensive language in strings, serving as a drop-in replacement for the unmaintained `profanity-check` library. It utilizes a linear Support Vector Machine (SVM) model trained on a large dataset of human-labeled text. The library is actively maintained, with frequent updates that align with its primary dependency, `scikit-learn`. The current version is 1.8.0, supporting Python 3.11 and newer.

Common errors

```
ModuleNotFoundError: No module named 'profanity_check'
```
cause The `alt-profanity-check` library is not installed, or Python cannot find the installed package.

fix Ensure `alt-profanity-check` is installed in your active Python environment using `pip install alt-profanity-check`.
```
AttributeError: module 'profanity_check' has no attribute 'predict'
```
cause This typically occurs if `alt-profanity-check` failed to install correctly or its internal data files (`model.joblib`, `vectorizer.joblib`) are missing/corrupted, often due to dependency conflicts with `scikit-learn` or `joblib`.

fix Reinstall `alt-profanity-check` in a fresh virtual environment to ensure all dependencies are correctly resolved: `python -m venv .venv && source .venv/bin/activate && pip install alt-profanity-check`.
```
RuntimeError: Python version 3.10.x is not supported by scikit-learn 1.8.0
```
cause You are trying to use `alt-profanity-check` v1.8.0 (or newer) with an unsupported Python version. `alt-profanity-check` v1.8.0 requires Python >= 3.11.

fix Upgrade your Python environment to 3.11 or newer, or downgrade `alt-profanity-check` to a version compatible with your current Python (e.g., `pip install alt-profanity-check==1.7.2` for Python 3.10).

Warnings

breaking Python version requirements have changed significantly across minor versions of `alt-profanity-check` due to its tight coupling with `scikit-learn` dependencies. For `alt-profanity-check` v1.8.0, Python >= 3.11 is strictly required.
Fix: Ensure your Python environment meets the `python_requires` specification for the `alt-profanity-check` version you are installing. For v1.8.0, use Python 3.11 or newer. For older Python versions, pin `alt-profanity-check` to an earlier compatible version (e.g., v1.7.x for Python >=3.10, v1.4.x for Python >=3.8).
gotcha The library is installed via `pip install alt-profanity-check`, but imports use the original module name `profanity_check` (e.g., `from profanity_check import predict`). This is by design, as it's a drop-in replacement.
Fix: Always use `from profanity_check import ...` in your code, even though the installed package name is `alt-profanity-check`.
breaking Major underlying dependency (`scikit-learn`) version upgrades can introduce breaking changes. `alt-profanity-check` explicitly pins `scikit-learn` versions per release. Mismatched `scikit-learn` versions between what `alt-profanity-check` expects and what is installed in your environment can lead to runtime errors or incorrect predictions.
Fix: Always install `alt-profanity-check` in a clean virtual environment to ensure its `scikit-learn` and `joblib` dependencies are correctly installed and matched. Avoid manually upgrading `scikit-learn` independently of `alt-profanity-check`.

Install

pip install alt-profanity-check Install with pip

Imports

predict
wrong:
```
from alt_profanity_check import predict
```
correct:
```
from profanity_check import predict
```
Despite the package name being 'alt-profanity-check', it is a drop-in replacement for 'profanity-check' and maintains the original internal module structure for imports.
predict_prob
wrong:
```
from alt_profanity_check import predict_prob
```
correct:
```
from profanity_check import predict_prob
```
The library is designed as a drop-in replacement for 'profanity-check', hence imports are from 'profanity_check' internally.

Quickstart

This example demonstrates how to import and use the `predict` and `predict_prob` functions to check a list of strings for offensive language and get their respective profanity probabilities.

from profanity_check import predict, predict_prob

# Predict profanity (returns 1 for profane, 0 for clean)
texts = ['hello world', 'go to hell', 'this is a bad word']
predictions = predict(texts)
print(f"Predictions: {predictions}")

# Predict probability of profanity
probabilities = predict_prob(texts)
print(f"Probabilities: {probabilities}")

view raw JSON →