Toxic Triggers

Toxic Triggers is a Python package that provides access to a dataset of text triggers that have been categorized as offensive or harmless to minorities. This package can be useful for researchers and developers who want to understand how offensive language can affect marginalized groups and how it can be detected and prevented.

Features

• Access to a dataset of text triggers categorized as offensive or harmless to minorities
• Provides easy access to subsets of the dataset based on categorization
• Built with Pandas for easy data manipulation
• Analyze the text for the presence of toxic triggers

Installation

To install the package, you can use pip install toxicTrig

Usage

Here's an example of how to use the package:

from toxicTrig import toxicTrig
import pandas as pd #type: ignore
import time

df = pd.read_csv("example_data/ND_founta_trn_dial_pAPI.csv")
text = df["tweet"].tolist() # A list of strings as the input

tt = toxicTrig()
start_time = time.time()
tt.text_analysis(text, batch_size=100)
end_time = time.time()

time_taken = round(end_time - start_time, 2)
print("Total time taken: ", time_taken, " seconds")

# Count tokens in text
total_tokens = sum([len(t.split()) for t in text])
print("Total tokens in text: ", total_tokens)

Contribution

Install dev options

pip install -e ".[dev]"
mypy --install-types --non-interactive toxicTrig
pip install pre-commit
pre-commit install

New branch for each feature

git checkout -b feature/feature-name and PR to main branch.

Before committing

Run pytest to make sure all tests pass (this will ensure dynamic typing passed with beartype) and mypy --strict . to check static typing. (You can also run pre-commit run --all-files to run all checks)

Citation

If you use this package in your research, please cite the following paper:

@inproceedings{zhou-etal-2020-debiasing,
  title = {Challenges in Automated Debiasing for Toxic Language Detection},
  author = {Zhou, Xuhui and Sap, Maarten and Swayamdipta, Swabha and Choi, Yejin and Smith, Noah A.},
  booktitle = {EACL},
  year = {2021},
}

License

This package is released under the MIT License. See the LICENSE file for more details.

Credits

This package was developed by Xuhui Zhou and Maarten Sap.