2.3.1 — PyPI sdist fix

Fixes

• PyPI: LICENSE file is now included in the source distribution (sdist). This fixes installation on Python 3.10+ when no prebuilt wheel is available.
• PyPI: Release workflow now builds and uploads sdist automatically.

Added

• examples/python/ml_moderation.py — example of ML-based toxicity detection with badwords-py[ml].

v2.3.0 — ML Module & Benchmarks

Optional ML-based toxicity detection and benchmark updates.

Packages

• Python: pip install badwords-py
• Python (ML): pip install badwords-py[ml]
• Rust: cargo add badwords-core
• Node.js / Browser: npm install badwords-wasm @badwords/languages

Highlights

• badwords-py[ml] — optional ML toxicity predictor (XLM-RoBERTa ONNX)
• Model storage — GitHub Releases, cache ~/.cache/badwords/ml/, or BADWORDS_ML_PATH
• Benchmarks — make bench / make bench-compare (vs glin-profanity)
• Throughput — ~130–400 K/s (rule-based), ~150–220/s (ML)

ML

from badwords.ml import ToxicityPredictor
p = ToxicityPredictor() # downloads from GitHub Releases on first use
prob = p.predict("Текст") # 0.0–1.0

Assets

• badwords-ml-model.zip — ML model for badwords-py[ml] (download on first use)

v.2.2.0

v2.2.0 — Rust Core & Multi-Platform

Major release: core rewritten in Rust, bindings for Python, Rust, and JavaScript (WebAssembly).

Packages

Platform	Install
Python	pip install badwords-py
Rust	cargo add badwords-core
Node.js / Browser	npm install badwords-wasm @badwords/languages

Highlights

• badwords-core (crates.io) — Rust library
• badwords-py (PyPI) — Python bindings via PyO3
• badwords-wasm (npm) — WebAssembly for JS/TS
• @badwords/languages (npm) — 25 language packs
• GitHub Actions CI (tests, format, release)
• Docs: badwords.flacsy.dev

Breaking changes

• Project layout: badwords/ → python/badwords/, new rust/ crate structure

v2.1.0

BadWords 2.1.0

• Добавлен мощный текстовый процессор (TextProcessor) с расширенными возможностями:

  • Транслитерация между кириллицей и латиницей
  • Нормализация Unicode-символов
  • Обнаружение и обработка гомоглифов
  • Агрессивная нормализация текста
  • Частотный анализ символов
  • Многослойная фильтрация

• Добавлена поддержка Python 3.10 и выше