Skip to content

v0.1.2

Choose a tag to compare

@taylorwilsdon taylorwilsdon released this 24 Feb 14:52
· 91 commits to main since this release

🚀 reddacted v0.1.2 Release Notes

We're excited to announce reddacted v0.1.2, which includes various improvements, bug fixes, and enhancements to the AI-powered Reddit privacy suite. This release focuses on refining the installation process, improving test coverage, updating dependencies, and enhancing error handling.

🔥 What's New

✅ Improved GitHub Actions Workflows

  • Updated lint, unit, and smoke test workflows to support both main and master branches.
  • Upgraded Python version in workflows to 3.12.
  • Improved dependency caching for faster CI runs.

📖 Documentation & Readability Enhancements

  • README.md has been significantly improved:
    • Clearer installation instructions.
    • More structured FAQ section.
    • Updated CLI usage examples.
    • Improved formatting for better readability.

🛠️ Code & Dependency Updates

  • Migrate away from /api/tags to /v1/models for model lookup, refactor exception handling for cleaner output
  • fix: issues/6 by @taylorwilsdon in #8
  • Refactored pyproject.toml:
    • Updated dependencies to latest stable versions.
    • Added httpx==0.27.2 for improved HTTP handling.
    • Improved test dependencies (pytest, black, mypy, ruff).
  • Removed setup.py in favor of pyproject.toml for modern package management.

🧪 Testing & Stability Improvements

  • Refactored tests/test_llm.py:
    • Migrated from unittest to pytest for better async test handling.
    • Added new test cases for LLM-based PII detection.
    • Improved error handling tests (e.g., invalid API keys, rate limits, malformed responses).
    • Introduced batch processing tests for concurrent text analysis.
  • Enhanced exception handling in reddacted/utils/exceptions.py:
    • Improved traceback formatting.
    • More robust error messages with better debugging context.

🏗️ Other Improvements

  • Better local LLM support:
    • Fixed incorrect API endpoint references in reddacted/cli.py.
    • Improved model selection logic.
  • Automatic NLTK data download:
    • vader_lexicon is now downloaded automatically if missing.
  • More robust CLI error handling:
    • Improved error messages when connecting to local LLMs.

📦 Installation & Upgrade

To install or upgrade to the latest version:

pip install --upgrade reddacted

Full Changelog: v0.1.1...v0.1.2