- Add assignee benchmarks to manifest.in
- Bug fixes
- Compatibility with er-evaluation>=2.0.0
- FIX: Unavailable function in all.
- Add inventor_summary_trend_plot() and inventor_estimates_trend_plot() functions.
- Update setup.py with find_packages()
- BREAKING: Major refactoring of the package. Core evaluation functionality has been moved to the er-evaluation package.
- FIX: Replace sklearn dependency with scikit-learn. Add scipy dependency.
- Added assignee benchmark datasets
- Added patents links and preview tooltips to automated reports.
- Added estimates table and plots.
- FIX: Install from Github now includes necessary csv files.
- Added script for processing hand-disambiguated inventor benchmark.
- Removed Harvard inventors benchmark (low data quality).
- BREAKING: change "mention-id" to "mention_id" and "unique-id" to "unique_id" in datasets.
- Added inventor names to raw csv benchmarkd datasets (useful for validation and debugging).
- Added cluster precision and cluster recall estimators (for cluster block sampling).
- Added additional inventor benchmark datasets.
- Added license file (GPL3)
- Initial release