News and Changelog

• 0.6.0 (26/05/22)

  • Fixed bug where the prediction visualizer repeats text when there is overlapping span. and intensifiers were considered outside of the sentences boundaries. Adressing #57

  • To use the new visualizers asent now required spacy version >= 3.3.0. However, can be run with spacy>=3.0.0.

  • Updates to documentation

• 0.5.3 (26/05/22)

  • Fixed bug where negation and intensifiers were considered outside of the sentences boundaries. Adressing 58.

  • Improvements to the documentation. Thanks to @tomaarsen for the pull request.

• 0.4.2 (28/05/22)

  • Added new Danish Dictionary from AFINN

  • Asent now support 81 languages using the autogenerated lexicons from Chen, Y., & Skiena, S. (2014).

  • Fixed minor bugs related to the Visualizer

  • Fixed spaCy pipelines used in text to allow for more consistent tests going forward.

  • Added pre-commit hooks to ensure future code quality.

• 0.1.3 (02/01/22)

  • 96% Coverage on pytest 🎉

  • New Visualizer for visualizing the “analysis” of the model, using spaCy dependency Visualizer

  • New CI for Tutorials to ensure that they always run

• 0.1.0 (26/12/21)

  • Major updates to the documentation along with two new tutorials 📖

  • Cool new infographics 😎

  • Changes:

    • make_token_polarity_getter no longer takes the is_negation_getter, but the is_negated_getter to avoid it setting the is_negated setting itself, potentially overwriting custom changes.

  • bug-fixes 🪲:

    • Lexicons is now properly installed with the package

    • Empty string no longer causes a bug

    • All extentions are now overwritten when force=True

    • And many more

• 0.0.3 (05/12/21)

  • First version of asent launches

    • with four different languages supported, some of which have never had rule-based sentiment models 🎉

    • Tutorials allowing you to get started using and customizing your sentiment pipeline 🔧

    • Visualizers for examining the predictions of the model 🌟

    • An extensive docuementation 📖