Released: Oct 10, 2023

Mises à jour de v1.4.0

Fonctionnalités

• Performance gains in the .to_js_literal() method. Implementation seems to improve performance by 50 - 90%.
• Performance gains in the .to_json() method. Implementation seems to improve performance by 30 - 90%.
• Significantly simplified use of the .from_pandas() method to support:
  • Creation of multiple series from one DataFrame in one method call.
  • Creation of series without needing to specify a full property map.
  • Support for creating series by DataFrame row, rather than just by DataFrame column.
• Added the .from_pandas_in_rows() method to support creation of charts and series from simple two-dimensional DataFrames laid out in rows.
• Added one-shot chart creation and rendering from Series objects.
• Added one-shot chart creation using series and data/series_type keywords.
• Added .convert_to() convenience method to Series objects.
• Added CallbackFunction.from_python() method which converts a Python function to its JavaScript equivalent using generative AI, with support for both OpenAI and Anthropic.

Correctifs

• Fixed instability issues in Jupyter Notebooks, both when operating as a Notebook (outside of Jupyter Lab) and when saved to a static HTML file.