Initialise a Jupyter notebook with useful extensions and reasonable defaults.
%load_ext jupyter_interactive
- Enable autoimport of missing modules.
- Enable the
%autoreload
keyword to reload all modules in the current session. - Enable the
%%R
and%%coconut
magic to execute specific cells using R and Coconut. - Configure plot output formatting:
- Display Matplotlib plots as interactive, resizable widgets (using
ipympl
). - Display Matplotlib plots at a higher resolution.
- Zoom in and out of Plotly figures using the mouse scroll wheel and/or two-finger scrolling.
- Download Plotly figures at their currently rendered size.
- Ensure that dragging Plotly figures results in panning (rather than zooming).
- Display Matplotlib plots as interactive, resizable widgets (using
- Configure the output formatting of dataframes:
- Display dataframes as interactive HTML tables (using
itables
). - Increase the table size limit to 8 MiB.
- Adjust default colours for better contrast.
- Display cell contents in a monospace font.
- Extend the range of available table lengths.
- Make Polars dataframes output Markdown when printed.
- Show all dataframe columns, even if there are many of them.
- Display dataframes as interactive HTML tables (using
- Try to make results reproducible by setting the seed for the built-in
random
module, NumPy, and, if available, PyTorch and Tensorflow.