Problem
As soon as one example has more than one sub-example, e.g. here
https://matplotlib.org/stable/gallery/text_labels_and_annotations/font_family_rc.html#sphx-glr-gallery-text-labels-and-annotations-font-family-rc-py
it is inconvenient to iterate on that example with python files (*.py), because it's not straight forward to run a python file cell in sections.
Suggested improvement
A good format to support section wise code execution would be .ipynb.
What do you think about supporting .ipynb for creating docs examples?
I think that would make iterating over examples easier, and it would encourage contributors to create examples of shorter code blocks with multiple sub-examples. This could also be very useful in the tutorial section.
In my opinion, a bunch of small examples each showing only one feature is much better than one big example.
Matplotlib has currently many examples that have long and convoluted code (e.g. 1,2,3,4), and I think that might be the case because it's difficult to split examples into smaller chunks with the current infrastructure.
Problem
As soon as one example has more than one sub-example, e.g. here
https://matplotlib.org/stable/gallery/text_labels_and_annotations/font_family_rc.html#sphx-glr-gallery-text-labels-and-annotations-font-family-rc-py
it is inconvenient to iterate on that example with python files (
*.py), because it's not straight forward to run a python file cell in sections.Suggested improvement
A good format to support section wise code execution would be
.ipynb.What do you think about supporting
.ipynbfor creating docs examples?I think that would make iterating over examples easier, and it would encourage contributors to create examples of shorter code blocks with multiple sub-examples. This could also be very useful in the tutorial section.
In my opinion, a bunch of small examples each showing only one feature is much better than one big example.
Matplotlib has currently many examples that have long and convoluted code (e.g. 1,2,3,4), and I think that might be the case because it's difficult to split examples into smaller chunks with the current infrastructure.