René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Implement wrapper functions with decorators instead of __getattribute__","articleBody":"So far, I have been implementing all plot method wrappers by applying them in `__getattribute__` on the axes classes. This is pretty \"hacky\" and unexpected for new contributors. On the \"[Axes and related wrappers](https://proplot.readthedocs.io/en/latest/axes.html)\" page of the documentation, I defend this choice as reducing the source code size. Originally, I also avoided this because the [automodapi sphinx extension](https://proplot.readthedocs.io/en/latest/axes.html) used to generate ProPlot documentation inherits docstrings from identically-named methods on parent classes. In my [fork of automodapi](https://proplot.readthedocs.io/en/latest/axes.html), this is no longer the case -- if the docstring is empty (i.e. `func.__doc__` is empty), automodapi skips that method. I prefer to document ProPlot enhancements of matplotlib axes methods separately, so users calling e.g. `help(ax.contourf)` get the native matplotlib docstring.\r\n\r\nIn the future, I'd like to implement all wrappers using decorators, which is more verbose but probably falls under the [Simple is better than complex](https://www.python.org/dev/peps/pep-0020/) category. There are also several places in the wrapper functions where I apply function-specific behavior -- e.g. in `cmap_wrapper`, various `if func.__name__ == 'contour'` statements. With decorators, it will be easier to implement method-specific behavior and easier to debug issues.","author":{"url":"https://github.com/lukelbd","@type":"Person","name":"lukelbd"},"datePublished":"2019-09-11T17:52:07.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":1},"url":"https://github.com/37/proplot/issues/37"}