René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Native support for showing OOP-created figures","articleBody":"### Problem\r\n\r\nOften, I will create figures using matplotlib using the ``Figure`` class explicitly, because either I want to make a figure and repeatedly tweak it to see how it looks, or I am creating a lot of figures and don't need / want to deal with the pyplot interface. However, as far as I can tell, there is no native matplotlib way to show those figures since they weren't created with the pyplot interface.\r\n\r\nAn example of what I would like to do:\r\n```\r\nfig =. Figure()\r\nax = fig.subplots()\r\nax.plot([1, 2, 3, 4])\r\nfig.show() # This gives an error (below)\r\n```\r\n\r\nError:\r\n```\r\nTraceback (most recent call last):\r\n  File \"\u003cstdin\u003e\", line 1, in \u003cmodule\u003e\r\n  File \"/Users/jordan.e/opt/miniconda3/lib/python3.7/site-packages/matplotlib/figure.py\", line 408, in show\r\n    \"Figure.show works only for figures managed by pyplot, \"\r\nAttributeError: Figure.show works only for figures managed by pyplot, normally created by pyplot.figure()\r\n```\r\n\r\nThe reason I don't want to use ``plt.subplots()`` followed by ``plt.show()``, is, as mentioned above, because I often want to close and re-show the same figure. It seems really odd to me that matplotlib has a way of creating figures via a pure-OOP interface, but no easy way to display them. This is very frustrating and basically leads me to either repeatedly re-create the same plot with the pyplot interface instead of just tweaking the one I already have. I wrote a very hacky method for showing created figures based on a StackOverflow answer to this question which is below, and does not work reliably (the figure doubles in size every time I show it):\r\n\r\n```\r\ndef show_figure(fig):\r\n    \"\"\"\r\n    create a dummy figure and use its manager to display \"fig\"\r\n    \"\"\"\r\n\r\n    dummy = plt.figure()\r\n    new_manager = dummy.canvas.manager\r\n    new_manager.canvas.figure = fig\r\n    fig.set_canvas(new_manager.canvas)\r\n    plt.show()\r\n```\r\n\r\n### Proposed Solution\r\n\r\nEither support ``fig.show()`` directly for figures created using the OOP-interface, or provide a method like ``pyplot.show_figure()`` which takes as an argument an existing figure and prints it to the screen (doing the exact same thing that a ``plt.plot()`` followed by ``plt.show()`` would do)","author":{"url":"https://github.com/edmundsj","@type":"Person","name":"edmundsj"},"datePublished":"2021-04-14T01:44:30.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":13},"url":"https://github.com/19956/matplotlib/issues/19956"}