matplotlib documentation from python notebook?

tacaswell · August 31, 2015, 3:14am

A tad dry, but through, example of what the notebook backend can do is to
look at the UAT at

matplotlib/matplotlib/blob/main/lib/matplotlib/backends/web_backend/nbagg_uat.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imp import reload"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## UAT for NbAgg backend.\n",
    "\n",
    "The first line simply reloads matplotlib, uses the nbagg backend and then reloads the backend, just to ensure we have the latest modification to the backend code. Note: The underlying JavaScript will not be updated by this process, so a refresh of the browser after clearing the output and saving is necessary to clear everything fully."
   ]
  },

This file has been truncated. show original

Tom

···

On Mon, Jun 8, 2015 at 5:28 PM Jody Klymak <jklymak at uvic.ca> wrote:

Hi Ben,

My idea was to just make the notebook the ?example?. But nbconvert-ing
the notebook makes static images that then need to be checked into the
repository, and take space, so I wasn?t sure how desirable that was.

It would be fun to have the documentation script accept ipython notebooks
and run nbconvert on them. Being able to save state as you work through
examples is quite nice, versus creating five or six standalone *.py files
that then get run at build time. Of course you are adding a dependency to
anyone who has wants to build the docs.

Thanks, Jody

On 8 Jun 2015, at 13:35 PM, Benjamin Root <ben.root at ou.edu> wrote:

No, there isn't an accepted way to do that AFAIK. However, it doesn't seem
like it is all that far off. Our doc-build process will create the images
from the examples automatically, so you don't need to include the image
tag. It is sort of a way to make sure the examples work and that the image
matches the code correctly.

On Mon, Jun 8, 2015 at 1:06 PM, Jody Klymak <jklymak at uvic.ca> wrote:

Hi all,

If I want to contribute *.rst files to the matplotlib documentation, I
can see a few styles already contributed, at least one of which makes
extensive use of ipython (http://matplotlib.org/users/image_tutorial.html).
However, even it makes use of `.. sourcecode:: python` and `.. plot::`
blocks.

If I convert an ipython notebook to rst, it formats as: `.. code::
python` and instead of making plots it loads images:
`.. image:: MyExample_files/MyExample_1_0.png`

So, is there an acceptable way to directly make matplotlib documentation
directly from a notebook? I didn?t see anything, but wanted to check, as
that would by far be the easiest way to make a *.rst that had structured
text, code, and plots.

Thanks, Jody

--
Jody Klymak
Jody M. Klymak - UVic Ocean Physics

------------------------------------------------------------------------------
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at lists.sourceforge.net
matplotlib-users List Signup and Options

--
Jody Klymak
Jody M. Klymak - UVic Ocean Physics

------------------------------------------------------------------------------
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at lists.sourceforge.net
matplotlib-users List Signup and Options

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20150831/402d78cb/attachment.html>