Converting between ipython notebooks and sphinx docs
See the nb2plots documentation for more information.
What it does
Nb2plots assumes that the ReST document will become the source for your Sphinx web pages, but also for future versions of the notebook. The notebook may serve as a draft for the polished ReST page, and an output format from the Sphinx build. Why? Read on.
Why convert Jupyter notebooks to Sphinx?
Jupyter notebooks are just what the doctor ordered when hacking up a quick tutorial, or preparing a software demo. The problems start when you want to do not-trivial edits to the notebooks, or you need features that notebooks don’t have, such as flexible cross-referencing, extensible markup, and so on. Notebooks are also painful to use with version control. These times make you wish your notebook was in a standard extensible text format, such as ReST.
You could convert your notebook to ReST using the standard nbconvert command, but this gives rather ugly ReST, and you lose all the nice code execution and figure generation that the notebook is good at.
Enter Nb2plots. The nb2plots command converts notebooks to specially formatted ReST pages. Use with:
nb2plots notebook.ipynb > with_plots.rst
Nb2plots converts your notebook to not-very-ugly ReST, where the code cells become nbplot directives in your ReST file.
Specifically, a notebook code cell like this:
a = 1
becomes (in the ReST document):
.. nbplot:: >>> a = 1
The nbplot directives run the contained code when Sphinx builds your ReST files, and embed the results of any plots that your code makes. Actually, nbplot is an extended and edited version of the matplotlib plot directive. Building your pages runs all the code and regenerates the figures, and you get much of the reproducible goodness of the notebook experience.
You can also run the standard Sphinx doctest extension over your pages to check the doctest output of the code cells.
The ReST version of your notebook has many advantages - it is easier to edit in your favorite text editor, and you can extend and configure the execution and display of the code in several different ways. For example, you can hide some code cells (Nbplot directives) if the code is not interesting to your point, but you still want the generated figure. You can configure your Nbplot directives to run different code for different configurations. For these options, see nbplots documentation. But - what do you lose, when going from a notebook to a Nb2plots ReST document?
I want notebooks and .py files too
You may also want a version of your document that your users can execute. Perhaps the page build is generating some tricky errors or warnings, and you want to experiment with the code in the page interactively. Perhaps your users are used to notebooks, and prefer the code in that format.
Nb2plots also contains Sphinx extensions that cause the Sphinx build to generate Python code files and Jupyter notebooks from the ReST source. When you add the Nb2plots ReST directive code-links to your ReST page, it will cause the Sphinx build to create a Python code file and notebook versions of your page, and adds download links to these versions:
See code-links documentation for details.
Show me what it looks like
For a very simple example, see worked example.
For a moderate-sized teaching site that makes extensive use of Nb2plots, see https://matthew-brett.github.com/teaching.
pip install nb2plots
You will need Pandoc installed and available as the pandoc command.
To install Pandoc on OSX, we recommend homebrew:
brew install pandoc
Add the following to your Sphinx conf.py file:
extensions = ["nb2plots"]
See nbplots documentation for the various conf.py configuration settings.
Released under the BSD two-clause license - see the file LICENSE in the source distribution.
travis-ci kindly tests the code automatically under Python versions 2.7, and 3.3 through 3.5.
The latest released version is at https://pypi.python.org/pypi/nb2plots
Install the pytest testing framework, the mock package, and the scripttester package.
pip install pytest mock scripttester
Run the tests with:
py.test --pyargs nb2plots
Please put up issues on the nb2plots issue tracker.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size pymor-nb2plots-0.7.tar.gz (122.5 kB)||File type Source||Python version None||Upload date||Hashes View|