Simplify process to generate plotting documentation

[jenhagg]

Purpose

The plots used on the docs website require specific scenarios as well as an environment to run the code snippets. The goal here is to simplify that as much as possible, so we can recreate the content as needed, e.g. to take advantage of new functionality.

What the code is doing

What's involved:

• zip file with scenario data uploaded to blob storage
• docker-compose file for self contained environment
• notebook which generates the code used to run/save each code snippet

Testing

I ran the notebook and compared the results with currently checked in files. There are a few minor differences, but that is due to inconsistencies in the code snippets, which we can fix separately if needed.

Usage Example/Visuals

If github will render the notebook, that is probably the most useful place to look

Time estimate

20 min

[rouille]

I think we can use the following:

from bokeh.io import export_png

export_png(plot, filename="plot.png")

to automatically save the plot as png

[jenhagg]

Couple of issues.. we can get the plot object to pass in using curstate() if output_notebook() was called too, but the export also required pip install selenium, and after that I got an error that firefox isn't supported. At that point I gave up.

[rouille]

Should we add a paragraph in index.rst that would explain how to download the zip file and spin up the container so users can run the code snippets? Plots would not be saved but would display only.

[rouille]

Now that we have docs/run_snippets.ipynb, we probably don't need postreise/plot/demo/plot_tutorial.ipynb. Correct?

[jenhagg]

Now that we have docs/run_snippets.ipynb, we probably don't need postreise/plot/demo/plot_tutorial.ipynb. Correct?

Might not hurt to have both.. the tutorial is more readable, while run_snippets is easier to run and save repeatedly. Maybe we could delete it down the road if it turns out we don't use it.

[rouille]

Thanks