PEP8 and black

[bryanwweber]

• Variable names conform to PEP 8 style
• Run the Black code formatter on the notebooks
• Show the currently running Cantera version in all the Notebooks
• Switch all plots to %matplotlib inline because:
  1. JupyterLab does not support %matplotlib notebook and I found the recommended replacement %matplotlib widgets to be unreliable in terms of showing up after I closed the notebook
  2. This should give a consistent look to the output on the website

[bryanwweber]

This is simply ridiculous! It must be from removing all the %matplotlib notebook JS garbage from the output.

[ischoegl]

Indeed impressive 😄. You could save even more with non-executed notebooks … which would also make changes tractable; otherwise I’m not sure that it’s reasonable to go through this line-by-line 😆

[ischoegl]

But out of curiosity: would SVG images be more economical to save than png?

[bryanwweber]

You could save even more with non-executed notebooks … which would also make changes tractable; otherwise I’m not sure that it’s reasonable to go through this line-by-line 😆

Yah, reviewing these PRs is super hard. I don't think this really needs a review, unless we want to bikeshed about Black's formatting style. I'm posting it here mainly to have a place to discuss anything that might come up, and so there's some record of this happening aside from the commit messages, even if I hit the big green button myself 😄

I executed everything so I know it all works, and so that it would show up in the static view on GitHub itself. The website build runs the Notebooks (I think), so that's fine.

But out of curiosity: would SVG images be more economical to save than png?

I'm not sure, possibly. But I don't know how to generate those automatically with a magic like %matplotlib inline 😄

[ischoegl]

I believe there’s this magic … not sure though as I just ‘borrowed’ this from some recent notebook from our group

%config InlineBackend.figure_formats = ['svg']

[bryanwweber]

FYI, you can comment/review the Notebooks in this PR over here: https://app.reviewnb.com/Cantera/cantera-jupyter/pull/34/

[ischoegl]

I mainly had a cursory look and as this is mostly formatting this looks good to me. (Fwiw, if we want to update content of individual notebooks, this is probably not the place.)

So I overall have no issues with adopting as is (although it would be great if the issues I pointed out in Cantera/cantera-website#171 could be addressed. Notably, the LaTeX formatting glitch in equations_of_state.ipynb doesn't show up here, which is odd).

[ischoegl]

@bryanwweber ... I cannot reproduce the LaTeX artifacts, so from my perspective it's ok to

[bryanwweber]

@bryanwweber ... I cannot reproduce the LaTeX artifacts, so from my perspective it's ok to

👍 I'm planning to add those few fixes plus the fix for SolutionArray extra elements

[speth]

@bryanwweber - Unless whatever other changes you want to make are as pervasive as these, can we handle them in a separate PR? Otherwise, waiting on this makes any other PRs against this repo impossible without running into merge conflicts. Also, any additional changes made in this repository will be difficult to review for the same reason.

[bryanwweber]

@speth sure, the other changes can go in a new PR. This is ready for review then, or merging if it's good.

[speth]

Having to switch to %matplotlib inline is a little disappointing -- I think the figures end up looking a lot worse than they do with the notebook or widget backends, but I'm also seeing the issue with restoring plots in a saved notebook. Two things that would improve things are

• setting plt.rcParams['figure.dpi'] to a higher value than the default of 72 -- maybe 96 or even 120 since the figures are pretty tiny by default.
• Using @ischoegl's suggestion of %config InlineBackend.figure_formats = ['svg']. On high-DPI displays, this seems to result in much crisper figures than the PNG images that end up looking fuzzy after they get blown up.

For once I'm OK with running black on the repository -- I think these notebooks have been harder to keep consistent formatting in, and I'd say the result here is clearly an improvement.

[bryanwweber]

Thanks for the review @speth! I'm out of town for the weekend, so if you want to push one of those options to this branch, please feel free. Otherwise, I'll pick it up on Monday.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[bryanwweber]

@speth I updated based on your suggestions! Figures look much better now.

[ischoegl]

Too bad that figures don't render on ReviewNB (I checked only one of the files though) ... per discussion elsewhere, I hope that display works as intended once things are uploaded to GH / embedded in the website.

[speth]

I just tested this with a local checkout of both the cantera-jupyter and cantera-website repositories, and all of the figures look great.