Plot gallery

[sr-murthy]

• Add kaleido to package deps + orca to docs deps
• Add a plot gallery in the Sphinx docs - currently covers bar charts, bar line charts, descriptive tables, dual stack pyramid plots, frequency charts, Sankey diagrams, sunburst plots, upset plots (more will be added over time).

Visualisation page (RTD build version)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Files with missing lines	Coverage Δ
isaricanalytics/analytics.py	11.33% <ø> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[read-the-docs-community]

Documentation build overview

📚 IsaricAnalytics | 🛠️ Build #34172387 | 📁 Comparing 0222549 against latest (e2be39f)

  🔍 Preview build  

27 files changed · + 12 added · ± 13 modified · - 2 deleted

+ Added

• sg_execution_times.html
• sources/visualisation/bar.html
• sources/visualisation/bar_line.html
• sources/visualisation/dual_stack_pyramid.html
• sources/visualisation/frequency.html
• sources/visualisation/index.html
• sources/visualisation/sankey.html
• sources/visualisation/sg_execution_times.html
• sources/visualisation/sunburst.html
• sources/visualisation/table.html
• and 2 more...

± Modified

• genindex.html
• index.html
• _modules/index.html
• sources/analytics.html
• sources/api-reference.html
• sources/contributing.html
• sources/redcap-data.html
• _modules/isaricanalytics/analytics.html
• _modules/isaricanalytics/visualisation.html
• sources/isaricanalytics/analytics.html
• and 3 more...

- Deleted

• sources/visualisation.html
• _modules/scipy/stats/_mannwhitneyu.html

[alasdairwilson]

This is fantastic! I had fun messing with the sankey plot in the rendered docs.

There were some minor typos (there might be more I didn't see) and I would reccomend storing the data used, especially the big csv tables as sample data or test data rather than.

The only other thing I'd add is make absolute sure this is squashed on merge since there are some very large files that were subsequently removed, so would be a shame to not prevent them from being added to the git repo forever.

[sr-murthy]

Fine, the next PR update - probably early next week - will address these.

P. S. We can certainly store the figure CSVs as static files, perhaps in test data, and load them in the figure scripts, but to keep it simple we can keep the hard-coded RST tables that appear as comments (required by Sphinx gallery). If you want to make the RST comments dynamic you'd have to dynamically write the figure script files (examples/*.py), and that would involve more work that is probably worth at this point.

There is a PR setting for squash on merge, and the large files to which you refer were added then removed in earlier commits. So, squashing will not result in these files being added back.

[alasdairwilson]

There is a PR setting for squash on merge, and the large files to which you refer were added then removed in earlier commits. So, squashing will not result in these files being added back.

You misunderstand, I wasn't asking them to be removed from the current state of main, I wanted to keep them out of the history, as if this was not squashed they would forever live in reachable history of the repo as blobs.

But yes squash on merge will prevent that, and this is what I was asking to be done.

[sr-murthy]

You misunderstand, I wasn't asking them to be removed from the current state of main

I do understand - I removed the files myself here - but squash on merge is already enabled (screenshot below from the PR ruleset):

This is one of the reasons why we're enforcing squash merging (with rebase) in all the key repos - you get a much smaller, more linear, cleaner Git history, which makes Git log easier to read for a start.

[sr-murthy]

@alasdairwilson As you may have noted, the refactored plot gallery that now depends on Sphinx gallery only works because I'm using this Plotly fork to deal with a Sphinx gallery <-> Plotly problem I reported here. The solution is being reviewed in this Plotly PR.

We shouldn't obviously merge this PR at least until that Plotly PR is merged, and a new Plotly release, which we will need to pin. This will almost certainly also trigger a Sphinx gallery release, so we should wait for that too.

[sr-murthy]

@alasdairwilson Please review again. Fixes made + two more example scripts added for bar charts, and bar-line charts. Also, all static CSV data defined in the scripts have been moved to CSVs in a new subfolder (examples/csv) organised by figure.

It is certainly possible to also have dynamic generation of the RST tables of the synthetic datasets appearing in the plot pages - I think that seemed to be one of your suggestions / comments in the last review - but that would require additional changes that could be done at a later point.