Gallery: show colour bar stealing space from multiple axes

[rcomer]

🚀 Pull Request

Description

It was embarrassingly recent that I realised you can get Matplotlib to automatically place a colour bar across multiple axes. Probably because I originally came to the Iris gallery to learn this sort of thing. This PR modifies one example to show such placement rather than manually placing a colour bar axes. Note that we also have the GloSea example showing manual colour bar placement.

I have two versions of this in the two commits. In the first commit, I created Cartopy GeoAxes up front so that I could pass the original array of axes to colorbar. I the second commit, I reverted that and found a different way to get hold of the GeoAxes that Iris auto-creates. I am in two minds which is the right version to keep: I think the first version is Better Object Oriented Code, but the second version better illustrates Iris' capabilities (and gotchas).

---

Consult Iris pull request check list

[rcomer]

Needs SciTools/iris-test-data#86

[codecov]

Codecov Report

All modified lines are covered by tests ✅

Comparison is base (58220d1) 89.41% compared to head (e9bc206) 89.41%.
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #5537   +/-   ##
=======================================
  Coverage   89.41%   89.41%           
=======================================
  Files          89       89           
  Lines       22539    22539           
  Branches     5387     5387           
=======================================
  Hits        20154    20154           
  Misses       1639     1639           
  Partials      746      746           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[rcomer]

I was totally expecting a failure because the reference image needs updating 😕

[trexfeathers]

I was totally expecting a failure because the reference image needs updating 😕

#5539

[trexfeathers]

Looks good! I'm happy that this works, and it's definitely more elegant / more idiomatic Matplotlib. I prefer the second approach - the one currently in the PR - as it is more typical Iris use with less to explain.

Got one thing I'd be keen for you to change in the What's New.

[rcomer]

I think I also need to update the test data version?

[trexfeathers]

I think I also need to update the test data version?

Yes please, thanks for the reminder

[rcomer]

Thanks @trexfeathers 😀