PR Summary

Closes #16153, and does some more.

Three fixes/improvements to contour labeling, grouped together to minimize baseline images changes. Code-wise I think this is ready for review, but marking as draft PR just so that I can still have a look at whether we can combine/delete some of the baseline images. I also took advantage of the baseline regen to merge two tests (test_contour_manual_labels and test_contour_label_size_color).

Commit 1: Improve estimation of label text width, as explained in #16153.

Commit 2: Remove

which shifts the labels a bit but has no bad effects (there is no division in print_label and division by zero warning is expressly suppressed in locate_label.

Commit 3: Switch the default formatter to ScalarFormatter and use format_ticks to format all ticks at once, allowing the reuse of auto-precision choice by ScalarFormatter. Ultimately we could even use someting like #12473 (reusing colorbar logic to format values) but this shouldn't change any baseline images as none of them uses log scale (so they would still all use ScalarFormatter).

Commit 4: Update baseline images all at once. See the precision change and unicode minus sign in e.g. test_contour/contour_test_label_transforms.png and the tighter fit of labels in e.g. test_patheffects/collection.png.

I'm not sure why CI is failing even though I updated the baseline images locally, though. Figured out the test failure.

PR Checklist

• Has Pytest style unit tests
• Code is Flake 8 compliant
• New features are documented, with examples if plot related
• Documentation is sphinx and numpydoc compliant
• Added an entry to doc/users/next_whats_new/ if major new feature (follow instructions in README.rst there)
• Documented in doc/api/api_changes.rst if API changed in a backward-incompatible way