[Feature] Text alignment

[cvanelteren]

Closes #753 for text alignment

I implemented a KD tree approach to determine the closest neighbor between text elements to prevent overlap. How this works is that by default it would keep track of texts and aligns them on the final draw. Nothing is registered unless you pass avoid_overlap=True (or flip rc['text.align']), so the solver never even imports on an ordinary figure. The naive approach would be to do this with O(n^2) pairwise comparison but that scales poorly and is not necessary.

A plot with 30 labeled texts with run in about 80ms compared to 50ms natively without any draws. This approach will scale not properly for large number of labelled elements (say > 500) but I don't think that will ever occur for plots as it will be unreadable anyway.

[codecov]

Codecov Report

❌ Patch coverage is 96.16909% with 29 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
ultraplot/textalign.py	91.50%	16 Missing and 10 partials ⚠️
ultraplot/tests/test_textalign.py	99.50%	1 Missing and 1 partial ⚠️
ultraplot/axes/base.py	97.82%	0 Missing and 1 partial ⚠️

📢 Thoughts on this report? Let us know!

[cvanelteren]

TODO: add tests