RFC/DEPR: deprecate yt.funcs.levenshtein_distance in favor of difflib.get_close_matches

[neutrinoceros]

PR Summary

I just learned about difflib.get_close_matches, and it seems like a nice way to remove a custom function of ours.

PR Checklist

• New features are documented, with docstrings and narrative docs
• Adds a test for any bugs fixed. Adds tests for new features.

[cphyc]

We are loosing some relevant suggestions (when there are more than 3 possible matches, try e.g. querying ad["index", "d"] (it should suggest x,y,z,dx,dy,dz at least. With the current PR, it only suggests the last three.

With the changes, I get more irrelevant (but also more relevant ones) suggestions:

ds = yt.load_sample("output_00080")
ad = ds.all_data()

ad["gas", "densty"]

Previously:

YTFieldNotFound: Could not find field ('gas', 'densty') in info_00080.
Did you mean:
	('gas', 'density')

This PR:

YTFieldNotFound: Could not find field ('gas', 'densty') in info_00080.
Did you mean:
	('gas', 'density')
	('gas', 'dy')
	('ramses', 'Density')

[neutrinoceros]

rebased to resolve merge conflicts

[chrishavlin]

@cphyc does your prior approval on this still stand? I think it was dismissed just from a merge.

I'll try to give this a second review soon.