RF: Rely on stdlib implementations for tmpdirs tools

[effigies]

I noticed that we have almost an identical implementation of TemporaryDirectory as the stdlib (https://docs.python.org/3/library/tempfile.html#tempfile.TemporaryDirectory), so I scheduled it for deprecation and removal.

With the coming chdir context manager, the remaining tools can be made convenience wrappers on stdlib 1-or-2 liners, so I went ahead and did it. IMO it's a bit cleaner to read a single function with the setup-yield-teardown structure than a two-or-three method class.

[codecov]

Codecov Report

Base: 92.38% // Head: 92.40% // Increases project coverage by +0.01% 🎉

Coverage data is based on head (27b757b) compared to base (9000943).
Patch coverage: 100.00% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1172      +/-   ##
==========================================
+ Coverage   92.38%   92.40%   +0.01%     
==========================================
  Files          98       98              
  Lines       12257    12248       -9     
  Branches     2520     2524       +4     
==========================================
- Hits        11324    11318       -6     
+ Misses        612      611       -1     
+ Partials      321      319       -2     

Impacted Files	Coverage Δ
nibabel/tmpdirs.py	100.00% <100.00%> (+7.89%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[ZviBaratz]

LGTM.
If you'd like, I think using pathlib instead of os for everything besides chdir would be preferable.

[effigies]

From an API perspective, it's good to accept either str or Path, but maintaining existing return types is important. These will accept either input, but have been emitting strings. We would need to raise a FutureWarning to switch to returning Paths, which hardly seems worth it when people can already wrap with Path() if they want it.

Without that change, using the pathlib API internally seems like it would make this more complicated, not less.