Add autopep8 git hook

[brryan]

Background

• flake8 is effective for enforcing python syntax, but it doesn't apply inline updates.
• autopep8 does provide inline fixes as an option

Purpose of Pull Request

• Add autopep8 git hook to apply inline formatting fixes to modified python code prior to calling flake8.
• Fixes re-git issue #18

Description of changes

• Add pre-commit-autopep8 script and update pre-commit and install-hooks.sh to Draco/environment/git.

Status

• Reference:
  • Pre-Merge Code Review
  • CDash Status
• Checks
  • Travis/Appveyor CI checks pass
  • Code coverage does not decrease
  • Valgrind test passes
  • Toss3 checks pass
  • Trinitite checks pass
  • Code reviewed/approved, sufficient DbC checks, testing, documentation

[brryan]

@KineticTheory Here you go. I tested this with some simple garbled formatting in a python file and it worked for me as a commit hook.

I'm worried that at some point autopep8 will make a change that flake8 flags as an error, which will make committing with active hooks impossible to get past the CI... These two formatters should be following the same standards, but I've already encountered scenarios where autopep8 doesn't fix things that flake8 complains about. Maybe their options can be exactly synchronized? Another option would be to run flake8 on each file and record the errors, then run autopep8 to make changes, and then run flake8 again, checking that no new errors were introduced by autopep8. A bit involved...

[KineticTheory]

Do we still need pre-commit-flake8 if we adopt pre-commit-autopep8?

[brryan]

If we don't mind switching to autopep8 from flake8, that would solve the above consistency issue I brought up. I do like that autopep8 will automatically apply formatting changes, but I don't know if there are reasons for us to still prefer flake8 at the CI level

[KineticTheory]

The superlinter CI check that we use will run flake8. I want to keep that CI because it has a very low maintenance cost. It sounds like we need to keep both for now.