Run checks locally

[centreboard]

The PR checks run linting, type checks and tests.

wheatley/.github/workflows/main.yml

Lines 24 to 33 in 801e3d1

	- name: Run linter
	run: python3 -m pylint wheatley/**/*.py
	- name: Run typechecker
	run: python3 -m mypy wheatley --pretty --disallow-incomplete-defs --disallow-untyped-defs
	- name: Run unit tests
	run: python3 -m pytest
	- name: Run integration tests
	run: ./doctests
	- name: Run fuzzer
	run: ./fuzz

mypy and pylint are in the requirements, we just need a way of running these commands locally on Linux or Windows machines.
To cope with venvs, perhaps a python script that could be invoked? Looks like mypy and pylint both can be called from python
https://github.com/python/mypy/blob/master/docs/source/extending_mypy.rst
http://pylint.pycqa.org/en/latest/user_guide/run.html

[kneasle]

Running from python sounds like a good plan. Actually, what we could probably do is to make a python script which reads .github/workflows/main.yml and executes the shell commands. This way, there's no possibility for divergence between the two checking methods and we can keep the workflow split up into a separate job per check (which makes it nice and easy to find out what's failed).

[centreboard]

To combine them it would be better to have the github workflow call the python script

[kneasle]

True, but it would make the PR check into one monolithic job, which would make detecting what fails very pesty. At the moment, it's really easy to see which job fails:

[centreboard]

I was imagining a script which ran all checks by default, or add an argument to run only specified

- name: Run linter 
  run: python3 ./runchecks.py --linter
- name: Run typechecker 
  run: python3 ./runchecks.py --types
- name: Run unit tests 
  run: python3 ./runchecks.py --unittests

[kneasle]

Oooh - OK yes that would make much more sense.