MAINT - Simplify repo and workflow

[trallard]

This PR introduces:

1. Remove abandoned POC
2. Restructure some bits of the repo, namely: move tools to the top of the repo, align READMEs with content
3. Tidy up the setup.cfg and pyproject.toml files
4. Refactor our GH action workflow to use the contents of ./testing/jupyterlab

Closes #3
Closes #9

[trallard]

This is almost done - except for the issue with a missing dependency, it seems (see https://github.com/Quansight-Labs/jupyter-a11y-testing/runs/8156261806?check_suite_focus=true)

@gabalafou, any insights on why this might be happening would be great 🙏🏽 - the dep seems to be in package.json, and yet it is missing?

Affecting:

1. CI (see actions log)
2. Local development - (following instructions in readme) throws the same error

code: 'MODULE_NOT_FOUND',
  path: '/home/runner/work/jupyter-a11y-testing/jupyter-a11y-testing/testing/jupyterlab/node_modules/expect-axe-playwright/package.json',
  requestPath: 'expect-axe-playwright'

[gabalafou]

any insights on why this might be happening

It was because of a difference in the way that Yarn (1.x) and NPM handle install. My fork added a prepare script to package.json. When you run npm install (in testing/jupyterlab) and it comes across a package (gabalafou/expect-axe-playwright) whose package.json has a prepare script, it downloads both the dependencies and dev dependencies of that included package, which is necessary in order to build the package from source.

So basically, when you run npm install and you do ls node_modules/expect-axe-playwright, you'll see a lib/ directory, but when you do yarn install it doesn't properly execute the prepare script and so there's no lib/ directory, which is why the import from expect-axe-playwright was failing.

I don't know why there's this difference. It's stupid and frustrating, and I think now in 2022, maybe it's no longer beneficial to use Yarn over NPM, so I replaced yarn everywhere with npm.

[trallard]

This should be ready for a final review and merge

[trallard]

Testing artefacts: https://github.com/Quansight-Labs/jupyter-a11y-testing/actions/runs/3061380280

[steff456]

I left a couple of minor comments, this PR is looking really good!