We're not using these so I'm not sure we need them here?

I don't think this does anything?
And by removing it we can avoid having to call the Github API except
when seeding data which makes running the tests a bit easier

This adds the packages, config and a github action to run the tests

This adds a seeding job.
Using live data is hard to assert against and means we have to run the
seeding job which takes time and uses rate-limit tokens.
Instead we copy across some test data we can use. This means the data
will always be consistent and we can test against it more easily.
We can't do this in a setup job [as Playwright
recommends](https://playwright.dev/docs/test-global-setup-teardown)
because the app depends on the JSON existing before it can start.
Instead we have to copy and paste the seed data across before we start
the app.

This is trivial for now but once we have some functionality it will be
more interesting