Add new playwright test for project creation flow

[KatieB5]

Closes #5621

Summary

This PR introduces a basic 'happy path' Playwright test for the project creation flow.

Context: Slack thread with discussion about how to structure the test and what to include.

WIP notes about the factory fixture to create a login context for the user:

I originally built the factory fixture yield _login_context_for_user with a variable in the outer function to keep track of multiple contexts in case we wanted this functionality in future, and used .close() to close the browser contexts we created at the end of a test. We don't use .close() in these opencodelists or airlock examples, but the example test in the Playwright documentation shows closing a new context that we have created once it's no longer needed, so I used it. Playwright also advises that we use .close() for any new contexts that we create ourselves (not the Playwright default).

new_context
Added before v1.9
Creates a new browser context. It won't share cookies/cache with other browser contexts.

note
If directly using this method to create BrowserContexts, it is best practice to explicitly close the returned context via browser_context.close() when your code is done with the BrowserContext, and before calling browser.close(). This will ensure the context is closed gracefully and any artifacts—like HARs and videos—are fully flushed and saved.

But then I changed my mind and paired the approach back:

• The brief in Create project form - test project creation flow #5621 is to write a basic Playwright test. The definition of basic is subjective, but I interpret this to mean keep the test as simple as possible, whilst providing sufficient coverage of the workflow under test
• Multiple browser contexts: YAGNI - currently, the basic playwright test for the project creation workflow is a 'happy path' test requiring one user and one new context for that user; we aren't creating multiple contexts in one test that we need to track
• .close(): I'm semi-confident (i.e. still somewhat unsure) the Playwright documentation on test isolation means that each test gets it's own browser, and therefore any browser contexts created will not be shared between tests; thus we can be reasonably confident that even if we don't implement context.close(), the contexts will not impact other tests in our test suite

Questions:

• Do we need to create a new context (context = browser.new_context(locale="en-GB")) in the factory fixture? Could we use the default context that Playwright gives us per test? We currently don't need multiple users, or multiple browser contexts.
• There is a (very small, I think) chance that we have a flaky assertion where we check Created at date. Should we implement control over the time/date in the test, or is the risk small enough that we don't need to do this?
• Do we want to test that the correct breadcrumbs (title and link) are showing on each page the user sees or is this unecassary)?

AOB

• Run just test-functional --headed --slowmo 1000 tests/functional/test_project_creation.py to see what's happening in the test; 1000 is the number of milliseconds to slow each Playwright operation down by
• There are some other useful CLI options you can pass in listed here.