Add some example servers

[alcarney]

I'm concious of the fact #418 is a rather large PR!

I've also realised that there's a growing amount of work in #418 that is not directly dependant on the new architecture. So, this PR is a first pass at breaking some of that work out into a separate PR which we can merge now.

As part of refactoring the tests under tests/lsp to be end-to-end tests, I have been introducing small example servers which not only help with the tests, but which should help anyone new to pygls/lsp get to grips with how things work. This PR adds the following example servers:

Filename	Works With	Description
code_lens.py	sums.txt	Evaluate sums via a code lens
colors.py	colors.txt	Provides a visual representation of color values and even a color picker in supported clients
goto.py	code.txt	Implements the various "Goto X" requests in the specification for a dummy programming language
hover.py	dates.txt	Opens a popup showing the date underneath the cursor in multiple formats
publish_diagnostics.py	sums.txt	Use "push-model" diagnostics to highlight missing or incorrect answers
pull_diagnostics.py	sums.txt	Use "pull-model" diagnostics to highlight missing or incorrect answers

Side note: I'm thinking we should start moving away from json_server.py as our flagship example for two reasons.

1. Small focused examples are probably easier to parse than one big one.
2. Editors like VSCode have a lot of built-in support for JSON - it might not be obvious to tell which features are from pygls and which are from the editor.

Code review checklist (for code reviewer to complete)

• [ ] Pull request represents a single change (i.e. not fixing disparate/unrelated things in a single PR)
• [ ] Title summarizes what is changing
• [ ] Commit messages are meaningful (see this for details)
• [ ] Tests have been included and/or updated, as appropriate
• [ ] Docstrings have been included and/or updated, as appropriate
• [ ] Standalone docs have been updated accordingly

[tombh]

This is epic, so many useful examples and tests 🥰 I'll take a good long look soon.

[tombh]

Sorry for the delay, I've had a good look and think about this now. Again, really epic work here. This is so much more than just adding great new tests, it's self-documenting and generous to newcomers and experts alike.

Also, as always, thank you so much for the self-contained commits, they really make understanding the changes here more digestible.

I'd be happy to merge this as is! But here are some comments, they're very minor, so unless they resonate, don't worry about them:

• Could the example server's README be linked to from the main README? And what about linking each example server from the bottom of https://pygls.readthedocs.io/en/latest/pages/getting_started.html#quick-start?
• You have added some comments to some of the example servers, and generally I roughly understand what each one does. But I'm wondering if a brief description could be added to each explaining what the in-editor experience of using the server might be like? For example, for the hover server: "When the cursor is placed over a date, you should see a small popup describing the recognised Python strftime formatting for the date"

• The following seems to be a common pattern:

  client, response = await get_client_for("...py")
  yield client, response
  
  await client.shutdown_async(None)
  client.exit(None)
  
  await client.stop()

  To what extent can it be refactored?

• Could assert test_uri is not None live in the uri_for function itself?

Dwi'n dysgu am englyn yn fy dosbarthiau cymraeg ddiweddar. Yr PR yma yn atgoffa o y llinell glasurol:

Campwaith dewin hynod

🪄

[alcarney]

Could assert test_uri is not None live in the uri_for function itself?

Yes, good spot

The following seems to be a common pattern: To what extent can it be refactored?

Perhaps get_client_for can handle it and the fixture just becomes yield from get_client_for(...), I'll have a play and see if it will work

[alcarney]

Turns out you can't yield from an async generator, but an async for loop seems to work, so the fixtures now have the form

@pytest_asyncio.fixture()
async def json_server(get_client_for):
    async for result in get_client_for("json_server.py"):
        yield result

And what about linking each example server from the bottom of https://pygls.readthedocs.io/en/latest/pages/getting_started.html#quick-start?

I'll probably handle this in a separate PR - perhaps #427

[tombh]

Great 😊