Add REPL plugin hooks; Add output, output-mode, stderr attributes

[JeffersGlass]

This PR does four major things

• Add two new plugin lifecycle methods, beforePyReplExec() and afterPyReplExec(), which which are called immediately before and after a py-repl tag is evaluated, respectively.
• Restore the output attribute of <py-repl> tags. If provided, the Element with the ID specified by output is used as:
  • The location to send writes to sys.stdout
  • The location where the result of evaluating the REPL (if any) is display()'d
• Restore the output-mode attribute of <py-repl> tags. If the output-mode == 'append', the contents of the output element (either specified by the output attribute or the default location as the next sibling of the REPL) are not cleared before writing new contents. This is not typical "REPL-like" behavior, but may be wanted if the user is already writing the REPL output somewhere else. (Removed in #881.)
• Add a stderr attribute for <py-repl> tags. If provided, the Element with the ID specified by the output is used as the location to send writes to sys.stderr

All writes to the above locations (using output or stderr attributes) are in addition to writing to the py-temrinal / ioMultiplexer.

---

This PR is a ready for review. The task list was:

• Add beforePyReplExec(), afterPyReplExec() to to plugins
• Expand StdioDirector plugin to accommodate new <py-repl> attributes
• Integration Tests:
  • New integration tests for new attributes
  • Update old plugin integration tests with new attributes
• Documentation
• Changelog Entries
• Review for completeness and correctness

Resolves #988, Fixes #1070, Resolves #975, Resolves #1074

[JeffersGlass]

Apologies that this has stayed in draft for so long - I hope to get this completed and ready to review sometime this week.

[FabioRosado]

Apologies for the LOOOOONG delay in reviewing this PR Jeff! Left some comments/suggestions and approving 😄

[marimeireles]

Just some minor comments here
🚀
Great work!

[marimeireles]

Oh no, a lot has changed 😕
@JeffersGlass lmk if you want I can take this over and rebase after I finish the event stuff.

[JeffersGlass]

@marimeireles @FabioRosado Finally got around to addressing the comments and merging from main! Would you mind doing a re-review?

For some reason, the Circle CI tests are stuck on a mergability check from a much earlier commit, and won't run the most recent tests at all. Any ideas how to clear those? Everything's merged and passing tests (locally) as of right now, for what it's worth. If they won't clear, I'll open a fresh PR with this branch again I guess.

[marimeireles]

No idea why @JeffersGlass, sorry I missed your comment here.
To me it shows a bunch of unresolved conflicts.
Guess I'll give it a try on these :)

[marimeireles]

Ok, I've rebased 👍
Got quite a few failing tests.

tests/integration/test_00_support.py ...............                                   [  8%]
tests/integration/test_01_basic.py ................s.                                  [ 18%]
tests/integration/test_02_display.py .......................                           [ 31%]
tests/integration/test_03_element.py ...............                                   [ 39%]
tests/integration/test_async.py .......                                                [ 43%]
tests/integration/test_importmap.py xX                                                 [ 44%]
tests/integration/test_interpreter.py ....                                             [ 47%]
tests/integration/test_plugins.py ...F.......                                          [ 53%]
tests/integration/test_py_config.py ..FF........                                       [ 60%]
tests/integration/test_py_repl.py ............F..F.FFFFFF                              [ 73%]
tests/integration/test_py_terminal.py .......                                          [ 76%]
tests/integration/test_splashscreen.py ...s..                                          [ 80%]
tests/integration/test_stdio_handling.py ...........                                   [ 86%]
tests/integration/test_style.py ..                                                     [ 87%]
tests/integration/test_warnings_and_banners.py .                                       [ 88%]
tests/integration/test_zz_examples.py F.....s............x.                            [100%]

Will make sure I've merged everything correctly cause it was a bunch of files and then force push it here.

[JeffersGlass]

Got quite a few failing tests.

This is very odd to me - I merged from main yesterday, all tests passing locally, and pushed. I think the merge conflicts here are potentially an old circle ci artifact that won’t close? Going to try opening a fresh PR from the branch and see what happens…

[marimeireles]

Hey @JeffersGlass, that would also work!
I've made a fresh PR here if you think it's easier and if it looks correct for you: #1301
But if the tests were passing locally to you then I might have merged something wrong... I was checking and it seems fine but will wait for you to come back to me so we don't dup. work.

[marimeireles]

Ok great :)
This looks great!
I'll close my dupped PR and we continue from here =)

[JeffersGlass]

Sounds good @marimeireles! I suspect I had messed up the inital merge/rebase, but this one is at least free of merge conflicts now as well. Looks like tests are passing here, but linting is not.

[marimeireles]

You can also see the local linting issues by running pre-commit run --all-files.

[marimeireles]

Went over it again and it looks good!
Thanks for coming back to this one @JeffersGlass, great work! 😊

[marimeireles]

Actually, there's one problem.. Let me check.
I'm looking into: https://github.com/pyscript/pyscript/actions/runs/4492024745/jobs/7901322111?pr=1106
That was introduced by me as I fixed linting.

[marimeireles]

Okay, this is good now!

[JeffersGlass]

Amazing! I'll get it merged later today!