Refactor config and pyinstaller build

[KIRA009]

What kind of change does this PR introduce?
This PR changes the config to use json files, and exposes some configuration for modifications through the dashboard. This PR also adds build scripts to create executables for both macos and windows os-es.

Screen.Recording.2024-04-20.at.10.58.44.PM.mov

Summary

This PR adds a page to the dashboard that will allow users to alter certain configuration settings for the project.

It also modifies the release workflow to add a build step that builds the executables for windows and macos, and then uploads them as artifacts to the latest release

Checklist

• My code follows the style guidelines of OpenAdapt
• I have performed a self-review of my code
• If applicable, I have added tests to prove my fix is functional/effective
• I have linted my code locally prior to submission
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation (e.g. README.md, requirements.txt)
• New and existing unit tests pass locally with my changes

How can your code be run and tested?
To build the project into an executable app, run the following command in a poetry shell

python -m openadapt.build

To run the project, you can now use the following command in a poetry shell

python -m openadapt.entrypoint

[KIRA009]

@abrichr This should be ready for review now, I have rebased all the functional changes to a separate commit - ded9c66, and the subsequent commit only contains changes of updating the references of the config object in various files

[abrichr]

Excellent work! I left a few comments, please let me know if you would like to chat about any of it 🙏

[abrichr]

This appears to be a typo

[KIRA009]

Yep

[abrichr]

Why is noqa required here?

[abrichr]

What do you think about renaming this terminate_process?

[KIRA009]

While it makes sense, this flag is mostly used as a way to stop from reading and processing events, so maybe terminate_event_processing?

[abrichr]

Or terminate_processing to be more terse, either works

[abrichr]

What do you think about this slight refactor:

def is_running_from_executable() -> bool:
    """Determine if the Python script is running as an executable."""
    return getattr(sys, 'frozen', False)

class RedirectOutput:
    """Context manager to redirect stdout and stderr to /dev/null."""

    def __enter__(self) -> 'RedirectOutput':
        if is_running_from_executable():
            self.old_stdout = sys.stdout
            self.old_stderr = sys.stderr
            null_stream = open(os.devnull, 'a')
            sys.stdout = sys.stderr = null_stream
        return self

    def __exit__(self, exc_type: type, exc_value: Exception, traceback: type) -> None:
        if is_running_from_executable():
            sys.stdout.close()
            sys.stderr.close()
            sys.stdout = self.old_stdout
            sys.stderr = self.old_stderr

def redirect_stdout_stderr() -> RedirectOutput:
    """Get the RedirectOutput instance for use as a context manager."""
    return RedirectOutput()

• Open /dev/null once and assign it to both stdout and stderr to avoid opening it twice.
• __enter__ properly returns an instance of the class itself ('RedirectOutput').
• __exit__ ensures that the original streams are restored even if an exception occurs within the block.

[abrichr]

@KIRA009 there appears to be a regression affecting the pipes at the end of a recording:

bug.mov

The progress bar appears to complete before all events are saved. This behavior does not happen on main.

[abrichr]

LGTM! 🚀