Feature/botmessage

[ilikecubesnstuff]

Addresses #9, along with other changes. Notable changes include:

• Updated test suite following the move to tortoise ORM v1.0.
• Updated workflows for pushes and pull requests, including running tests.
• Renamed cog from "Embeds" to "BotMessage", with the following interface:
  • &send for sending arbitrary messages, including those with embeds.
  • &edit for editing an existing bot message, including those with embeds.
  • &embedraw legacy command, which sends warnings and updated JSON syntax when invoked.
  • 3 context menu commands for retrieving JSON, adding embeds, and editing messages.

[ilikecubesnstuff]

Note for future: how to run tests is not documented anywhere at the moment?
uv run pytest will only work once the docker-compose.test.yml file is run and the test database is active to interact with.

[archivehaze]

I'm happy with this PR, thank you for adding the workflows and changing the tests so they run under tortoise orm. Just one question, did you add the changes that we discussed about the error message when attempting to edit a message that the bot didn't send? Also, is the check to see if they manage the guild necessary? As far as I'm aware, that checks if the user is an admin, but won't the admins also have the moderator role?

[ilikecubesnstuff]

Thanks for the comments! Addressing each point:

Just one question, did you add the changes that we discussed about the error message when attempting to edit a message that the bot didn't send?

Turns out I managed to delete the code that actually registers the context menu commands to the command tree. That's added back in now, but there were hiccups in the checks I had implemented for this. I've now shifted this validation logic inside the callback, and it should work as you described above.

Also, is the check to see if they manage the guild necessary? As far as I'm aware, that checks if the user is an admin, but won't the admins also have the moderator role?

From my understanding, this ensures that the command works when invoked by an admin, even if they happen to be missing the moderator role. This reflects the logic in src/ib_py/utils/checks.py:

def staff_command():
    return commands.check_any(commands.check(is_moderator), admin_command())

Unless there is a need to change this, I think that this fallback behavior is what is preferred/expected.