Added typing to functions and improved docstrings to be more descriptive.

[BeatsuDev]

Added typing to all the functions and parameters in __init__.py. Also added Sphinx-styled documentation to explain the parameters of the functions.

[Danielhiversen]

Thanks.
Please fix the failing tests

[BeatsuDev]

I think it should be good now 👍

[BeatsuDev]

I read through the code again and made some small changes/fixes (ref: last two commits on this PR). These shouldn't break anything or throw any new warnings or errors.

[Danielhiversen]

Great.
Mypy should also be added to test the typing

https://github.com/Danielhiversen/pyTibber/pull/131/files#diff-51335077549f8f95664f409b515ccf2019991e216d97f29b790ddde6c4be4033R46-R48

[BeatsuDev]

Didn't know about mypy. Seems like a great tool! I'll look into it, thanks! 😄

[BeatsuDev]

Notable changes after using mypy:

• I changed the declaration of variables from self.foo: str = None to self.foo: str = "". This was only changed in the last commit and can be easily reverted if it causes problems.

• Adding to the previous point, I changed the initial value of self.last_data_timestamp, self.peak_hour_time and self.last_cons_data_timestamp to dt.datetime.min. I checked the first and last properties for usages and since they are only used to check if they're lower than another value, it should be fine.

• Some places I had to add return None instead of simply return to stop mypy from complaining.

• Since Tibber.execute can return None, I added an if statement and only set the value of self.info if the return value of .execute is defined. I've added TODO comments for potential error handling in the code.

[BeatsuDev]

Do you think it's okay to just return None instead of overwriting self.info with a potential None value? Does this break any functionality? There are 2 other places where I've done this (3 in total)

[BeatsuDev]

@Danielhiversen ^ The question is really: When a user fetches data, but it fails to get that information because of a request error, should the method return None, overwrite self.info with value None, or should an exception be raised?

[Danielhiversen]

return None is ok for this pr

[Danielhiversen]

Have you tested your changes locally?

[BeatsuDev]

Have you tested your changes locally?

Yep! The only workflow I could not get to run was the isort workflow. I don't think that's a problem with the repo however, but rather on my side...

isort.exceptions.InvalidSettingsPath: isort was told to use the settings_path: C:\Users\Beatsu\Documents\Coding\GitHub\pyTibber\** as the base directory or file that represents the starting point of config file discovery, but it does not exist.

[BeatsuDev]

Running isort . however (which should be the equivalent to isort **/*.py?) changes the imports in setup.py (which I haven't touched) and test.py (which I also haven't touched) as well as __init__.py. Shall I apply this change?

[BeatsuDev]

I actually double-checked and just realised that the tests are being skipped because pytest-asyncio is not installed... The third test then fails because at assert tibber_connection.name is None, because the name is an empty string instead (""). I'll look into changing all/most the properties into Optional[...] types instead.

[BeatsuDev]

I added a new dependency (pytest-asyncio) so that the tests could be run and not just be skipped (see issue #178). There were some issues with the cache so the setup of the workflow failed. Therefore removed the caching step and used python-setup's built-in caching feature instead. This required a requirements.txt to install from afaik, so I added that. In short, the last changes done in this PR will fix #178.

[BeatsuDev]

I could revert these last changes (requirements.txt and pytest-asyncio) and add them in a different PR if you'd like. Shall I do that?

[Danielhiversen]

Fixed by #181

[BeatsuDev]

Not sure why pylint did not use the setup.cfg in this branch, but it worked fine in the master branch... Anyway, I fixed it by shortening the line 👍