Clean up battery, alert and hidpp related code

[MattHag]

Simplifies hierarchy, adds type hints and enhances readability. Fixes capitalization of constants, which is violated by NamedInts.

Related #2273

[MattHag]

Don't know what's the issue with the more or less unrelated test failure

[pfps]

The first two commits look fine.

[pfps]

The third commit looks fine.

[pfps]

I do not understand the rationale for creating fake classes for a small part of Device.

[MattHag]

Which fake classes?

[pfps]

Can IntEnum be used instead of NamedInt everywhere in Solaar? If so, then it is a good idea to do the replacement. If not, then Solaar should stick with NamedInt throughout.

The NamedInts that need to be checked are the ones that have default values, I think.

[pfps]

DeviceProtocol

[MattHag]

Can IntEnum be used instead of NamedInt everywhere in Solaar?
Sure, here it was useless to have special implementation, loose proper type hints and linter checks.

The protocol makes clear which functions are used in that module. And that's the foundation for further decoupling to simplify the code complexity and enable boring, but effective tests and implementations.

[MattHag]

@pfps ready

[pfps]

I am not convinced of the need to add the protocols. If there is a PR without these commits, I'll merge it in.

[MattHag]

The protocols are important for further clean up of the data read/write related code, as they indicate, which actions are needed in a module. The hidpp, device and base modules have interesting data paths and thus dependencies, which can be tackled on top of that.

[pfps]

If the prototols are important for future work split them off for now. When the future work is being performed they can be re-evaluated.

[MattHag]

Fixed and simplified

[pfps]

I'm seeing extra messages when running bin/solaar -dd and a device not showing up.

2024-05-11 11:50:34,378,378     INFO [MainThread] hidapi.udev: Report Descriptor not processed for DEVICE /dev/hidraw0 BID 0003 VID 0000046D PID 0000C52B: open.<locals>.matchfn() takes 3 positional arguments but 5 were given

[pfps]

@MattHag Is there a better way to set up the Linux vs MacOS hid code? Would it be a good idea to move the hidapi directory to be a subdirectory of logitech_receiver?

[MattHag]

Yes, there's certainly a different one, but it's not the most relevant.

No, all those low level communication interface should be self-contained and have no dependency or tight connection to the main Solaar code /solaar.
The HID package should probably be used like a pip dependency and all Solaar specific code from there should be moved to the solaar package, as described in #2480. The addition of #2481 will make it easier to talk about the purpose of the packages and reason about improvements.

[pfps]

The problem is that currently Solaar installs hidapi as a top-level module in .../site-packages and there is a separate hidapi package in PyPI. So either Solaar should just use that package or the Solaar HID code should be moved inside the low-level part of Solaar, i.e., inside logitech_receiver. The question is which way to go and if the HID code is retained how best to interact with the two back-ends.

At the same time, it is probably a good idea to rename logitech_receiver as it isn't just about receivers.

[MattHag]

I'm seeing extra messages when running bin/solaar -dd and a device not showing up.

2024-05-11 11:50:34,378,378     INFO [MainThread] hidapi.udev: Report Descriptor not processed for DEVICE /dev/hidraw0 BID 0003 VID 0000046D PID 0000C52B: open.<locals>.matchfn() takes 3 positional arguments but 5 were given

Kept the fileopen alias, now it will work again.

[pfps]

Looks acceptable