raise validation error for all zero response data

[VadimKraus]

This is a possible solution for #67.

After this PR:

• the response dict will be validate using a schema definition, therefore the access to the dict properties are more safe now
• the check includes a validation that at least one value in 'Data' is not 0

Consequences for Home assistant:
If an zeros data event occurs, the validation error will cause the sensor update to fail and thereby the inherent home assistant retry will be triggered. No Changes to the solax integration are required.

[Azelphur]

Any way we could get some movement on this? happy to contribute, this has been broken for over a year now and it is bugging me :)

[VadimKraus]

Reworked based on current master.

[joseal]

Hi VK, thanks for the push! Quick question: do you know the version/build when it will hit GA?

[VadimKraus]

Hey @joseal, that depends on @squishykid. He might have different plans about his issue :)

[joseal]

@squishykid , any concern? This issue is disturbing a ton of people for a year... it was great if it could be fixed. Thank you.

[joseal]

Any news about this issue? I still need to be constantly manual fixing the values...

[joseal]

Hi @squishykid any news?

[squishykid]

Thanks for the reminder!

I will also cut a new release, so it can be updated in home-assistant

[joseal]

Brilliant! Thank you very much.

[squishykid]

@joseal can you please take another look at this? I think CI was broken, but I had to revert this commit because it broke discovery

[joseal]

@joseal can you please take another look at this? I think CI was broken, but I had to revert this commit because it broke discovery

If I'm not wrong, this change was made by @VadimKraus... VK, could you please check it? Thank you!

[VadimKraus]

The discovery error is unrelated to this change.
Try running the test multiple times, you will find them failing eventually.
pytest --count=10 tests/test_discovery.py

[VadimKraus]

Did some testing: the discovery has been in this broken state since a7a95f7.

This causes a race condition where 'solax.inverters.x1_boost.X1Boost' and 'solax.inverters.x1_mini_v34.X1MiniV34' are being confused with each other. They have basically an overlapping schema...

@squishykid I will reopen this PR, since the problem is unrelated.

However, the discovery needs to be fixed ...

• added "Update" raise validation error for all zero response data #142
• added Discovery non deterministic results #143