Support Smart-Home-Interface Part 5: Main interface

[Guzz-T]

Fifth pull request to gradually integrate the smart home interface. This pull request includes:

Contiguous blocks: Helper classes to group read/write operations
Interface: Main interface for the user
Pytest units for new methods and classes

Part 5 for #190

Extensions not yet included:

• Scripts
• Integration into the luxtronik interface

[github-actions]

Coverage Report

File	Stmts	Miss	Cover	Missing
luxtronik
__init__.py	184	11	94%	42–43, 46–51, 257–258, 263
__main__.py	21	21	0%	3–49
datatypes.py	325	1	99%	114
discover.py	58	1	98%	62
luxtronik/scripts
dump_changes.py	43	43	0%	7–85
dump_luxtronik.py	26	26	0%	6–52
TOTAL	1811	103	94%

Tests	Skipped	Failures	Errors	Time
299	0 💤	0 ❌	0 🔥	5.657s ⏱️

[kbabioch]

@Guzz-T Since you've pushed some new commits: Is this still work-in-progress, or ready to be reviewed?

[Guzz-T]

It should be finished. There might be a few minor changes later, but those could also be included in the next pull request.

[Guzz-T]

Since there are still no review comments, I’ve (once again) pushed a small change to reduce the overall effort required for the whole review 😅

[Guzz-T]

@kbabioch: Would you have time for a review?

[kbabioch]

Ah, sorry, totally forgot about this one. As always, super clean pull request, nothing to complain from my point of view.

Obviously, you're breaking the whole thing down into small / reviewable pieces and are following good practice in terms of software engineering. Just wondering how "complex" the overall component / interface will become and if it might be over-engineered :-). But this is just my split personality, since I like well architectured software ;-).

[Guzz-T]

Ah, sorry, totally forgot about this one. As always, super clean pull request, nothing to complain from my point of view.

Obviously, you're breaking the whole thing down into small / reviewable pieces and are following good practice in terms of software engineering. Just wondering how "complex" the overall component / interface will become and if it might be over-engineered :-). But this is just my split personality, since I like well architectured software ;-).

Much of the complexity stems from the fact that the controller/pymodbus interface aborts reading/writing with an exception if a non-existent register is addressed. Therefore, it must be divided into many small operations. This also causes problems with older controller versions that does not support some registers.

Maybe I'll find a way later on, which would simplify the whole process.

But yes, I have a tendency to make things a little more complex than is strictly necessary.

[Guzz-T]

Could someone please merge?