Rounding of Taxes and Transfers

[MImmesberger]

Closes #88

Todo

• Wohngeld in steps of 5€.
• Income deduction from Kinderzuschlag in steps of 10€ until 07/2019.
• Vorsorgeaufwendungen are rounded up.
• Income tax is rounded down.
• Unterhaltsvorschuss is rounded up.

[codecov]

Codecov Report

Attention: 2 lines in your changes are missing coverage. Please review.

Comparison is base (d2c717d) 90.96% compared to head (d4b8091) 91.01%.

Files	Patch %	Lines
src/_gettsim/interface.py	60.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #659      +/-   ##
==========================================
+ Coverage   90.96%   91.01%   +0.05%     
==========================================
  Files          48       48              
  Lines        3131     3140       +9     
==========================================
+ Hits         2848     2858      +10     
+ Misses        283      282       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[MImmesberger]

The PR is ready for review.

1. I couldn't find the legal basis for Wohngeld being payed out in 5€ steps, so I did not implement it. See discussion in Rounding is missing for several taxes / transfers #88
2. The rounding of zu versteuerndes Einkommen includes an additive part (e.g. round down to a number divisible by 54 and then add 18€). A summary is here. Because this wasn't possible with the current machinery, I added an "add" key to the rounding specification that adds a float after the rounding step.

[hmgaudecker]

I couldn't find the legal basis for Wohngeld being payed out in 5€ steps, so I did not implement it. See discussion in

Yes, let's leave it at the nearest Euro!

[hmgaudecker]

Very sweet! I think I would get rid of the case distinction for the add parameter (and rename it), just requiring that it is zero if not used.

[MImmesberger]

Thank you for the comments! I hope I addressed all your concerns.

[MImmesberger]

Sorry, I still have trouble interpreting codecov outputs:

From the output I figured that _add_rounding_to_one_function is not properly covered by a test (only implicitly because it is used in the wrapper _add_rounding_to_functions which is tested). But testing directly via test_rounding_callable doesn't help.

Maybe @lars-reimann knows whats going on?

[hmgaudecker]

Oh, codecov happens to be off every now and then. If you are sure it is being run, that's fine!

[hmgaudecker]

Looks great, just one tiny comment left.