Define functions on scalars and vectorise them

[ChristianZimpelmann]

What problem do you want to solve?

• Make functions easier to read and understand
• Make code much faster

Todo

• Implement functions to aggregate values on hh or tu level
• Rewrite all functions such that they are defined on scalars
• Implement vectorization (for numpy , numba, and jax)
• Rewrite interface.py such that all columns are on individual level
• If it is a relevant contribution, put it in CHANGES.rst.

Closes #281

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (main@9a626a9). Click here to learn what that means.
The diff coverage is 93.65%.

❗ Current head 0b9338d differs from pull request most recent head a5419b4. Consider uploading reports for the commit a5419b4 to get more accurate results

@@           Coverage Diff           @@
##             main     #356   +/-   ##
=======================================
  Coverage        ?   93.09%           
=======================================
  Files           ?       77           
  Lines           ?     3478           
  Branches        ?        0           
=======================================
  Hits            ?     3238           
  Misses          ?      240           
  Partials        ?        0           

Impacted Files	Coverage Δ
gettsim/synthetic_data/synthetic.py	94.54% <ø> (ø)
gettsim/taxes/soli_st.py	100.00% <ø> (ø)
gettsim/tests/test_interface.py	96.66% <ø> (ø)
gettsim/transfers/benefit_checks/benefit_checks.py	100.00% <ø> (ø)
gettsim/visualization.py	52.67% <27.27%> (ø)
gettsim/config.py	85.71% <60.00%> (ø)
gettsim/aggregation.py	67.39% <67.39%> (ø)
gettsim/transfers/unterhalt.py	84.61% <71.42%> (ø)
gettsim/functions_loader.py	85.71% <79.16%> (ø)
gettsim/typing.py	81.25% <80.00%> (ø)
... and 57 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 9a626a9...a5419b4. Read the comment docs.

[hmgaudecker]

Also still needed: Search for "Series" (uppercase) and replace appropriately at least in the GETTSIM functions (not necessary to go in detail through interface.py right now).

There are quite some instances in the docs that were not using "FloatSeries" and friends so far.

[hmgaudecker]

On the phone and cannot directly reply at the right spots: - I'd stick to vermietung and also vor_vermög_check (or the like) — simply not reason to rename in this PR. Sorry, I see why that was confusing. -17/24 + höchstalter, unless you come up with a really good name (I did not, but that does not mean anything) Thanks!

[ChristianZimpelmann]

I'd stick to vermietung and also vor_vermög_check (or the like) — simply
not reason to rename in this PR. Sorry, I see why that was confusing.

I am still not sure what your prefered name is. We have two variables:

• input variable on monthly basis: current version and version before this PR: eink_vermiet_m
• defined function which calculates the value on yearly basis: before this PR eink_vermietung, current version: eink_vermiet

Just let me know how you would like to change it.

-17/24 + höchstalter, unless you come up with a really good name (I did
not, but that does not mean anything)

How about altersgrenze? I still think we should, if possible, use the parameter values mentioned in the law.

• Otherwise, all comments should be addressed now.
• I haven't implemented the user-provided aggregation-dictionary, yet. Should we postpone?
• I tried to add python 3.10, but didn't get to run a working conda environment, yet. So let's postpone this for a bit.

[hmgaudecker]

I'd stick to vermietung and also vor_vermög_check (or the like) — simply
not reason to rename in this PR. Sorry, I see why that was confusing.

I am still not sure what your prefered name is. We have two variables:

input variable on monthly basis: current version and version before this PR: eink_vermiet_m

defined function which calculates the value on yearly basis: before this PR eink_vermietung, current version: eink_vermiet

Ahh, I did not remember the former. Then we forgot to rename that, I believe. eink_vermietung_m / eink_vermietung, if that is not too much of a hassle

How about altersgrenze? I still think we should, if possible, use the parameter values mentioned in the law.

Very good.

Otherwise, all comments should be addressed now.

Perfect!

I haven't implemented the user-provided aggregation-dictionary, yet. Should we postpone?

Sure, just open an issue. Or a PR directly, should be just a couple of hours' work, right? Then would be nice to have in release.

I tried to add python 3.10, but didn't get to run a working conda environment, yet. So let's postpone this for a bit.

Sure.

[hmgaudecker]

Excellent!