Refactor example models

[timmens]

In this PR, I refactor the example models.

Previously, the example models were located in src/example_models but were not used in any of the code but only in the tests. One example model did not have a meaningful test but existed only for experimenting on the server.

To change this, I created two new folders:

1. tests/test_models contains all the models used in the tests.
2. examples contain example models that can be used elsewhere or as a reference but do not have a corresponding test.

I also added comments and cleaned up the example models in the process. Moreover, while deleting the module get_model.py, I updated the regression tests.

Closes #46

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.27%. Comparing base (4f97784) to head (48e6070).

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #56      +/-   ##
==========================================
+ Coverage   97.69%   98.27%   +0.57%     
==========================================
  Files          47       42       -5     
  Lines        2521     2433      -88     
==========================================
- Hits         2463     2391      -72     
+ Misses         58       42      -16     

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

[hmgaudecker]

Much clearer, thanks a lot!

I have a bunch of comments, most of which are unrelated to the changes made here, though 😇

[hmgaudecker]

Just small stuff left, thanks a lot!

If other models enter the tests, we may want to rename deterministic to base_deterministic etc., but no need to do that preemptively.

[hmgaudecker]

There seems to be quite some repetetive code setting up the models/params. In test_analytical_solution, you are now using _get_iskhakov_2017_model and _get_iskhakov_2017_params, which seems very nice.

Would it make sense to re-factor that into a module that is used to generate fixtures across multiple test files?

[hmgaudecker]

Why don't we make all parameters modifiable and leave the defaults at the values found in the main specification of the paper?

[hmgaudecker]

Same comment as for the params below, but less important here.

[hmgaudecker]

See main comment.