[x] Refactoring of enrichment classes: create a parent class that can be inherited to implement other enrichment methods (e.g uniform) (https://github.com/tellae/bhepop2/issues/17)
[x] Improve testing: add more unit tests, less macro tests. Evaluate macro tests on aggregated results instead of exact comparison with a stored table.
[x] Create "accessibility" methods: avoids creating the class, calling optimize and assign_feature_value_to_pop. Facilitates exterior use. Try to answer the question how do we want to call the bhepop2 library (in Eqasim for instance) (https://github.com/tellae/bhepop2/issues/26)
[x] Improve/complete documentation, adapt to qualitative approach: maybe explain the package on a qualitative example, simpler (https://github.com/tellae/bhepop2/issues/21)
[x] Improve example notebooks (rename with distribution type quantitative/qualitative), make them simpler, lighter
[ ] Have a single, smaller synthetic population for examples and tests. For distributions, we have one real life data set (Filosofi) and a distribution that comes from a population, then used to re-enrich this population. Heavy and realistic VS no distribution data and almost meaningless.
optimize
andassign_feature_value_to_pop
. Facilitates exterior use. Try to answer the question how do we want to call the bhepop2 library (in Eqasim for instance) (https://github.com/tellae/bhepop2/issues/26)