structural changes: uniy sorting_key and observable
Closed this issue · 3 comments
The file pool_model.py has the function sorting_key(x) and in the file src/solving.py, we have to supply the observable which is often simply given by one of the predefined functions convert_S_matrix_to....
Additionally the function calculate_Fischer_observable in file src/solving.py needs to be adjusted such that all arguments times, Q_arr, P, Const, S are given to the observable function.
The file src/optimization.py uses this sorting_key function which is pointless since we always want to sort by the observable itself (which is simply a number).
Does it mean that we want in the functions convert_S_matrix_to... straightaway to divide the determinant (or other observable) by effort? And save already this new value in fischer_results, right?
The goal is to get rid of the sorting_key and only use ˋobservableˋ. We will provide some functions such as calculating the Fischer determinant or sum of Eigenvalues. Custom observable are to be defined in the corresponding model files.
In general: everything in src/... is model-agnostic!
Since commit 7b47808, sorting key is now a model-specific (to pool_model.py) quantity and not part of the abstraction.