You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Following the EKO transition from an internal Numba-version of ekore, to a separate expressions' library, available to external users without the need of specific dependencies, I'd like to propose the same process to happen in yadism, for its coefficient functions library.
We could call it yadkore or whatever, but the main goal would be to distribute it separately from the main package.
This would involve a bit of refactoring, in particular:
the current classes-based structure should be dropped in favor of instances
the instances should be labeled by label objects, containing the information classifying the expression in a structured way (to allow us
this will be done on two levels: collections and individual elements
collections will be related to (observable, flavor, ...)
elements will be addressed with (pto, llo) (perturbative and resummation order) as integers, as opposed to the current one dimensional structure based on strings
we should keep room for SIA (polarized coefficients already live in a partitioned space of observables, so they do not need special care)
The added benefit of points 2. and 3. is that constructions like FONLL will become much simpler then they currently are.
The text was updated successfully, but these errors were encountered:
Following the EKO transition from an internal Numba-version of
ekore
, to a separate expressions' library, available to external users without the need of specific dependencies, I'd like to propose the same process to happen in yadism, for its coefficient functions library.We could call it
yadkore
or whatever, but the main goal would be to distribute it separately from the main package.This would involve a bit of refactoring, in particular:
(observable, flavor, ...)
(pto, llo)
(perturbative and resummation order) as integers, as opposed to the current one dimensional structure based on stringsThe added benefit of points 2. and 3. is that constructions like FONLL will become much simpler then they currently are.
The text was updated successfully, but these errors were encountered: