-
Couldn't load subscription status.
- Fork 58
Description
PR #452 uncovered this commented out block of code.
pyGSTi/pygsti/evotypes/densitymx/opreps.pyx
Lines 730 to 745 in 6a3ebd5
| #TODO: can add this after creating OpCRep_IdentityPlusErrorgen if it seems useful | |
| #cdef class OpRepIdentityPlusErrorgen(OpRep): | |
| # cdef public object errorgen_rep | |
| # | |
| # def __init__(self, errorgen_rep): | |
| # self.errorgen_rep = errorgen_rep | |
| # assert(self.c_rep == NULL) | |
| # self.c_rep = new OpCRep_IdentityPlusErrorgen((<OpRep?>errorgen_rep).c_rep) | |
| # self.state_space = errorgen_rep.state_space | |
| # | |
| # def __reduce__(self): | |
| # return (OpRepIdentityPlusErrorgen, (self.errorgen_rep,)) | |
| # | |
| # #Needed? | |
| # #def errgenrep_has_changed(self, onenorm_upperbound): | |
| # # pass |
During review of that PR, Corey asked
Which rep type currently covers the identity plus error generator object? Is this already covered, or are we deficient in this regard in which case should this be re-enabled?
Corey subsequently looked into this, and came back with this assessment.
This should be resuscitated. If I understand the code correctly then this will result in an error when using this modelmember if the representation type is '1+L'. There is an implementation of this representation in
densitymx_slow, so an error would get hit when using the cython module (though that is the default when present).