Internally, PyWiFeS is still using a 25+ year old Python conversion of a FORTRAN implementation of Levenberg-Marquardt minimisation (mpfit), which comes bundled with PyWiFeS. It was minimally updated to be Python 3-compatible in 2019, but otherwise has been left more-or-less alone.
GIven PyWiFeS is now pinned to a fairly modern version of SciPy (1.15.1), which includes a vast array of minimization functions (https://docs.scipy.org/doc/scipy/reference/optimize.html), is there any particular reason to stick with this internal mpfit implementation, or should we consider moving to a more flexible SciPy fitting routine?
Internally, PyWiFeS is still using a 25+ year old Python conversion of a FORTRAN implementation of Levenberg-Marquardt minimisation (
mpfit), which comes bundled with PyWiFeS. It was minimally updated to be Python 3-compatible in 2019, but otherwise has been left more-or-less alone.GIven PyWiFeS is now pinned to a fairly modern version of SciPy (1.15.1), which includes a vast array of minimization functions (https://docs.scipy.org/doc/scipy/reference/optimize.html), is there any particular reason to stick with this internal
mpfitimplementation, or should we consider moving to a more flexible SciPy fitting routine?