search

tBuLi / symfit

Symbolic Fitting; fitting as it should be.
http://symfit.readthedocs.org
MIT License
233 stars 17 forks source link

Remove leastsqbound, use scipy equivalent #253

Closed tBuLi closed 1 year ago

tBuLi commented 5 years ago

Rehash of #67: we should really remove leastsqbound.py in favor of the scipy equivalent, since that is peer reviewed whereas our current file hasn't undergone any changes since its creation.

See https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.least_squares.html

Should be a matter of deleting leastsqbound.py and changing all references to the function leastsqbound in symfit.core.minimizers.MINPACK to scipy.optimize.least_squares.

tBuLi commented 5 years ago

@DuckNrOne, I think this might be an interesting first step in your quest to increase coverage since this will simply knock out an ill-tested component without adding any new code. But no pressure, if it doesn't have your fancy just have a look at the unit tests, that would already be great.

antonykamp commented 4 years ago

while A uses multiple Bs, only 2 elements are allowed on scipy. Also, they have to be ordered, but that's not a problem compared to the first one.

tBuLi commented 4 years ago

Sorry, could you elaborate on that? I'm afraid I don't understand your comment. What are A and B?

antonykamp commented 4 years ago

I don't know why I used A and B :D While leastsqbound uses multiple boundaries, only 2 elements are allowed on scipy. Also, they have to be ordered, but that's not a problem compared to the first one.