Sign error in fourier_shift function?

[mperrin]

I've just noticed what I believe is a sign error in the fourier_shift function:

pandeia-coronagraphy/jwst_pancake/transformations.py

Line 106 in 6b30b0b

offset = fourier_shift( np.fft.fftn(image), (-yshift,xshift) )

Why is there a negative sign on the y shift but not on the x shift?

Note, this is a benign error if you avoid calling fourier_shift directly and instead invoke it through the register_to_target function (since the optimization in that function just solves for the required value with that sign error). But it's confusing if you try to directly access the fourier_shift function, or even if you just look at the output shifts as reported from register_to_target as a sanity check.