Allow shifting a detector, in alternative to shifting a source

[mperrin]

Finish implementing the long-delayed option to shift a detector, rather than shifting the point source. This is implemented as an option for offsetting a detector relative to the nominal center of a beam. The underlying math for this has been implemented for years in the matrixDFT.py functions; this PR adds the small amount of code to the Detector class to pass through parameters to the matrix DFT functions.

Conceptually, shifting a source (+dX, +dY) should be equivalent to shifting the detector (-dX, -dY). There's a unit test implemented that verifies that is the case. (In fact, that unit test is >50% of the code in this PR)

This option may be used in certain future features that may be added to webbpsf/stpsf particularly in the context of coronagraph simulations. (We can already shift the source, and shift the coronagraph mask; this added the option to shift the detector also).

To Do:

• add astropy units handling

[mperrin]

@BradleySappington this one is a bit more technical optics in part, but hopefully the existence of a test function that passes should suffice to verify the physics part of this works as intended.

[BradleySappington]

Code is clean/commented and passing tests