Add robust fitting back to fit_origin

[sezelt]

Fixes #548

[bsavitzky]

Looks good, thanks @sezelt. Not sure how things wound up in this strange state.

There will be conflicts with the edits I made in my GTO branch to add the masking functionality correctly back in to this method, but I'll just pull from dev and handle them locally before PRing those changes. I'll likely modify the arguments at that point as well (mask_check_data is nutso), so small modifications to user-side code may be necessary then, but these should be small things.

Lmk if you still need the masking mods given this PR. Since I'll need to handle conflicts before I can push that stuff, I can try to get to that sooner if you need it or a bit later if not.