Possible duplicate reproject() function

[emolter]

Is there any meaningful difference between stcal.alignment.util.reproject() and stcal.outlier_detection.utils.reproject()? Could the one in outlier detection be removed? @mcara @nden @braingram

[mcara]

is it used anywhere?

[mcara]

calc_gwcs_pixmap and calc_pixmap are no longer used. these were the main clients. All those utils modules need a clean-up based on what is used.

[emolter]

calc_gwcs_pixmap is used inside gwcs_blot, which is called from the outlier detection step in both jwst and romancal. That function currently uses the reproject version that is in stcal.outlier_detection.utils.

I agree RE calc_pixmap and the reproject version that is in stcal.alignment.util - these appear unused.

There is also a calc_gwcs_pixmap and reproject in Romancal: https://github.com/spacetelescope/romancal/blob/7484c3870708e5b00ab389e5ade94150373fd9fe/romancal/resample/resample_utils.py#L253, and that version of calc_gwcs_pixmap is currently used in two places in Romancal's resample step. It appears that those will be removed by spacetelescope/romancal#1634, and Romancal will start calling gwcs_blot from stcal.outlier_detection.utils.

None of the three reproject functions (Romancal, alignment, outlier_detection) are identical, although it looks to me like they accomplish the same thing.

I would suggest that we decide which of these is the "best" and then remove the others. Do you have any opinion on which implementation is preferred?