The ARD pipeline uses the idl-functions package for a trivial reason: histogram of discrete classification classes. I tested that the behavior of numpy.histogram with proper arguments is identical to the idl-functions version in this particular case.
Some possibly unused scripts reference it, but its usage their also does not require IDL-like behavior. I don't think my version of wagl.scripts.wagl_residuals is identical to the idl-functions version, but it is morally the same (the difference has to do with how numpy calculates bin boundaries).
The ARD pipeline uses the
idl-functionspackage for a trivial reason: histogram of discrete classification classes. I tested that the behavior ofnumpy.histogramwith proper arguments is identical to theidl-functionsversion in this particular case.Some possibly unused scripts reference it, but its usage their also does not require IDL-like behavior. I don't think my version of
wagl.scripts.wagl_residualsis identical to theidl-functionsversion, but it is morally the same (the difference has to do with hownumpycalculates bin boundaries).One less thing to worry about going forward.