Open teuben opened 1 year ago
aha, karma doesn't know how to deal with bitpix=-64 data, which is what we got by default.
Setting dtype=np.float32 as extra argument to griddata() will be the new default in the pipeline, the default being np.float64
A temporary solution is NEMO's conversion:
fitsccd file64.fits - | ccdfits - file32.fits
it will have lost some of the WCS, but kpvslice is not complaining.
there is some issue with the FITS WCS in final cubes:
WARNING: Expected '=' at column: 9 in line: "CONTINUE 'BPA=0.0 deg'" WARNING: Expected '=' at column: 9 in line: "CONTINUE '[(None, None, 2), (None, None, None), (None, None, None)]]'" Unknown unit: "deg" for: "CUNIT1", dropping Unknown unit: "deg" for: "CUNIT2", dropping Unknown unit: "km s-1" for: "CUNIT3", dropping Data type: 1 not found