resolve competition between hsd2nc.py, ahi2cmi.py, and cmi_changer
Right now we have several variants of CMI and L1B conversion for AHI. Look at reducing this to two, the "retro" hsd2nc and the "neo" cmi_changer; this requires cmi_changer have a radiance option like ahi2cmi (which was originally a test pattern going into cmi_changer development)