IngestLDD problems remaining for Alice and LORRI additions

[acraugh]

• [x] There has been no addition to the comments at the top of the file indicating what was added to this dictionary and when. These need to be added.

The following definitions are problematic for the reasons given and should be improved before the next version of the dictionary is released:

• [x] aperture value "Airglow": The definition of this permissible value is circular, and the meaning of the word "Airglow" does not seem to be appropriate at all in the context of an aperture description. The shape and/or nature of the aperture needs to be described, and the permissible value itself should be reconsidered to provide some meaning to users who know nothing specific about the instrument or mission.
• [ ] approx_target_name/line/sample: "Approximate" seems like the wrong concept for these keywords. What is LORRI targeting that would lead to an observation of a different target? Is this actually indicating a "planned" target that might or might not have been acquired, or a pointing direction that was not successfully achieved? If these names are being held over from PDS3 for some reason, then that should be stated - but I don't see a lot of point in retaining deceptive keyword names as attributes in PDS4.
• [ ] approx_target_name: The definition says, "If the best guess at the target is unknown, the attribute will be UNK i.e. unknown. If there was no target, i.e. for some instances of instrument calibration activities, the attribute will be nilled." but this is inconsistent with PDS4 usage. If the target is unknown, it should be nilled with a reason code of "Unknown". In particular the string "UNK" should never be used in a PDS4 label.
• [ ] LORRI_Target_Information: Given that all the attributes in this class begin with "approx", the name of this class is misleading and the class definition is not accurate. Are additional attributes expected for the actual target?

[acraugh]

@cgobat and @benjhirsch: I need some sort of response on these issues. If this needs to go back to the science team, I need to know that ASAP.

[benjhirsch]

There's a PDS3 keyword called NEWHORIZONS:APPROX_TARGET_NAME which is defined in labels as "the best guess of the intended target of the observation made by the automated data processing pipeline." The dataset.cat file for LORRI expands on that, saying:

The downlink team on New Horizons has created an automated system to take various uplink products, decode things like Chebyshev polynomials in command sequences representing celestial body ephemerides for use on the spacecraft to control pointing, and infer from those data what the most likely intended target was at any time during the mission.

As it turns out, however, TARGET_NAME and NEWHORIZONS:APPROX_TARGET_NAME have identical values for all LORRI products (with the exception of a couple names that are formatted differently), so it's probably unnecessary to have the latter as an attribute in the PDS4 label.

The approx_target_line and approx_target_sample keywords do seem potentially useful, though, if they make it easier for users to find the target in an image. On the other hand, the definition of approx_target_line from the PDS3 data dictionary says:

Used in New Horizons (NH) mission instrument observations labels as an approximate line location in an image of the center of a target. This is only an estimate of the location and may be off by a large fraction, or multiple, of the size of the image, especially for flyby encounters.

So perhaps they're not that useful at all, except perhaps with those caveats spelled out. With the limitations identified here, I would be fine removing the whole class for now and possibly finding a way to implement it later if these keywords are thought to be important.