Add releaseTime to NCEP ADPUPA converter

[haydenlj]

Description

The variable releaseTime is required by Skylab, however, IODA files produced by the NCEP ADPUPA converter do not contain this information. The variable should be added to the converter output

Requirements

Determine release time from the prepbufr data

Acceptance Criteria (Definition of Done)

code outputs ioda file with the release time included and the file is accepted by Skylab

Dependencies

none

[haydenlj]

@emilyhcliu @PraveenKumar-NOAA @gthompsnJCSDA @rmclaren
How can the release time be determined in the prepburf files? The code in bufr_ncep_prepbufr_adpupa.py seems to be expecting a 2 dimensional array (e.g. line 110 num_locs = lat.shape[0]), however, when I print the shape, the array is 1 dimensional. Is there a way to determine where each profile starts so that we can take the time at that point as the release time?

[rmclaren]

The reason that you get 1 dimensional arrays here is because of the group_by fields specified in the call to the getter
r.get('latitude', 'prepbufrDataLevelCategory'). If you want to see the original shape you could do r.get('latitude').shape.

It might be that you might not want to apply group_by fields to anything, as this will have the effect of flattening the arrays (Remove/rotate away the PRSLEVEL dimension). So you could just leave out the second argument from the r.get('example') calls. In this case each "row" in the resulting data set will correlate to a specific PRSLEVEL. It all depends on what you want to do with the data....

[PraveenKumar-NOAA]

@haydenlj @huishao-r I am working on adding this to the ADPUPA BUFR Python API (#1298 (comment), NOAA-EMC/JEDI-T2O#78), but releaseTime/receiptTime can't be added to the existing YAML converter as it belongs to a different branch which does not follow group_by variable path.