Ungrib driver is hardcoded to GFS

[jderrico-noaa]

Expected behavior
The ungrib driver can ungrib files from different sources, e.g. RAP, RRFS, HRRR, etc

Current behavior
The driver is hardcoded to GFS, and returns errors when called with a different data source

Machines affected

To Reproduce
experiment directory using RAP ICs for mpas on Jet:
/lfs5/BMC/wrfruc/jderrico/mpas/dev/conus_rt/ush/conus_test

[maddenp-noaa]

It looks like the config YAML key ungrib.gfs_files was named short-sightedly and should be generalized (and similarly some variable names in the driver code, though users wouldn't normally see that), but looking at /lfs5/BMC/wrfruc/jderrico/mpas/dev/conus_rt/ush/conus_test/log/ungrib_2024123000.log, I see

[2025-01-23T22:36:25]  WARNING File /lfs5/BMC/wrfruc/jderrico/mpas/dev/conus_rt/ush/conus_test/2024123000/RAP/rap.t00z.pgrb2.0p25.f000: State: Not Ready (external asset)

while the files in that directory are actually named

$ ls -l /lfs5/BMC/wrfruc/jderrico/mpas/dev/conus_rt/ush/conus_test/2024123000/RAP/
total 905552
-rw-r--r-- 1 Janet.M.Derrico wrfruc 302067750 Dec 30 01:23 rap.t00z.wrfnatf00.grib2
-rw-r--r-- 1 Janet.M.Derrico wrfruc 311164120 Dec 30 01:24 rap.t00z.wrfnatf06.grib2
-rw-r--r-- 1 Janet.M.Derrico wrfruc 314035224 Dec 30 01:26 rap.t00z.wrfnatf12.grib2

The ungrib: section in /lfs5/BMC/wrfruc/jderrico/mpas/dev/conus_rt/ush/conus_test/experiment.yaml has

    gfs_files:
      ...
      path: '{{ user.experiment_dir }}/{{ cycle.strftime("%Y%m%d%H") }}/{{ get_lbcs_data.external_model }}/rap.t{cycle_hour:02d}z.pgrb2.0p25.f{forecast_hour:03d}'

Would changing that path: value to

{{ user.experiment_dir }}/{{ cycle.strftime("%Y%m%d%H") }}/{{ get_lbcs_data.external_model }}/rap.t{cycle_hour:02d}z.wrfnatf{forecast_hour:02d}.grib2'

be a quick fix? We should still generalize the YAML key name and driver code, but would the driver make progress with a path: value matching the actual filenames?

[maddenp-noaa]

I'm about to merge #688 to generalize what might have been misleading apparent specificity about ungrib and GFS but still don't see a problem not under user control to fix, as described above. But please let me know if I missed something and I'll address it with a separate PR.