Update master from gsd/develop 2020/06/30 #307
Conversation
…from new to old metadata
gsd/develop: update from master 2020/01/27
Bugfix for converting optional attribute from new metadata for ccpp_prebuild.py
gsd/develop: update from NCAR master
…CCPP_VARIABLES_*.pdf)
…ersions K<->C and g kg-1 <-> kg kg-1 scripts/common.py: additional definitions required for handling blocked data structures
…iminary name) from new metadata; remove support for old metadata standard; add legacy extension code to handle inconsistent use of horizontal_dimension versus horizontal_loop_extent
…and 'active' (preliminary name)
…n _init and _finalize routines, support for handling conditionally allocated/used variables when converting blocked data structures or units
…uctures gsd/develop: handle blocked data structures in init and finalize routines, remove pset logic
…er_20200616 Update gsd/develop from master 2020/06/16
|
All, we are hoping to merge this tonight. Given that you reviewed these changes when they were merged into gsd/develop, I would go ahead without approval if needed. |
| " with 'horizontal_dimension' in table {}".format(table_name)) | ||
| standard_name = 'horizontal_dimension' | ||
| legacy_note = ' replaced by horizontal dimension (legacy extension)' | ||
| elif standard_name == 'horizontal_dimension' and scheme_name and table_name.endswith("_run"): |
There was a problem hiding this comment.
Since this is being handled in the CCPP framework here, do we not need to clean up the scheme metadata? Or should that still be done to avoid future confusion (and to set the right example for future developers)?
There was a problem hiding this comment.
Correct, this is a temporary workaround - we need to do a full cleanup of the metadata in GFS_typedefs.F90 and other places. I didn't want to do this as part of this commit, because it would have required a huge amount of changes on top of all that is in the current PRs already. Making this metadata change should be a self-contained PR that does nothing but to remove the workaround and correct the metadata in ccpp-physics and fv3atm.
There was a problem hiding this comment.
I don't feel super qualified to approve this, but I did look it over. It seems like the most "interesting" change is in commit 1975041 which I still need to digest. I'm sure that it could benefit from some more documentation on the implications of these changes, but I don't want to get in the way of this either, especially if this is made obsolete by capgen?
Yes, this will be taken over by cap_gen hopefully rather sooner than later. |
…elop_20200630 Update master from gsd/develop 2020/06/30
Changes in this PR:
_initand_finalizeroutineshorizontal_dimensionversushorizontal_loop_extentThese changes were reviewed by all reviewers on this PR when merged originally into gsd/develop.
Associated PRs:
NCAR/ccpp-physics#465
#307
NOAA-EMC/GFDL_atmos_cubed_sphere#30
NOAA-EMC/fv3atm#134
ufs-community/ufs-weather-model#156
For regression testing information, see ufs-community/ufs-weather-model#156.