Updating realm format

[durack1]

In preparation for changes to source_id some tweaks to realm are planned. These changes will update the current CMIP6_realm.json, to:

{
    "realm": {
        "aerosol":"aerosol",
        "atmos":"atmosphere",
        "atmosChem":"atmospheric_chemistry",
        "land":"land_surface",
        "landIce":"land_ice",
        "ocean":"ocean",
        "ocnBgChem":"ocean_biogeochemistry",
        "seaIce":"sea_ice"
    },
    "version_metadata":{
        "author":"Paul J. Durack <durack1@llnl.gov>",
        "creation_date":"Wed Apr 12 18:44:20 2017 -0700",
        "institution_id":"PCMDI",
        "latest_tag_point":"3.2.2 (175; g0c97a12)",
        "note":"Add missing activity_id values",
        "previous_commit":"c50f25d4e3856590e867ba9b36773625a7b33ca3"
    }
}

Note that the key ocnBgchem will be updated to ocnBgChem.

@taylor13, @dnadeau4 pinging you here.

@ehogan, @jkettleb, @alistairsellar, @senesis, @mbittner1, @jypeter, @ARadhakrishnanAtGFDL, @martinjuckes please let us know if this planned change will cause any problems. Please forward on to others if you think they may be affected

[durack1]

@momipsl and @eguil pinging you both here - I am uncertain how much ES-DOCs is dependent on this repo

[jypeter]

LGTM !

Having a translation dictionary of the realm names somewhere seems like a useful thing and I don't use _CMIP6realm.json directly so I should not be affected by the change. As an end-user, I mostly use realm as some extra information I can tell other end users to look for in the DR, CMOR tables or metadata in the nc files. I'm not sure I have ever explicitly used the realm information in a script

To be safe, I have also just checked the realm related information in the recently updated CMIP6_global_attributes_filenames_CVs document

I expect the ocnBgchem=>ocnBgChem change will require an update of the DR

[asladeofgreen]

@durack1:

1. What is the rationale behind this change ?

2. Impact upon ES-DOC is limited as it reformats the WCRP CMIP6 vocabularies to a normalised format using pyessv, thus only the reformatting script will need a tweak.

[durack1]

@momipsl

1. There are changes planned for the table_id which expands on the CV just being table names and includes the associations with the data request, Omon:monthly etc, see #166. This change is being made to make the CV more readable, and then allows us to use just the shortened identifier (rather than the expanded entry as currently in the source_id)
2. Ok great, I assume you'd need to know about changes upstream however - I suppose you are "watching" this repo for changes?

[durack1]

@jypeter thanks for reviewing, yes we are further checking for consistency between the descriptive document CMIP6_global_attributes_filenames_CVs and these entries.. We've noted a number of inconsistencies and are making these changes to bring things more inline.

@martinjuckes has been pinged on this request, so we're hoping that the change (ocnBgchem=>ocnBgChem) will propagate quickly across to the dataRequest, and that any further clarifications can be completed quickly before we make the change in this repo

[durack1]

@martinjuckes

On 4/14/17, 1:48 AM, "martin.juckes" wrote:

    Hi Karl,

    could you use "Atmospheric Chemistry" etc instead of "atmospheric_chemistry" etc? The short terms "atmosChem" etc are suitable for using in code and elsewhere when spaces are inconvenient, so we could use natural word spacing in the descriptors.

    regards,
    Martin

[durack1]

All after reviewing this, it seems the ocnBgchem -> ocnBgChem change is spurious, as the full descriptor 'Ocean Biogeochemistry' does not lead to camelCase for the chemistry 'C' (it's part of the complete second word), so will implement the list -> dictionary change but not the camelCase change as suggested above.

[asladeofgreen]

@durack1: Yes ES-DOC is watching ... closely.