search

USGS-CMG / usgs-cmg-portal

do_convert.sh
6 stars 13 forks source link

what value goes into the `id` field for model output #307

Open zdefne-usgs opened 5 years ago

zdefne-usgs commented 5 years ago

@rsignell-usgs

How id is used with the model output: In the yaml file to generate metadata for model output we use assign a value to id:

dataset:
    id: "USGS_COAWST_MVCO_CBLAST_Ripples_SWAN_40m"

    title: "USGS-CMG-COAWST Model: CBLAST2007 Ripples with SWAN-40m res"

, which then gets transferred to the ncml 

<ncml:attribute name="id" value="USGS_COAWST_MVCO_CBLAST_Ripples_SWAN_40m"/>

, and to the ISO xml as (after prefixing the domain)

<gmd:fileIdentifier>
<gco:CharacterString>gov.usgs.cmgp.whsc:USGS_COAWST_MVCO_CBLAST_Ripples_SWAN_40m</gco:CharacterString>
</gmd:fileIdentifier>

How id field is used for measurements: For the measurement it looks like only the data filename is used ( #289):

global:id - A unique identifier for this file. Historically set to the filename since that is already identifiable and recognizable (9651hwlb-a, 10831dw-a, 9431rb-a, etc.). This is used in the portal as a discriminant to distinguish variables of the same type being measured differently at the same mooring.

1) Should we try to use the same approach with the model ids to be consistent with the naming of measurement ids? 2) Is the use of only the file name going to be problematic in the case of two identical file names in different folders? 3) In thatcase should we consider using the full path and the file name to make sure it is unique?