Key "http://rs.tdwg.org/dwc/terms/geodeticDatum" in occurrence API
15628
Reporter: rdmpage
Assignee: mdoering
Type: Bug
Summary: Key "http://rs.tdwg.org/dwc/terms/geodeticDatum" in occurrence API
Priority: Major
Resolution: Fixed
Status: Resolved
Created: 2014-05-20 11:51:44.863
Updated: 2014-05-23 15:07:16.946
Resolved: 2014-05-23 15:02:48.331
Description: Some occurrence records retrieved using the API have the key "http://rs.tdwg.org/dwc/terms/geodeticDatum, for example, see http://api.gbif.org/v0.9/occurrence/477903170
Every other key is a simp string, why is this a URI? Furthermore, when it occurs there is already a "geodeticDatum" key, which often contradicts the value in http://rs.tdwg.org/dwc/terms/geodeticDatum (GBIF seems to say that every occurrence, regardless of whether it is georeferenced or not, is WGS84). For example, http://api.gbif.org/v0.9/occurrence/686418295 has both WSG84 and NAD27 datums!
Might be an idea to suppress "geodeticDatum":"WGS84" unless record has been georeferenced, and to respect http://rs.tdwg.org/dwc/terms/geodeticDatum if that contradicts WGS84 (and to convert the URI http://rs.tdwg.org/dwc/terms/geodeticDatum to a string key).]]>
Author: mdoering@gbif.org
Created: 2014-05-20 12:04:38.84
Updated: 2014-05-20 12:04:38.84
we apply WGS84 as a static value right now for all records as we intended to do the reprojection - which we never did :( .
Ideally we should:
- show WGS84 in case we have coordinates (small API getter change)
- do the reprojection during interpretation (Id be real happy to implement that now, the basics are real simple)
- surpress the verbatim datum in the interpreted version (bug in our current code)
Alternatively we could just show the verbatim one and do nothing about the datum.
Author: rdmpage
Comment: I'd be in favour of keeping the verbatim datum somewhere in the record so that people can see that a different datum was used in original record. Would be handy if people are trying to reconcile GBIF data with, say, value reported in original publication.
Created: 2014-05-20 12:07:14.193
Updated: 2014-05-20 12:07:14.193
Author: mdoering@gbif.org
Created: 2014-05-20 12:18:16.193
Updated: 2014-05-20 12:19:18.685
[~rdmpage], we keep all verbatim values, but we have 2 distinct versions of a record (actually 3, we also keep the raw fragment as it came in). The interpreted record and the verbatim one. We decided not to lump them together into a single representation and remove the verbatim values from the interpreted version in case we used them for interpretation. This has been done primarily to avoid utter confusion if various term values contradict each other.
For example here is the interpreted and the verbatim version:
http://api.gbif.org/v0.9/occurrence/477903170
http://api.gbif.org/v0.9/occurrence/477903170/verbatim
Author: mdoering@gbif.org
Comment: For geodeticDatum interpretation see http://dev.gbif.org/wiki/display/DEV/Growing+Occurrence+Width
Created: 2014-05-20 12:43:29.198
Updated: 2014-05-20 12:43:29.198
Author: mdoering@gbif.org
Comment: show datum only with coord: https://github.com/gbif/gbif-api/commit/066d933116182d5c4e3bd9f7102995df92f150c8
Created: 2014-05-23 15:07:04.321
Updated: 2014-05-23 15:07:04.321
Author: mdoering@gbif.org
Comment: dont show verbatim datum in interpreted record: https://github.com/gbif/occurrence/commit/4ab9b135ca48e995194ba7872707c38f185ed6d7
Created: 2014-05-23 15:07:16.946
Updated: 2014-05-23 15:07:16.946