search

iodepo / oih-ui

source code for the Ocean InfoHub (OIH) website
https://oceaninfohub.org/
MIT License
5 stars 4 forks source link

handle WOD partner records #47

Open jmckenna opened 1 year ago

jmckenna commented 1 year ago
jmckenna commented 4 months ago

@emarzini @pbuttigieg @Marco-C97

I fixed a big issue with coordinates for the WOD records, it was a coordinate swapping issue. The WOD sitemap & JSON-LD have been fixed. For example, here is how one cast, wod_osd_1915.nc looks in the current TrustIT instance, versus in the OIH spatial search results (with the coordinate fix) (see staging server which currently only has the WOD data in Solr) It is the elongated circle geometry.

Current view of wod_osd_1915.nc

1915-old

Updated view of wod_osd_1915.nc

1915-new

More background

There are a lot of -999 coordinates and empty (blank) coordinates inside the raw NetCDF cast files hosted by WOD, causing many to appear as 'global' geometries etc. It will really come down to the WOD team fixing their source metadata.

related to https://github.com/iodepo/odis-arch/issues/126

jmckenna commented 4 months ago

For example, this shows all of the WOD records, in the old interface - notice the deep blue from all of the so-called 'global' geometry records, overtaking the map. If we want a positive, it is that now the points do look like valid cast locations.

global-issue

jmckenna commented 4 months ago

In summary, the WOD cast metadata for geometry is so problematic, that it really affects our interface. Should our UI interface ignore any 'global' reocrds? Not sure.

pbuttigieg commented 4 months ago

Thanks @jmckenna - please tag the WOD team here and ping them via email with Lucy and I in CC

pbuttigieg commented 4 months ago

I think we should ignore "global" records - this is not accurate and gunks up the interface.

We should perhaps have a validation step to detect and throw warnings on those @fils