Layer that is a data.frame

[maelle]

library("EMODnetWFS")
wfs <- emodnet_init_wfs_client(service = "biology_occurrence_data")
#> Loading ISO 19139 XML schemas...
#> Loading ISO 19115 codelists...
#> Loading IANA mime types...
#> No encoding supplied: defaulting to UTF-8.
#> ✓ WFS client created succesfully
#> ℹ Service: 'http://geo.vliz.be/geoserver/Dataportal/wfs'
#> ℹ Version: '2.0.0'
simple_filter_sf <- emodnet_get_layers(
    wfs = wfs,
    layers = "taxa",
    cql_filter = "scientificname='Abra'",
    reduce_layers = TRUE
)
#> Warning: Download of layer 'taxa' failed: Error in UseMethod("st_crs<-"): pas de méthode pour 'st_crs<-' applicable pour un objet de classe "data.frame"

^{Created on 2022-03-02 by the reprex package (v2.0.1)}

[maelle]

@annakrystalli for a change would you like to try running this one? 😅

[annakrystalli]

Same error for me in http-tests branch.

I think it's because the package does some post processing to the crs that fails if the response is not an sf

[maelle]

the layer type in the layer medata from wfs info was sf, should it indicate something else?

[maelle]

I see "sf" is hard-coded, maybe I can get it from the web service 👀

[maelle]

I modified ows4R code

https://github.com/eblondel/ows4R/blob/8068e7284e3c0c99f54457579602bc33174f134d/R/WFSFeatureType.R#L328

to quiet = FALSE

I get

no simple feature geometries present: returning a data.frame or tbl_df

[maelle]

I wonder whether we should expect more metadata before the download.

[maelle]

And when we do get a data.frame, the layers can obviously NOT be reduced.

[maelle]

So, TODOS

• Find more info on data.frame layers in WFS in general.
• Tweak the code so that we simply return the data.frame when that's what was returned by the web service.

[annakrystalli]

And when we do get a data.frame, the layers can obviously NOT be reduced.

That's ok because the function only tries to reduce and if it can't it returns a list

[maelle]

https://geo.vliz.be/geoserver/Dataportal/wfs?SERVICE=WFS&REQUEST=DescribeFeatureType&VERSION=2.0.0 does not have info that it's a table

[maelle]

Related eblondel/ows4R#65

[maelle]

For any layer if we get its description and see no element is a geometry, we can know it will be a data.frame rather than an sf object.

[salvafern]

I think @maelle 's approach in eblondel/ows4R#65 (comment) would be great. There are layers in there with no geometry as they are required or used in the infrastructure of EMODnet-Biology, so we cannot really get rid of them.

In any case I will check why some layers don't have a geometry. For instance Dataportal:eurobis-obisenv_basic. Probably someone just forgot as the important fields when exporting to csv are decimalLongitude and decimalLatitude.

[maelle]

@salvafern #83 is meant to tackle this.

[salvafern]

Update: we should not change Dataportal:eurobis-obisenv_basic in geoserver as many other system are relying on this layer. We should go forward with #83 . @maelle are you following this up or shall I have a look?

[maelle]

@salvafern My main problem in #83 is the size of fixtures. So if you can have a look at the functionality, that'd be great, I can keep thinking about the testing.