factsmission / synospecies

Using Plazi Data to find currently accepted scientific names
https://synospecies.plazi.org/
MIT License
5 stars 1 forks source link

changed UI? #129

Closed myrmoteras closed 5 months ago

myrmoteras commented 5 months ago

@retog @nleanba

the UI seems to have changed and the original timeline with the green, blue, red dots disappeared. why?

image

https://synospecies.plazi.org/#Pipistrellus+pipistrellus

nleanba commented 5 months ago

The timeline was not removed, but here it has been pushed to the side and squished to an invisible width of 0px...

It has been pushed to the side as an unfortunate consequence of displaying all given authorities verbatim (the light gray text).

Synospecies should maybe cut the authorities off here to keep the timeline visible;

but this suggest an underlying data problem: this means that the treatments all assign a (subtly) different authority to the taxon concept (many of which wrongly (?) contain the species name again)

nleanba commented 5 months ago

Displaying the verbatim authorities was supposed to improve the situation with regards to https://github.com/plazi/names_LOD/issues/154

nleanba commented 5 months ago

@retog after quick investigation I found that some of these wrong authorities are no longer in treatments-rdf, but that change seems not to have been picked up by treatment.ld.plazi.org

(compare also the results of this search with the Lindas endpoints, which are more up to date it seems)

can you figure out how to efficiently reimport all data into treatment.ld.plazi.org?

retog commented 5 months ago

@retog after quick investigation I found that some of these wrong authorities are no longer in treatments-rdf, but that change seems not to have been picked up by treatment.ld.plazi.org

(compare also the results of this search with the Lindas endpoints, which are more up to date it seems)

can you figure out how to efficiently reimport all data into treatment.ld.plazi.org?

This should be feasible using the SARQL GraphStore protocol, anallogously to how this is done for lindas:

#!/usr/bin/env bash
set -eo pipefail

set +u
ENDPOINT=${ENDPOINT:=https://stardog-test.cluster.ldbar.ch/lindas}
set -u

cat input/col.nt > upload.nt
cat input/data.nt >> upload.nt

echo "Posting to endpoint: $ENDPOINT"
curl -n \
     --netrc-file config/dot.netrc \
     -X PUT \
     --http1.1 \
     -H Content-Type:application/n-triples \
     -T upload.nt \
     -G $ENDPOINT \
     --data-urlencode graph=https://lindas.admin.ch/foen/plazi

echo "Executing post-insert SPARQL on endpoint: $ENDPOINT/update"
curl -D - -n \
     --netrc-file config/dot.netrc \
     -X POST \
     --http1.1 \
     -H Content-Type:application/sparql-update \
     -T scripts/post-insert.sparql \
     -G $ENDPOINT/update
nleanba commented 5 months ago

UI is fixed to always have some part of the timeline visible (and all of it reachable by scrolling horizontally), truncating the authority if necessary (full authority is visible below the timeline either way).