More than the simple Eucalyptus test included in this PR, I did a new tool inside the la-toolkit to compare some random occurrences in a LA portal against the equivalent occurrences in GBIF.org.
Testing ~400 occurrences of our ~50M records with an ala-namematching-service without patching:
and with an ala-namematching-service using this patch:
We discovered recently that some species are incorrectly matched when the
rank
is filled in the occurrence and correctly matched when it's not.More info in: https://atlaslivingaustralia.slack.com/archives/CCTFGEU1G/p1719471737011039
Also described in: https://github.com/AtlasOfLivingAustralia/Taxonomic-Issues-Register_new/issues/161
More than the simple
Eucalyptus
test included in this PR, I did a new tool inside thela-toolkit
to compare some random occurrences in a LA portal against the equivalent occurrences in GBIF.org.Testing ~400 occurrences of our ~50M records with an
ala-namematching-service
without patching:and with an
ala-namematching-service
using this patch:And the last report of issues:
https://datos.gbif.es/others/202407120244_la_gbif_comparative_issues_report.html
As you can see that the number of differences and nulls in the scientificNames are significantly lower.
This comparative tool will help us also to debug aligning issues in
pipelines
.