plazi / treatmentBank

Repository devoted to house keeping of treatmentBank
0 stars 0 forks source link

Zookeys stubs #118

Open myrmoteras opened 1 month ago

myrmoteras commented 1 month ago

@gsautter these ZooKeys articles have created stubs, and many of the treatments have no article DOI. Why?

https://tb.plazi.org/GgServer/srsStats/stats?outputFields=doc.uuid+doc.updateDate+lnk.zenodoDepId+lnk.gbifTaxonId+bib.source+pubLnk.articleDoi&groupingFields=doc.uuid+doc.updateDate+lnk.zenodoDepId+lnk.gbifTaxonId+bib.source+pubLnk.articleDoi&orderingFields=-lnk.zenodoDepId&limit=100&FP-bib.source=ZooKeys&format=HTML

e.g. https://tb.plazi.org/GgServer/html/0074F87B464364F7CA9E8404A20A37D7

gsautter commented 1 month ago

If you look at the "Timestamp of Last Update", these treatments have all remained unchanged since 2015 ... sorting by "Zenodo Deposition ID" (which is an integer field) in descending order will actually give you the lowest values first ... unintuitive, admittedly, but remember the legacy of the stats, namely their original purpose to populate dashboards, and for those you always need the high values at the top, not the long tail, and hence this inversion ... works great for counts and aggregate numbers, where you're hardly ever after the long tail, but is a bit of a nuisance with integer IDs ... here the other way around: https://tb.plazi.org/GgServer/srsStats/stats?outputFields=doc.uuid+doc.name+doc.uploadUser+doc.uploadDate+doc.updateUser+doc.updateDate+lnk.zenodoDepId+lnk.gbifTaxonId+bib.source+pubLnk.articleDoi&groupingFields=doc.uuid+doc.name+doc.uploadUser+doc.uploadDate+doc.updateUser+doc.updateDate+lnk.zenodoDepId+lnk.gbifTaxonId+bib.source+pubLnk.articleDoi&orderingFields=lnk.zenodoDepId&limit=100&FP-bib.source=ZooKeys&format=HTML

Regarding the ZooBank stubs, those do get replaced when the actual treatment comes in ... provided it has the same ZooBank LSID attached to it ... and to prevent typos (we've had these occasionally), there is a verification lookup to ZooBank before the LSIF is actually used ... now if that verification lookup fails because ZooBank happens to time out or be simply unreachable (which I've been seeing a good bit in the logs over the past decade, and still currently see as well), the stub doesn't get replaced because the LSID is unconfirmed ... that said, the stubs prevailing is simply down to ZooBank outages on the arrival or the actual treatments, and this is by no means specific to Pensoft imports.