convert table into minitreatments: concept

[myrmoteras]

Example 1 https://tb.plazi.org/GgServer/tables/DF3752A9F34EFFE1178A4FC66A43FF24?format=XML in FFD8CB4FF353FFFC170B4F6B687FFFC0

Dijkstra, Klaas-Douwe B. & Kalkman, Vincent J. & Dow, Rory A. | Redefining the damselfly families: a comprehensive molecular phylogeny of Zygoptera (Odonata) | 2014

does not show up directly in GGI

alternatively Example 2

has the advantage that it shows up in GGI

• at the moment, this needs to be marked manually, but could be trained, for this we need to build a gold standard
• needs table QC - this needs an analysis of existing tables
• when import into TB then when the header has taxon.header.row then this could be imported as treatment.
• each row shows up as individual treatment Should we ask BDJ how they format tables in this respect? Do they have similar tables.

What we need:

• Exemplar data
• Semantics for the table columns
• QC criteria
• get "nexus" file for phylogenetic tree

[gsautter]

• each row shows up as individual treatment Should we ask BDJ how they format tables in this respect? Do they have similar tables.

Actually, if multiple subsequent rows have the same taxon name, I fancy bundling the associated materials data into one treatment with multiple materials citations ... but that sure is a detail.

[myrmoteras]

Table examples

• https://tb.plazi.org/GgServer/summary/D16CFFF4D47E293CFFBA697BFFEBFFCB includes synoymy in tables (Tables 8-1, explained in the text

• tables as treatments https://tb.plazi.org/GgServer/summary/FFBCFFEEEA1BFFB18E69FFE1FFD5FF96 https://tb.plazi.org/GgServer/html/03858796EA19FF998DEAFE40FBBBF8E6

[myrmoteras]

Table example

https://tb.plazi.org/GgServer/summary/2F6EFFA8FF9A1571EE5EFF83FFFDFFCA including synonyms

[myrmoteras]

define in <taxonRecordFieldHeader type="matCit.collectionCode">Repository</taxonRecordFieldHeader> the type that can occur in tables.

https://tb.plazi.org/GgServer/tables/DF3752A9F34EFFE1178A4FC66A43FF24?format=XML

[myrmoteras]

table add taxonRecordFieldHeader in the table

import into SRS

the gadgets extracts the taxonRecordTreatments in SRS, which has the type "record", parallel to stubs, treatment

a record treatment then is also exported to GBIF

needs additional tables to make sure the bandwidth of possibliities is covered.

[gsautter]

needs additional tables to make sure the bandwidth of possibliities is covered.

In particular, we need to collect a sufficient number of examples before we can think about automating (a) recognition of tables whose rows are "record" treatments and (b) assigning the specific field type to columns.

Also, we'll need to establish proper QC for table structuring before we can automate this, as otherwise broken records might end up going to GBIF, etc.

[gsautter]

UUIDs of underlying table elements loop through to tables now (see https://tb.plazi.org/GgServer/tables/DF3752A9F34EFFE1178A4FC66A43FF24?format=XML), an essential building block for keeping UUIDs of treatments and materials citations stable across multiple generations, even through the exact UUID generation procedure is yet to be defined.

[myrmoteras]

Additional tables:

file name	article UUID	table UUID	taxonomic name	accession code	specimen code	verbatim specimen	linked to phylogeny	reference	muliplte specimen of he same taxon
RBZ-2021-0007			X	X		X	X
studiesInMycology.99.1.100117-100117			X	X	X	X	X	X	X