The goal of this project is to understand how accurately Plazi is marking up material citations in documents. Parsing specimen information in publications is a challenging problem. Ideally we would have a way of checking whether the extracted data is correct. In the apparent absence of such tools I've created this web site.
.This site takes a Plazi treatment in XML form and displays the XML as a marked up web page. In other words, it displays the tags assigned to the text by Plazi. These tags are colour-coded, and there is a tooltip to show the corresponding tag name. The interpretation of the markup by Plazi is shown as a table of key-value pairs.
To use the web site simply paste in the Plazi treatment id (typically a UUID) and the site will display the marked up document. Below I have listed some examples where Plazi appears to have failed to correctly extract specimen data.
| Treatment ID | Issue |
|---|---|
| 74CF9B715E205704A730270742F139CA | Misses specimen |
| 9319C4492C272A7EFF288BEBFC1849AB | Splits one specimen into two |
| 03F33310DA38FF875DB5FDF3FC2AF8CD | Specimen code interpreted as GenBank accession number |
| 03FC87E61268FFD6D3E36CD2FE12DF29 | Cruise interpreted as collector, depth treated as elevation |
| DD4987C1FF96247CFF3DFC385F30FE6F | Collector treated as locality, datum WGS84 interpreted as speicmen code |
| 704ECD36FFF12173FD03EF56CEA6D97D | Specimen split, only first part parsed |
| 023487907016FF9EDAC8FF48FAC1DA10 | Specimen split in two |
| C51DEB611E4C5421FF13F9ABA73729F7 | Failed to parse specimen citation |
| 03E787D0FFB0FFDA7B8BD57CB2FE779E | Station treated as colelctor, depth treated as elevation |
| 03EB1A42265CFFE3FF1EF9B71C5AF92C | Missed most information, confused collectors with places |
| 6FE5DF2540375B7184495C6134814F7F | Biodiversity Data Journal paper with structured data misinterpeted (origjnal paper doi:10.3897/BDJ.9.e64505) |
| 252C87918B362C05FF20F8C5BFCB3D4E | Collector initial interpeted as collection code, specimen code treated as accession number |
| 6CA445F891624F8F86D89A7580F879D4 | No specimens found|
| 71F7952EADCA5025A7A45AE1C2811C3D | No specimens found |
| 03FCA61B9A735210FF52F99DFCC38B27 | Missed lat lon |
| 03A087FA2319FF86FF3D64CDFD9BBABC | Split one record into two, missed lat lon, duplicates in GBIF |
| 03936656F63EE0DE7AB0FAF8FDE1F93B | Multiple species run together |
| 03EF1D76C8269652FF2FE282FDA10C3D | Museum conflated with locality (Germany versus Ecuador) |
| 5244B05E-FFC8-E20F-7BC3-2056C178F496 | Type specimen for Bulbophyllum hamatum split into two specimens. |
| 0D228E6AFFC2FFEFFF4DE8118C4EE6B9 | coordinates not parsed correctly (missed the degrees) |
A version of the Plazi material citation parser is available as a web service. I have constructed a simple wrapper around that service so that you can paste in a specimen citation and see how Plazi interprets it.