Panopoda interpreted as Noctuidae

18378
Reporter: donald hobern
Type: Feedback
Summary: Panopoda interpreted as Noctuidae
Status: Open
Created: 2016-04-07 09:30:10.663
Updated: 2016-04-07 10:39:05.796
        
        
Description: See: http://www.gbif-uat.org/occurrence/search?HAS_GEOSPATIAL_ISSUE=false&GEOMETRY=-73.12%20-68.08%2C-73.12%20-60.53%2C-45.57%20-60.53%2C-45.57%20-68.08%2C-73.12%20-68.08&TAXON_KEY=7015

These are not insects ...

*Reporter*: Donald Hobern
*E-mail*: [mailto:dhobern]]]>
    

Author: rdmpage
Comment: It's a fuzzy match problem, verbatim scientificName "Pantopoda and Crustacea species" has been mapped to the genus _Panopoda_. Perhaps we need to filter names that are not properly formed before attempting fuzzy match? 
Created: 2016-04-07 09:52:23.433
Updated: 2016-04-07 09:52:23.433

Author: mdoering@gbif.org
Created: 2016-04-07 10:38:58.214
Updated: 2016-04-07 10:38:58.214
        
We did not match those in the old index, but I assume this has not happened because they were interpreted with an older code.
http://www.gbif.org/occurrence/344675417

The current matching service even in the old portal does link them to Panopoda:
http://api.gbif-uat.org/v1/species/match?kingdom=&family=&name=Pantopoda%20and%20Crustacea%20species

It boils down to 2 things, a fuzzy match of the genus Pantopoda to Panopoda and before that a cleaned and parsed scientific name resulting in "Pantopoda spec.": http://api.gbif.org/v1/parser/name?name=Pantopoda%20and%20Crustacea%20species

Detecting not well formed names isn't that simple unfortunately - unless we want to be very strict and reject all sorts of names we get. Case for example cannot be used as we see all upper case names. But "and" could be a good hint in this case