Closed djtfmartin closed 1 month ago
Parsing small 2 char epithets that resemble english words is hard to support. The year is also not within the brackets which might cause trouble, the GBIF parser struggles with the authorship too: https://api.gbif.org/v1/parser/name?name=Tethea%20or%20(Denis%20&%20Schiffermüller),%201776
For some manual tweaking the CLB parser can be manually configured for special cases like this. Basically you can configure expected results for specific names and authorships: https://api.checklistbank.org/parser/name/config
I pushed a config for this name to CLB: https://api.checklistbank.org/parser/name/config/Tethea%20or|(Denis%20&%20Schiffermüller),%201776
This works now: https://api.checklistbank.org/parser/name?q=Tethea%20or%20(Denis%20%26%20Schiffermüller),%201776 https://api.checklistbank.org/parser/name?q=Tethea%20or%20(Denis%20%26%20Schiffermüller,%201776) https://api.checklistbank.org/parser/name?q=Tethea%20or
well, if you use the parser outside of CLB we need to make sure to load the configs from CLB - which stores them in the db. Haven't paid much attention to that really
Parsing of
Tethea or (Denis & Schiffermüller), 1776
results in the specific epithet "or" being dropped and the name being interpreted as a genus.Parsed correctly in current GBIF API (older parser): https://api.gbif-uat.org/v1/species/match?name=Tethea or (Denis & Schiffermüller), 1776
CLB API (current parser): https://api.checklistbank.org/dataset/53147/match/nameusage?scientificName=Tethea%20or%20(Denis%20&%20Schifferm%C3%BCller),%201776
Debugging shows the Issue is at the parser level (as opposed to WS API impl).