Open seltmann opened 1 year ago
Thanks for sharing your detailed example of unexpected behavior of finding images for specific type specimen.
I ran:
./find-images.sh CASTYPE1503
and zipped up the dist folder and attached the results.
And yes, even for an untrained eye like mine, the included images do look like a mix of different specimen. (see attached)
A at a quick glance, the find-images.sh currently selects images from catalog number that start with the provided catalogNumber.
So, CASTYPE1503 would match CASTYPE1503 specimen data, but also, CASTYPE15037. This is supported by labels included in earlier comments. Am working on a fix.
After applying a fix to make the catalog number selector a little less lenient, I generated the following results via
./find-images.sh CASTYPE1503
Associated images include:
Please confirm that this addressed your concern.
Thanks again for providing such specific feedback!
@jhpoelen we are making great progress measuring bees downloaded using Preston and the bee-image-finder. There is a curious issue, where multiple specimens are being downloaded together. Some of them are 2 very different taxa (bee and leafhopper). See our working spreadsheet at:
https://docs.google.com/spreadsheets/d/1Kx_aWF5Sz-o62j3Utxp5rfzb6bTaT4my6GbTQbpXnD8/edit?usp=sharing
Folder CASTYPE1503 is a good example. It includes a bee, a spider and a beetle. I would not think these all have the same catalog number.