I don’t know if this is a “Bug Report” or a “Feature Request”. I would like some way to prevent the uploader from autofilling identifications from keywords associated with each image. See the attached screen captures which show what is happening.
Note that after loading metadata, the uploader has autofilled in a moth family name or an insect order for each observation in the identification box. This necessarily comes from my keywords associated with each image. (For instance, in the first screen capture, these images have my keywords of “Tortricidae”, “Geometridae”, and “Crambidae” (twice) associated with the four moth images.) Prior to uploading, I typically fill in a family name (for Lepidoptera) or an insect order as a keyword for every image in iPhotos to help me sort my 100,000 images there. The uploader is detecting that keyword and placing it (against my desires) as a suggested ID. I understand that this might be useful for observers who use a species name (common or scientific) as a keyword in their local photo program but I don’t. The outcome of the uploader autofilling from my keywords is that I have to make an extra motion for every observation of selecting and deleting the family/order name in the species ID box before I can begin typing in the actual name (which then requires reselecting the box after deleting the autofilled name). I’d rather the species ID box start off blank so that I can just click in it once to either start typing my ID or get an ID suggestion from computer vision.
Is the autofill of a keyword as a suggested ID a purposeful feature of the uploader or is this a bug?
I use iPhotos 9.6.1 (910.42) […and yes I still use iPhotos instead of Photos on my Mac] running Mojave 10.14.5.
Well, I dont think it is a bug, in the sense of it is unexpected behaviour from the computer code. It is doing what it is designed to do, which is automatically assign keywords or tags in imported photos to applicable locations in the iNat data model.
If it is an appropriate design choice, or the options associated with it can be enhanced is a separate discussion.
Perhaps some clarity from the devs about how it decides which keyword to import in cases where there are multiple potential matches to choose from would help.
I usually enter family (or some other high level taxon) and species to my keywords. The uploader seems to consistently choose the finest taxon level that I provide.
I use Lightroom rather than iPhoto, which probably does not matter as far as keywords getting to the uploader.
So, if I understand, because of the way iNat works with keywords in metadata, all of your observation taxon fields are getting populated with names that you don’t want there, and you are finding it tiresome to go through and click the little X at the right of each taxon field to clear them all.
If you click ‘select all’ you can edit all the observations at once from the panel on the left. Just use this bulk-editing feature to set all your observations to the same taxon (any taxon you like, it doesn’t matter), and then the little X will appear in the taxon field of the bulk-editing box. Click this, and all of the taxon fields of all of your observations will be blanked in a single click. Is that what you wanted to achieve?
Daniel, I think I understand what you described. IF I am only clearing one species ID from a single observation, then the X is available to clear it. However, if multiple family names have been autofilled (the most common case), then the bulk-editing box says “Edit multiple species”, which as you suggest then requires that I change them all to the same thing, then delete them all with the X. Perhaps that solution saves me a several key strokes overall, but I would rather see the more fundamental solution implemented by leaving the species ID’s blank without autofill.
A related question: In what percent of all observations (and what portion of all observers) does this autofill of an ID from keywords correctly tap into the desired ID? I suspect VERY very few! So, if that assertion is correct, then the feature is creating more work than it obviates.
Still hoping to hear from the dev team if this is the intended behavior and why.
Chris, I was trying to change the topic category from “Bug Report” to “Feature Request” but I can’t seem to accomplish that. Does that require a Forum Moderator to do that?
For those of us submitting thousands of observations a year from photos, having to re-enter thousands of IDs again that are already given in the photo tags would cost us many hours of repetitive work for no gain. The present behavior is strongly desired, to prevent having to do that.
So I’d hope this feature request would at most be to add an account setting to disable reading IDs from photo tags for a single individual only.
I’d put the success rate around 90% for correctly reading IDs from photo tags in my observations, heavy in insects and birds from the United States, mostly a mix of English species names and scientific names of higher ranks.
I am sure it is an intended feature. Many people, myself included, write the taxon in the description field of the photo metadata specifically so that iNat will pick that up along with the location and date information automatically. I have never had iNat fail to pick up the correct ID except when I have made a spelling mistake. The issue here is that you are using that metadata field for something else that conflicts with what iNat intended. As @treichard says, the solution, if your issue is also affecting others, would be to have some user-level way of disabling this function in the settings. Turning it off for everyone would, I suggest, cause a lot of people a lot of extra work.
I find myself in full agreement with @treichard and @DanielAustin. My rant seems to represent the intersection of a useful feature in the upload process conflicting with the particular handling of my keywords in Apple’s iPhotos. There are finite (and clunky) limitations of how iPhotos handles key words. IF I were to begin putting the species names (scientific or English) as key words in iPhotos, it would quickly bog down and become unwieldy. My work-around has been to place the species name (scientific) as part of the file name, thus my IMG_1234 of Melipotis indomita has the title “Melipotis indomita_1234”. I’ve always wished the uploader could parse my photo title to extract that but that’s probably asking to much.
So my feature request should be modified as suggested by allowing an individual user to exclude certain metadata fields from being parsed for species IDs.
This is more motivation for me to eventually migrate my whole photo catalog to Lightroom or whatever!
I could change this to a Feature Request (I think), but since the nature of your request has also changed as part of this discussion, it might track better to just create a new Feature Request topic with exactly what you have in mind, and then put a link to it in this thread (and vice versa).
As for this thread, it can probably be changed to a “General” topic, since it seems pretty clear that this is design behavior and not a bug. Any objections?
It’s a feature. It’s for people who keyword their photos in their photo software before exporting to iNat upload, so they don’t have to ID them again, as @treichard noted.
@gcwarbler probably best to make a new Feature Request for this. Note, however, that at the moment we are a bit loathe to add more toggles and options to the Account Settings page until it’s redesigned and less messy. I’ll also move this topic to General since it’s not a bug.
In theory it should choose the keyword that matches the finest taxonomic level. So if I have a rattlesnake photo and I have keywords “serpentes”, “crotalus”, and “crotalus oreganus”, iNat should use that last one, since it’s at the species level. If it detects no scientific names, it will look for common name matches, but that’s a different can of worms.
If anyone finds a bug in this, please email help@inaturalist.org with an exact copy of the photo so we can take a look.
