Description of need: An API to fetch all comments made by a User, ideally with taxa filters
Feature request details: There was a request for a similar feature in the past,https://forum.inaturalist.org/t/search-comments-by-user-id/1009 but it was closed as will not do. In light of both recent advances in technology and iNat’s proposed usage of comments for training an LLM to help with identifications (which I fully support and I ask folks to please not make this post about the pros/cons/ethics/etc of that type of feature) I would very much like to access comments that have been made so that I can use this same data to build some experiments of my own.
I have not dug very deeply on the details of how comments are licensed on iNaturalist but this would need to be part of the overall plan. I can see constraining the API to only the caller’s own comments (unless there is already clarity on licensing of comments that I could not find).
Bottom-line, I would really like to access my own comments (at least) to leverage for experimenting and building out tools to help the larger naturalist and scientific community. In my immediate use case I am working on an Oak ID website and I really would like to mine my comments on the ~25,000 oak IDs I have made over the years.
If there is an existing way to accomplish this, I could not find it, but welcome any pointers to possible ways to accomplish this without a new API.
As I think you know, users can already view and search their own comments via the UI. (Quick note: iNat regards standalone “comments” as being distinct from text entered in support of an ID, and stores them as separate entities. The comments search only queries standalone comments.)
Staff and curators can view and search comments of other users, which may sometimes be needed to determine if users are adhering to community guidelines. iNat staff rejected the idea of allowing non-curator users to search other users’ comments, presumably on the basis that the potential for harassment, etc. was not outweighed by the benefit.
Right now, all these comments searches are pretty slow, which I think indicates that they were built quite a while back and have not been optimized since then. Also, AFAIK they’re not available via the API, which likely relates to the age of this part of iNat. However, it does seem that staff must have a better way of getting access to this data, as mining these comments is the foundation of the ID Summaries Project.
Overall, I don’t think @jeffdc’s request is likely to get much support from iNat staff. Searching other people’s comments seems already to have been rejected. And providing users with API access to their own past comments so they can train their own AI models seems to be too limited a use case to justify developer time.
P.S. If you do want to see something like this, make sure to vote for the feature request @jeffdc and @porroglossum.
I sincerely hope that this does get some attention from the staff. My belief is that without this there will be a general decline in useful comments. Maybe I am an outlier but I have actually started (a while back before the comments trained ID experiment was announced) tracking all of my ID notes outside of iNat and rarely comment in iNat about my IDs (unless someone specifically asks) due to the locked down nature of these comments. I have a Firefox extension that allows me to tie the observation to a comment (that is not in iNat and instead placed in a place that I control), but this does not help me with all of the comments that I made before I switched over to this mechanism.
My interpretation of the licensing situation for these comments is that they are All Rights Reserved of the creator, other than for iNaturalist itself who we grant a “worldwide, royalty-free, and non-exclusive license to reproduce, modify, adapt, and publish” the comments. This allows the site to legally display the comments, distribute them to other users, and maintain the functionality of the platform (e.g., displaying a comment thread). So iNat is well within their rights to ignore this request but I know that for myself this will mean that I will contribute far fewer comments directly to the site since they will essentially be forever locked away.
I suppose one could ask for the data via a CCPA request or similar but that would be a mess for all involved!