Best representation for tag / key value annotations in search ResultItem
Current situation: Both tags and key-value OMERO image annotations are represented as keywords. According to the first Hub Health group feedback, this might be confusing for the end-user and avert them from using the interface (see the screenshot for the clarification)
Requirements:
- Need for OMERO image items: represent unstructured annotations (child of #106)
- unstructured: representing them as fields of the record won't do
- Need to weigh the annotations in on search (indexer-related, currently keywords are more important for search than just the record text)
- Think about user perspective and make the representation not confusing
Problems:
- keywords are split on ","
- annotations have different formats (Tag, Key-Value), which are not well representable in the Keyword section
- keyword filters are automatically produced by concatenating all annotations, which is wrong
Activity
added UI/UX + 1 deleted label
assigned to @f.dmello
By keywords, we only mean tag annotations. Could you please explain why key-value annotations are being displayed as keywords in the image you attached? Is that what you expect as keywords in OMERO data?
If you had to do this because the items are search weighted on keywords that's a different issue, that I acknowledge, needs to be solved as you mentioned in the second point of the requirements. So, the problems in keyword filtering is already known but we are yet to decide how to solve this issue.
@f.dmello I guess I would be fine without "weighting benefits" in the search for the moment.
I use keywords for key-values because I don't know any other suitable place. I can not add those values as ResultItem fields, because the keys are highly variable and user-edited.
Maybe a way out would be to create a specific key-value representation for this regard?
marked this issue as related to #112
-
#109 (comment 6779890) keys are user-specified, I can not know what will be specified as a key even approximately