Hi, this idea might be stupid, impossible, or both.
I I have a rather futile obsession with achieving hierarchical tagging, but ofc have no ability to code something myself (it seems likely that number of relationships would kill a database but that's just a hunch).
Anyway, I figured you could use namespaces in hydrus to group indexed "entities" to achieve one level of tag grouping, but this is only useful if the client allowed a strange form of wildcard search.
Say I have two images, both with two people. (namespaces "entitiy1:" and "entity2:").
I want to search for an image where a single entity has both Blue Skin and Glasses.
I could group them like this:
(image 1)
entity1:blue skin
entity1:glasses
entity2:glasses
(image 2)
entity1:blue skin
entity2:glasses
Using traditional tag search, both images would match, they both contain both tags ("glasses" and "blue skin"). But I am only interested in the image that features both tags in the same namespace. So in this example the first image has a match for my search, a single entity with both tags under its namespace. BUT! How would I search for it? If I used a wildcard character in the namespace ("entity*:") that would still match both images, since I have now eliminated the indexing that differentiated between the enteties. And using a known unique namespace would defeat the whole exercise.
What would be handy would be a wilcard character that can match any "entityX" across a database, but will only return a match if each tag with the wildcard is under the same namespace on a single image. A search where I do not care what the wildcard/namespace match *is*, only that it is *the same* for all the specified tags.
I do not know if this is worthwhile or even possible to implement, but it has been churning around in my head and I though I should at least mention it.