Unify objects / aka uniquify [was: A task is needed to merge two TaxonNames]

[proceps]

Expected behavior. Before merging, all attributes from one name are transferred from one name to another. Duplicated attributes get destroyed. Self attributes (Author, parent, rank, etc.) stay unchanged. After that, one TaxonName get destroyed.

[debpaul]

This issue discussed today in the Nomenclature office hours. It's a larger issue too in that some steps to merge can be automated, other steps will have to manual, and it is not just for duplicate taxon names, but other duplicates (OTUs, family-group names, etc).

DD: suggest to put sources with nominotypical names to make this easier to merge in future (from DP, did I say this correctly @proceps?)

[mjy]

I think, if our model (and corresponding UI/UX) is setup correctly, we should only have to merge names if there are truly duplicates because of human error. Otherwise this voids the principle of "no changes, only new data".

[mabecabrera]

So needed for data curation purposes, including migration errors.

[mjy]

Endpoint:

/unify?remove_global_id=<global_id>&keep_global_id=<keep_global_id>&preview=true

[mjy]

To get a preview count of the numer of existing records use /unify/metadata?global_id=gid://taxon-works/Otu/1

[mjy]

• DataAttribute and Combination can be removed
• Radial navigator beside selected object(s)
• Radial annotator beside selected object(s)
• Object radial beside selected objects(s)
• Add radial navigator slice "Uniquify" that points an object to this task, and pre-initializes the left side
• Add headers and style(?) to left ('Keep') and right ('Remove') columns
• Maybe drop in Smart selector if it exists for the class, on choice of Model?

[mjy]

• Use name attribute in preview metadata for related row label
• Collapse smart selector, if present, and use undo button to clear OTU and make it re-appear
• Prevent identical object from being selected on both sides
• When tasked entered from URL and object(s) changed, update URL

[hhopkins77]

@mjy per your request: Here is just one example of something that comes up often: repositories that need unifying.

one ZMB + ZMHUB + BHU need to be unified

and

a second ZMB + ZMBH need to be unified

[hhopkins77]

@mjy I'm wondering if the author uniquify task could be subsumed into the new unify one? I think the new one might be easier.

[proceps]

I use people uniquify very often. And in most cases it is many to one. I do not think it is reasonable to replace it.

[hhopkins77]

@mjy I don't have an example just now but the ability to unify sources is needed.

[hhopkins77]

@mjy I need to be able to unify Taxon Name IDs. For example, I need everything associated with 1043436 (a ghost) to be unified with 1003420 (the "real" taxon name id). This example is in Plecoptera SF and this phenomenon appears to be a byproduct of migration. I don't know how many of these there are, (I'm guessing quite a few), and they can take forever to correct depending on the quantity of data attached to the ghost so a unifiy task for this would be a great timesaver.