Issue 17407

Backbone source empty

17407
Reporter: mdoering
Assignee: mdoering
Type: Bug
Summary: Backbone source empty
Priority: Critical
Resolution: Fixed
Status: Closed
Created: 2015-03-09 12:27:00.436
Updated: 2015-11-13 13:25:07.842
Resolved: 2015-11-13 13:25:07.705
        
Description: All backbone usages have no explicit, copied extension records (e.g. distribution, media, vernacular names, etc) so far, see ISSUE X. Instead we dynamically link all extension records based on the nub matching from various source checklists.

All extension record model classes have a source and a sourceTaxonKey property. For nub records the key is the name usage key of the source taxon and allows to link to the source. The source property should contain the checklist title in case of backbone records. It is used for other external sources, mostly a literature citation, in non nub checklists.

Since the larger refactoring of ChecklistBank there is no dataset table anymore that allows us to retrieve a dataset title via sql. So ALL nub extension records have an empty source property right now!

For example:
http://api.gbif.org/v1/species/7187198/distributions

In order to fix this situation I propose to add a simple dataset table with key & title for the time being, populating it manually for now from the current registry.

In the long run when we build a new backbone we are likely to copy extension records explicitly doing some quality control along the way. This allows us to read a dataset title from the registry and populate the source field with that.]]>
    


Author: mdoering@gbif.org
Comment: [~trobertson@gbif.org] do you agree to use a simple lookup table within CLB for the next few month until the new nub build is ready?
Created: 2015-03-09 12:28:33.721
Updated: 2015-03-09 12:28:33.721


Author: trobertson@gbif.org
Comment: +1
Created: 2015-03-09 13:40:59.804
Updated: 2015-03-09 13:40:59.804


Author: mdoering@gbif.org
Created: 2015-11-13 13:25:07.821
Updated: 2015-11-13 13:25:07.821
        
There is a new dataset table in clb now that is maintained by the clb-registry cli listening to changes.

A new clb-admin SYNC_DATASETS command will truncate that table and insert all the latest checklist titles.

Indexing works fine and the dataset titles show as sources for nub records now: http://api.gbif-dev.org/v1/species/2704922/vernacularNames