I am continuing work with integrating Aristotle with some other tools we have here in our department.
In the past I’ve done this mostly by using my own login credentials for adhoc bulk processing. However, as we proceed to automating certain tasks I would very much prefer to use a service account which can be assigned certain permissions. This assignment of permissions would, I think, be much in the same way we assign permissions to our own personal tokens we generate in Aristotle’s token management system. However also different, since a machine process is doing a driving not a person, so it’s a different sort of account with different needs.
Thanks for bringing this up - we’ve just added a new metadata harvesting tool and are building some new features at the moment for federation between Aristotle and other ISO-11179 metadata systems - so its an excellent time for us to look into this.
What we can investigate is add new “service users” that can be created by system administrators and can be assigned their own API keys. The new federation system is due for release in mid-September, so hopefully this will work for your internal development timeframes.
We’re also planning some upcoming webinars, so I’ll make sure that we include a webinar on all the new federation/harvesting/api improvements to show you how this can be managed.
The service accounts I set up, initially, would only need to be read only. However the option to bulk upload metadata using a service account would also be desirable.
Would these service accounts be able to be configured to write as well as read? It would be neat if you could control a service account’s reach in much the same way we can manage the access levels of our own tokens.
Always looking for ways to make the process of writing content faster. I have run custom API jobs before, using my own credentials not a service account, to read technical metadata from some source and then transform/load to Aristotle as a series of ‘bare-bones’ distributions. This saves much of the drudgery of manual entry, allowing our metadata specialists to focus their time on the job of linking those pre-filled distribution columns to data elements.
That has always been done on an adhoc basis, but soon I see the possibility of making this a regular thing. Lots of design work to do in the meantime, but having service accounts up our sleeves is one of the tools needed to make this happen.
Thanks for the fast turnaround, mid-September is impressive.