ORCID in Invenio 1

ORCiD

• every author/researcher gets a unique number, the ORCiD ID
• thus, a reliable primary key for every author across DB boundaries → easy data transfers
• ORCiD maintains profiles (publications lists, affiliations etc)
• ORCiD is non-profit, and is supported by many institutions

Goals

• collect ORCiD IDs of as many collegues as possible at Jülich, store them in our Invenio instance
• allow collegues to upload their Invenio publication list to their ORCiD profiles
• no manual intervention of us necessary

Constraints & peculiarities

• slightly lazy: own webserver (Flask-based) for simple implementation
• rather selfish: standalone publication uploader in Go
• absolutely hilarious: Python 2.4 for Invenio, Python 2.6 for ORCiD bindings

The webserver

• standalone, i.e. not part of the Invenio WSGI process, own port
• uses Invenio modules to access the Invenio DB directly
• caching of ORCiD IDs & access tokens in SQLite in /tmp

OAuth 2

URL endpoints

Publication upload

• ORCiD is sloooow
• publication upload is independent background process
• upload must be throttled
• until recently, one had to upload one-by-one when using ORCiD API 2.0
• still not possible to mass-upload all publications in one request

Technology is not everything!

• consider legal and regulatory implications
• plan thoroughly how to convince your scientists of ORCiD
• the Git repository contains a mass mailer, which can run as a cronjob (use it carefully); and it contains a success measuring program

ORCiD rate in Jülich

Future work

• merge Python 2.4 and 2.6 code into one 2.6 module!
• make use of recently added mass upload feature of the ORCiD 2.0 API
• if concurrency is not needed anymore, re-implement Go code in Python
• filter publications eligible for upload (e.g. by year)

Check it out!

https://git.io/orcid

Thank you