[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[EP-tech] ORCIDs - Crosswalks - EPrints

Morning All,

There's been a small problem on the repository since I started that I
decided to investigate yesterday. It involves the duplication of ORCIDS and
their associated email addresses. Sometimes they're duplicated in the
database and in the metadata of an eprint. This prevents the ORCID
hyperlink from working and prevents the small ORCID logo from appearing in
the EPrint citation.

To illustrate:

I expect in the database - 0000-0002-5069-1909

I get in the database - 0000-0002-5069-19090000-0002-5069-1909

I've identified the problem as Elements-Crosswalks related. The crosswalks
take every instance of an ORCID and user email address from each source and
then squash them together in their respective fields.

For example:

feed --> pubs:users - pubs:user - > pubs:identifiers -> pubs:identifier

and then it also grabs the ORCID this from the elements sources where it
exists (in this case from epmc).

I understand why it has to do it this way, because some sources have all
authors ORCIDs and some don't, so they all need to be picked up. I think my
solution is to alter the crosswalks behaviour.

Before I start down this route, has anybody encountered this previously and
is there a better/easier way to fix it?

Regarding fixing the database, I feel incredibly uncomfortable going in to
an EPrints table and directly altering stuff. Is there a way I can fix the
already duplicated ORCIDs? I did ask our Elements team if the data can be
prodded to re-send from Elements through  the crosswalks and into EPrints,
but it looks like a no go.

I was using version 1.3 of the ORCID support plugin, clicked upgrade, got
an error and now it appears to be 1.7, but I don't think my problem is
anything to do with the plugin.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ecs.soton.ac.uk/pipermail/eprints-tech/attachments/20190404/4c9a4ea2/attachment-0001.html