Scan and fix incorrect metadata

ben_a · November 11, 2016

Is there a way (built in or through a plugin) to automatically download missing metadata for a record? Or a way to scan a record and check the metadata against a database like PubMed or Google Scholar to check that the metadata is correct?

ben_a · November 11, 2016

I mean for records that already have metadata, I know that I can retrieve metadata from a PDF if there is nothing there already.

bwiernik · November 11, 2016

This feature is planned, but not currently available. Easiest way currently is to add the item a second time from PubMed, etc., then merge the two items.

ben_a · November 12, 2016

Ok, would be good to have this feature but your workaround isn't too inconvenient.

Thanks.

DWL-SDCA · November 12, 2016

In theory, the capacity to have this feature incorporated into Zotero or available as an add-on may be available in the not-distant future.

I built a system for my online database service (SafetyLit) to automatically query the PubMed system API and refersh volume-issue-pagination metadata and it works well. The problem is that having access of that kind requires a license. Although the license has no cost and can be had quickly, it must be renewed each year and there is a restriction based on IP range. Parts of the NCBI databases have become license free effective October 2016. I hope that access to refresh metadata will also soon no longer require a license.

For items not included in PubMed, we are working on a similar system that uses CrossRef, LoC, and other national libraries. Unless things change, the CrossRef and LoC will not require a license for datafase access. The main block from accomplishing this is our budget restraint and a few things that have a higher priority. I expect this by the end of March 2017.

My offer to share the code stands open, but until the license issues work themselves out, they aren't very useful to others.

adamsmith · November 13, 2016

(no need to hash this out here, but NCBI's e-utils, which Zotero currently uses for all interaction with pubmed, doesn't require a license and I'm pretty sure we should be able to make it work for this. As you say, the relevant CrossRef API as well as LoC's SRU are already license free. We know how to do this in general and the API calls are the easiest part -- harder are the GUI implementation and making merge behavior (if any) smooth and intuitive.)