API Update: Serving Up Author Affiliations

This week we have done more work on the API return format which now serves up author affiliation metadata. Author affiliations are optional for arXiv e-prints, and therefore we only give this information when it is present. The format is simple, and consists of an arXiv extension element <arxiv:affiliation> as a sub-element of the Atom <author> element. For example, the e-print seen at

http://arxiv.org/abs/0710.5765v1

has four authors and four affiliations. The API return for this record http://export.arxiv.org/api/query?id_list=0710.5765v1 has elements that look like

<author>
       <name>G. G. Kacprzak</name>
       <arxiv:affiliation xmlns:arxiv=...>NMSU</arxiv:affiliation>
</author>

Note that a single author can have multiple affiliations which are included as separate <arxiv:affiliation> elements. The API user’s manual has updated documentation and code examples in Perl, Python, Ruby and PHP that parse this information.

There has also been some recent requests for date sorting of API results. For example, you might like to retrieve “the most recent n items matching the query x”. Right now the API returns results in order of search relevance, and does not do date sorting or filtering. We are working to change this! In the mean time, you can always sort entries by date using the entry <published> or <updated> elements, although to do this correctly would mean you have to download all of the search results. In the future, the API will have a mechanism that lets you do date sorting/filtering on search results.

As always, we welcome your comments!