I'm chiming in here about another API that is being built at WMF that may be somewhat related.

On the Editor Engagement Experiments team Dario Taraborelli and myself have invested a good deal of effort in standardizing the way we gather user metrics [1]. The initial aim was to have a set of well defined metrics which could be used in the downstream analysis of experiments. It soon became clear that we would need, in addition to these definitions, a reproducible way to extract this data from the MediaWiki datastore. To that end we've been working on a project [2] that does just that (pushed to Gerrit also [3]), docs can be found at [4].

Further, this project has been deployed to our Wikimedia stats cluster where it is hosted at [5]. This is still a development instance and thus we cannot provide guarantees on it's uptime yet, and besides the project is currently under htaccess. However, we're working toward the point of having this tool more widely exposed in the coming weeks.

[1] http://meta.wikimedia.org/wiki/Research:Metrics [2] https://github.com/rfaulkner/E3_analysis [3] https://gerrit.wikimedia.org/r/gitweb?p=analytics/E3Analysis.git;a=summary [4] http://stat1.wikimedia.org/rfaulk/pydocs/_build/ [5] http://metrics-api.wikimedia.org/

On Sat, Feb 16, 2013 at 2:40 AM, Yuri Astrakhan yuriastrakhan@gmail.comwrote:

I have started an RFC, putting down my initial thoughts on how to make Wikidata API more seamless with the core API.

The goal is to minimize execution time, bandwidth, and server load, while making it play nicely with the rest of the action=query, allow for continuations and multi-values data request capabilities.

For now I only looked at the wbgetentities, but the RFC will be updated with the other modules. Please let me know what you think.

http://www.mediawiki.org/wiki/Requests_for_comment/Wikidata_API

---

Wikidata-l mailing list Wikidata-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata-l