We use the node.js html-metadata library to scrape metadata from webpages. This metadata is then used to generate a citation for the webpage using citoid. Editors can use this citation when they're writing or editing an article on Wikipedia as a reference.
For this you will need to install node and npm: https://www.joyent.com/blog/installing-node-and-npm/.
You will also need to fork the html-metadata library on github: https://help.github.com/articles/fork-a-repo/
Eprints tags are a type of metadata in a webpage that gives information about how to cite the resource. There is a good explanation about them here: https://www.google.com/intl/en/scholar/inclusion.html#indexing
Here is an example of a page with eprints metadata: view-source:http://eprints.rclis.org/13306/
An example of an eprints tag from that page:
<meta name="eprints.title" content="Die Informationsfachpersonen in einem digitalisierten Umfeld" />
You should also create tests for your new function. You can put tests of a live website in https://github.com/wikimedia/html-metadata/blob/master/test/scraping.js. You can also create tests of a static website in https://github.com/wikimedia/html-metadata/blob/master/test/static.js. You may find it helpful to write these tests early on in writing your function, as they will help you verify your function is working correctly. Tests can be run by doing