Interesting enough to find out the Reader function in Safari is actually Javascript and there are many interesting stuff from the 2000 line code:

* 5 main parts in the file:
* 1. define const
* 2. helper functions
* 3. CandidateElement
* 4. String Prototype extension
* 5. ReaderArticleFinder (main)
*
* findArticle: look for suggested candidate, otherwise look for others
* when looking for articles, check basic score of each elements (text length, text depth, id and class name, language)
* then cleanArticleNode: find pruneElements and remove them by shouldPruneElement: classname, textlength, imagearea, calculateLinkDensity
*

Way to call it:
* var ReaderArticleFinderJS = new ReaderArticleFinder(document);
* //var article = ReaderArticleFinderJS.findArticle();
* ReaderArticleFinderJS.isReaderModeAvailable();
* ReaderArticleFinderJS.prepareToTransitionToReader();

 

Blog also for this: http://joevennix.com/2011/05/09/Hacking-Safari-Reader.html