You'll need Python 2 with scikit-learn (and therefore numpy, scipy, and ctypes) installed. Anaconda is highly recommended.

$ ./dependencies

>>> import hanjatrans
>>> tr = hanjatrans.HanjaTranslator()
>>> print tr.translate(u'2011년 12월 26일')
2011年 12月 26日

Inferring correct hanja is somewhat difficult, and this library doesn't do it perfectly. See https://htmlpreview.github.io/?http://github.com/futurulus/infer-hanja/blob/master/sample_errors.html for more examples of the library's output, with errors highlighted.