A Python implementation of the algorithm used in arc90's readability bookmarklet:

>>> import urllib
>>> import readability
>>> url = 'http://www.nytimes.com/2010/09/07/health/views/07mind.html'
>>> html = urllib.urlopen(url).read()
>>> print readability.readable(url, html)[0]

To directly open the readable version of a URL in the web browser:

$ readability -b http://blog.doughellmann.com/2007/04/pymotw-linecache.html

readability.py is not released yet. To install the development version:

$ pip install -e git://github.com/srid/readability.git#egg=readability

Unfortunately this project, like others, uses an older version of the readability algorithm, Viz.

Matt: [...] the readability.js file that can be downloaded from the "downloads" section of the Google Code project is a year and a half old and only 8.9KB (about 250 loc), while the trunk version (presumably similar to what's used in the bookmarklet) has expanded to a whopping 73.5KB and 1825 loc.

Ideally, we should port this project to use the same algorithm as the bookmarklet.

readability.py adds several bug fixes and features to hn.py in the Readable Feeds project that adapted the original hn.py by Nirmal Patel. readability.py retains the original license (GPL3) chosen by its predecessors.