These are some experiments for working with the Calais web service from Reuters: http://developer.opencalais.com/ calais.py provides a wrapper around the REST api which lets you submit content and get back a rdflib Graph. Before you can start executing jobs you'll want to get a developer license key and store it in ~/.calais You'll also want to install rdflib: easy_install rdflib There are some sample scripts that use calais.py and some sparql to print out pieces of information extracted from text: - cities - people - provinces So you can do something like this: uqbar: ed$ ./cities data/ndnp\:1396148 | sort Boston Cincinnati Cleveland Fitchbiirg Fresno Great Barrington Mace Methuen Mullln Sacramento San Francisco WASHINGTON Washington If you execute calais.py directly it'll print out the entire rdf graph for the calais response as n3, which can be useful for discovering what other data is in the graph. uqbar: ed$ ./calais.py data/ndnp:1396148