Entity extraction:

Given a document, extracts points of interest from the document.

Points of interest might include:

* phone #s
* addresses
* names of people / companies / places
* sentence boundaries
* dates and times (both specific and relative)
* shipment tracking numbers
* flight confirmation numbers

The rough API is:

extract(str)

and returns

[{'start': 10, 'end': 18, 'type': 'phonenumber'}, ...]