Patentparse

A parsing library for patents.

This library is a shallow wrapper around ElementTree to provide easy access to Google's patent data archive (http://www.google.com/googlebooks/uspto-patents-grants-text.html)

This is currently only expected to parse patents granted after 2001.

Usage

A patent takes a raw XML string:

>>> import patentparse.patent
>>> patent = patent.patentparse.Patent(xml_contents)

It has some attributes to get at commonly-used information:

>>> patent.invention_title
"Method and system for preventing copying of information from previews of webpages"
>>> patent.doc_number
"08615810"
>>> patent.claims
['\n1. A method of providing a copy prevention feature...]

It also gives access to the ElementTree itself, for any information not available by the Patent API.:

>>> patent.tree.find('.//invention-title').text
"Method and system for preventing copying of information from previews of webpages"

And a JSON serialization:

>>> patent.json
'{"claims": ["\\n1. A method of providing a copy prevention feature...,
"doc_number": "08615810",
"invention_title": "Method and system for preventing copying of information from previews of webpages"}'

Bulk parsing

Google provides weekly dumps of all patent grants in a single XML file. `patentparse.parse` can be used to get an efficient Patent generator for all patents in the file.:

>>> import patentparse.parse
>>> patents = patentparse.parse('file.xml')