Python scripts for processing USPTO inventor and patent data

The following collection of scripts performs pre- and post-processing on patent data as part of the patent inventor disambiguation process. Raw patent data is obtained from Google Bulk Patent Download.

For a high-level overview of the patentprocessor toolchain, please see our technical report.

For a description of configuration of the patentprocessor toolchain, please see this technical report.

To follow development, subscribe to RSS feed.

Patentprocessor Overview

There are several steps in the patentprocessor toolchain:

1. Retrieve/locate parsing target
2. Execute parsing phase
3. Run preliminary disambiguations:
   • assignee disambiguation
   • location disambiguation
4. Prepare input for inventor disambiguation
5. Disambiguate inventors (external process)
6. Ingest disambiguated inventors into database

For the preliminary disambiguations, you need the location database. File requires 7zip to unpack.

Installation and Configuration of the Preprocessing Environment

The python-based preprocessor is tested on Ubuntu 12.04 and MacOSX 10.6. Any flavor of Unix with the following installed should work, though it is possible to get the toolchain running on Windows.

If you have pip installed, you can simplify the installation process by just running sudo pip install -r requirements.txt from within the patentprocessor directory.

Please file an issue if you find another dependency.

Ubuntu

sudo apt-get update
sudo apt-get install python-dev
sudo apt-get install python-setuptools
sudo easy_install -U distribute
sudo apt-get install -y python-Levenshtein make libmysqlclient-dev python-mysqldb python-pip python-zmq python-numpy gfortran libopenblas-dev liblapack-dev g++ sqlite3 libsqlite3-dev python-sqlite redis-server
sudo pip install -r requirements.txt