ConceptNet aims to give computers access to common-sense knowledge, the kind of
information that ordinary people know but usually leave unstated.
This Python package contains a toolset for building the ConceptNet 5 knowledge
graph, possibly with your own custom data, and it serves the HTML interface and
JSON Web API for it.
You don't need this package to simply access ConceptNet 5; see
http://conceptnet.io for more information and a browsable Web interface with an
API.
Further documentation is available on the ConceptNet wiki.
Licensing and attribution appear in LICENSE.txt and DATA-CREDITS.md.
Discussion groups
If you're interested in using ConceptNet, please join the conceptnet-users
Google group, for questions and occasional announcements:
http://groups.google.com/group/conceptnet-users?hl=en
For real-time discussion, ConceptNet also has a chat channel on Gitter:
https://gitter.im/commonsense/conceptnet5
Installing and building ConceptNet
To be able to run all steps of the ConceptNet build process, you'll need a Unix
command line (Ubuntu 16.04 works great), Python 3.5 or later, 30 GB of RAM, and
some other dependencies. See the build process on our wiki for instructions.
You may not need to build ConceptNet yourself! Try the Web API first.
Testing
Run pytest to test the ConceptNet libraries and a small version of the build
process.
Run pytest --quick to re-run the tests more quickly, with the assumption that
the small test database has already been built.
Run pytest --fulldb to run additional tests on the fully built ConceptNet
database.