Document similarity server
Index plain text documents and query the index for semantically related documents.
Simserver uses transactions internally to provide a robust and scalable similarity server.
Simserver builds on the gensim framework for topic modelling.
The simple way to install simserver is with:
sudo easy_install -U simserver
Or, if you have instead downloaded and unzipped the source tar.gz package, you’ll need to run:
python setup.py test sudo python setup.py install
This version has been tested under Python 2.5 and 2.7, but should run on any 2.5 <= Python < 3.0.
See http://radimrehurek.com/gensim/simserver.html . More coming soon.
Simserver is released under the GNU Affero GPL license v3.
This means you may use simserver freely in your application (even commercial application!), but you must then open-source your application as well, under an AGPL-compatible license.
The AGPL license makes sure that this applies even when you make your application available only remotely (such as through the web).
TL;DR: simserver is open-source, but you have to contact me for any proprietary use.
- performance improvements to sharding
- change to threading model – removed restriction on per-thread session access
- bug fix in index optmize()
- changed behaviour for very few training documents: instead of latent semantic analysis, use simpler log-entropy model
- fixed bug with leaking SQLite file descriptors
Copyright (c) 2011 Radim Rehurek