Rysearch is an exploratory search engine and recommender system. Based on BigARTM, open-source library for topic modeling, it takes into account latent topical structure of texts to achieve good results in both knowledge exploration and visualization.
Use our pre-configured Docker image for quick installation. Run:
docker run -t -p 3000:3000 tohnann/rysearch
And then open http://localhost:3000.
Everything is tested on Linux (NixOS) and Windows operating systems. If things don't work as described here — please submit us an issue.
Running a Rysearch server
cd server/ # Install Node.js libraries npm install
Rysearch server consists of two workers: ARTM_bridge and Node.js server. You have to run them as separate programs like this:
# Run ARTM_bridge python3 artm_bridge.py
# Run Node.js server npm start
 K. V. Vorontsov et al. Non-Bayesian Additive Regularization for Multimodal Topic Modeling of Large Collections, TM '15 Proceedings of the 2015 Workshop on Topic Models: Post-Processing and Applications, 2014.