BM25 for document ranking

This project implements BM25 algorithm described in this paper for ranking documents according to relevance.

Installing

Make sure to run the setup.sh script. It will install all required dependencies.

Running the sample

Use python sample.py to see it in action. It may take some time, but eventually it will print to the console retrieved documents matching the query, sorted by relevance.

Known issues

This implementation is best fit for small datasets because it lacks the inverted index required for fast querying in big datasets

Alternatives

https://github.com/nhirakawa/BM25