README.md

spectral_jaccard_similarity

Setting up the environment

The set of packages needed to run the experiments is provided here

If one is using the Anaconda package manager, one can set up the environment running the following commands

conda create -n sjs python=3.7 numpy=1.18.1
conda activate sjs
pip install -r requirements.txt

Then all the code in the repository can be run in the sjs environment created.

Running code

To create filtered datasets run Bacterial_pipeline_part1.py and Bacterial_pipeline_part2.py. The sjs environment has all the dependencies needed to run them.

To use our pre-filtered datasets, in the sjs environment run,

python pipeline_wrapper.py --num_jobs <number-of-jobs> --datasets NCTC_ds.txt