FairDiversity

Downloading datasets

There are two ways to get the datasets working:

Google Drive

Download the ziped file at: https://drive.google.com/file/d/1XuxgGWcZPNnUhNBBcHbxvYk3YCf_0ooU/view?usp=share_link

Extract and copy its contents into the ./datasets for all the imports in the code to work.

Script

Run ./datasets/init_datasets.sh to download and initialize all datasets. NOTE: This does not work for the YELP dataset

Validating Results from Paper Submission

To reproduce our results, run ./run_all.sh which will run all the experiments mentioned in ./publish/setup/.

To validate our graphs, the readings can be found in ./publish/results. To replot graphs run ./plot_all.sh within ./publish.

Install and run the virtual env.

Note: License for gurobi is required to run a few algorithms

pipenv install
pipenv shell

Run experiments

Experiement setup files can be found under ./experiments, an experiment can be run using runner_k.py or streamrunner_k.py. For example:

python runner_k.py ./experiments/equal_kis.json

The result files would also be generated under ./experiments with a timestamp

Plot experiments

python plot_k.py ./experiments/result_equal_kis/result_23_10_12_03_10_24.json