Install pip environment via pip install -r requirements.txt

Run with streamlit run badpun.py from root directory. When running two instances, specify port by streamlit run badpun.py --server.port 8051. For one high clustering and one low clustering instance run

$ streamlit run badpun.py --server.port 8051 -- high
$ streamlit run badpun.py --server.port 8052 -- low

The folder src/ contains all files code files required for the Application. experiments/ contains jupyter notebooks that were used to determine parameters and conduct experiments. In the folder data/ lies both example data for training of the model and for running the application. The folder submission/ contains our final report pdf file.

The file 'config.ini' contains all constants used throughout the system. In order to avoid spread of information, best add new constants to this file, as well as parameterization of methods etc. Usage:

import configparser
config = configparser.ConfigParser()
config.read('config.ini')
config['DATA']['TestUserEmbeddingPath']

On the main branch, the wordclouds are computed using the simple attentions due to time constraints. For the LRP implementation please switch to the high-dim-and-lrp branch. `