This directory contains classical unsupervised approaches, which do not utilize word embeddings.
These include YAKE!, KPMiner, MPRank, PositionRank, TopicalPageRank, SingleRank, TextRank and TopicRank.

In order to run this script you need to:

pip install pke
pip install pytextrank
pip install spacy
pip install git+https://github.com/LIAAD/yake

The en_core_web_sm model for the respective spacy version needs to be installed, since it is used by pytextrank.
TopicalPagerank and KPMiner use a lda_model_file and a weights_filerespectively, which can be obtained from the pke repo.
After they are obtained, their respective paths and the base_path for the dataset directory should be set in main.py.