A simple script for extracting plain text from arxiv dataset: https://www.kaggle.com/Cornell-University/arxiv - View it on GitHub