Highlighting PDF for CoreNLP

Prior to the release of 3.2, using PDF Highlighter with NLP (natural language processing) tools was inefficient and couldn’t guarantee 100% precision. Before 3.2, it was necessary to parse NLP output to extract key terms and phrases and then feed Highlighter a query file containing potentially hundreds of long phrases (e.g. whole sentences) you wanted marked in PDF. Thus, highlighting a single document with a few hundred pages could take minutes to complete—far from ideal.