ChemSum

We introduce a dataset with a pure chemistry focus by compiling a list of chemistry academic journals with Open-Access articles. For each journal, we downloaded full-text article PDFs from the Open-Access portion of the journal using available APIs, or scraping this content using Selenium Chrome WebDriver. Each PDF was processed with Grobid via a locally installed client to extract free-text paragraphs with sections. The table below shows the journals from which Open Access articles were sourced, as well as the number of papers processed. For all journals, we filtered for papers with the provided topic of Chemistry when papers from other disciplines were also available (e.g. PubMed).