This is great! I did something similar for myself that also renames the input pdf file based on metadata from CrossRef. Hopefully I can get around to contributing code to this project.