The script accepts a list of file_ids, and generates a list of documents from these ids. The documents are stored in a directory structure generated from the categories in the irb_documents spreadsheet