[ ] add doi's of all test articles to the code, so that others can check which articles were used for testing
[ ] check the "manual" csv files in test directory; these contain manually extracted stats for test articles but are not finished. these might be allowed to upload.
[ ] when reading in manual csv files, remove last two rows (only show total nr of extracted results)
[ ] rewrite tests to focus on pdftools as default method
[ ] write explicit tests to test difference in retrieval between pdftools and xpdf