I would like to limit the files loaded using langchain_community.document_loaders.pdf.PyPDFDirectoryLoader to a specific list. I tried using glob but there is some odd behaviour there

For instance, say I have

files_to_load = ["a.pdf", "b.pdf", "c.pdf"]

This does not work (loads 0 files)

loader = PyPDFDirectoryLoader("docs/", glob = "|".join(files_to_load))

This does not work either

loader = PyPDFDirectoryLoader("docs/", glob = "["+"|".join(files_to_load)+"]")

however running glob.glob with the same string does return the right files

Am I missing something obvious?