Have you tried using an ~~XML~~ PDF parser instead? /s

and tried to find which sentence without spaces matches your sentence