I know the pain. While there are definitely solutions that work sometimes, there's just no "one size fits all" that I'm aware of. PDFs can represent text very differently internally.
What I did for one project where extracting the text produced a complete mess was to convert the PDF pages to images and then OCR them...
It's a curse because it's used for things other than what it's intended to. It's doing a good job representing printed material, but unfortunately people very commonly expect it to be something more akin to a word processor file.