[huqedato]

I tried to parse a few hundreds pdfs with it. The results are pretty decent. If this was developed in Julia, it would be ten times faster (at least).