Lately, I was doing tasks on image and PDF parsing. And I figured out that a lot of users are using modules like Camelot, OpenCV, or Tesseract OCR for that. The only problem is that modules require installation and they are not that flexible.
But I was somehow sure that for that reason neural-networks are used.
So basically I'm just wondering what do you use for solving such a task. Do you use modules or functions maybe?