Added ocr lang config, psm and GCV support

[naourass]

• Can choose between tesseract or Google Cloud Vision as an OCR engine (default to Tesseract)
• Can specify languages and page segmentation mode (psm only for tesseract, languages should be expressed in the appropriate format : Tesseract / GCV langs refs)
• Can generate output with non-latin characters

[lolipopshock]

Thanks! Will take a closer look next week.

[naourass]

Hi there, I'll be adding following edits to this PR :

• Unknown option error message : removed gcv-credential option hint since the file should be dropped directly in working dir
• Renamed and refactored preprocessors/tesseract.py to preprocessors/ocr.py for better legibility
• Added comprehensive error message and handling when gcv credential file is not found
• Formatted code using black provider

Looking froward to your feedback about the implementation quality and if there's any relevant changes that should be done, so I can push the update and move forward FUNSD-like and IOB Format export features.