Please add support for tessdata_best OCR models

[voska7]

Hi, thanks for making Subtitle Edit - it's really useful!

I wanted to ask if you could add an option to use the tessdata_best OCR models from Tesseract.

Right now, Subtitle Edit uses the default traineddata files. But the tessdata_best files can give better results when converting .sup (PGS) subtitles to .srt, especially with hard-to-read fonts.

Could you please:

Add a way to choose between default, fast, and best models in the settings?

Or let us pick a custom tessdata folder path?

That would make OCR much more flexible. Thanks again for your great work!

https://github.com/tesseract-ocr/tessdata_best