Create Tesseract_jll package

[timholy]

Ideally the dependency on Tesseract OCR would be handled by creating a _jll library using BinaryBuilder. That way the entire package would always be available simply through Pkg.add.

In the meantime, #19 clarifies the requirements.

[leferrad]

In fact, this idea was also mentioned in #6, but I never did this task of creating _jll libraries. @timholy do you have some example of library or PR to check in order to work around that?

[timholy]

I've not done it myself but a lot of people have---at this point there's almost the equivalent of an entire Linux distribution's worth of _jll libraries available. There's a #binarybuilder channel on the Julia slack, if that helps.

[leferrad]

Luckily, Tesseract_jll already exists! The only thing is that now OCReract should be aware of the existence of the data for the language to use, something that so far is assumed responsibility for the user. Let me work around that to then submit a PR with the best changes I figured out.