Allow Tesseract OCR language files to be downloaded on a running instance

In order to work properly, Tesseract needs to have a language file to rely on.

This language file will change depending on the language of the recognized text.

Hopefully, Tesseract already provides training data here : https://github.com/tesseract-ocr/tessdata/tree/3.04.00

Note that those files are version-specific, therefore, if we plan to upgrade the currently provided version of Tesseract (3.04.01) to a higher version in the future, we will have to change the URL used to download those files.