Try to extract the text entities (thus the text) from a DjVu document.

At the moment the granularuty of the extracted text is "word", as "char" is quite difficult to find in documents.

Kind of works, let's hope it will like all the documents around.