Hello!
I've been looking through the tesseract code for some time now, but I couldn't find out, where tesseract actually goes through the ratings matrix to determine to which word(s) the ratings marix resolves.
I've found the rules by which you can wander through the matrix, but not where it actually does it, by which which mechanism it rates the results or how the result set is limited to keep the performance.
Does someone know where I can find that part in the source?