Adding option to check language in corpus data

Unfortunately, fasttext cannot natively detect sme, so it is only possible to check if a line matches either no, nn (the two versions of norwegian) or en.
If this is the case, it can be supposed that the line contains potentially a significant amount of text in this language.