For investigation. I am not aware of a converter for different language code standards such as ISO-639-1, ISO-639-2, and ISO-639-3.
If you use OpenNLP, for instance, it expects the language codes in ISO-639-3. But if you have language codes such as "en", then you will have to add a converter to "eng".
This might be moved to [lang] if that makes more sense.
Useful resources for this issue: