Decide Which Type

Run either byte NGramJ or character CNgram Here are some common cases.

You have files with text of unknown encoding.

Use byte NGramJ to determine both encoding and language.

You have files with text of known encoding.

Use CNgram to determine language or mixed language documents.

You don't have files but Strings within your Application.

Use CNgram to determine language or mixed language Strings.

You have structured files in XML/HTML.

Usually encoding is not the problem, but you need to get rid of the markup by using a parser first, then use CNgram to determine language or mixed language documents. Note: The parser has to be started somehow differently.