--tokenize
Instead of looking for grammar errors, split input into tokens. The tokens are prefixed by type:
"W" is a word, "P" is punctuation, "S" is whitespace, "U" is unknown and "E" is a prefix for error
messages.
--split-sentences
Instead of looking for grammar errors, split input into sentences. The sentences are prefixed by
type: "B" means that end of sentence is a probably correct, "P" means that end of sentence is a
possibly correct (but probably this and the next identified sentence should be joined) and "E"
means that sentence ends at the end of input.
-n Prefix all grammar checker messages with line number of input data.
accept_titles=naccept_unfinished_paragraphs=naccept_bulleted_lists=n
Set the value of the specified boolean option.
explanation_language=langcode
Print human readable error explanation in the specified language.