Automatic Annotation and Assessment of Syntactic Structures in Law Texts Combining Rule-Based and Statistical Methods