java - 斯坦福 nlp : Parse Tree

Question

我有这样一句话:我的狗也喜欢吃香肠。

我得到以下解析树:

(ROOT
 (S
   (NP (PRP$ My) (NN dog))
   (ADVP (RB also))
   (VP (VBZ likes)
     (S
       (VP (VBG eating)
        (NP (NN sausage)))))
(. .)))

如何只获取语法类别，即:NP、ADVP、VP 等？

我试过这段代码:

  Tree t=sentence.get(TreeAnnotation.class);
  t.labels();

Answer 1

从句子注释中，您可以获得各种类型的相关词集合。这可能是您正在寻找的“下一个级别”。

Tree tree = sentenceAnnotation.get(TreeAnnotation.class);                             
// print the tree if needed                                                           
SemanticGraph basic = sentenceAnnotation.get(BasicDependenciesAnnotation.class);      
Collection<TypedDependency> deps = basic.typedDependencies();                         
for (TypedDependency typedDep : deps) {                                               
    GrammaticalRelation reln = typedDep.reln();                                       
    String type = reln.toString();                                                    
}                                                                                     

SemanticGraph colapsed = sentenceAnnotation                                           
        .get(CollapsedDependenciesAnnotation.class);                          
Collection<TypedDependency> deps = colapsed.typedDependencies();                      
for (TypedDependency typedDep : deps) {                                               
    GrammaticalRelation reln = typedDep.reln();                                       
    String type = reln.toString();                                                    
}                                                                                     

SemanticGraph ccProcessed = sentenceAnnotation                                        
        .get(CollapsedCCProcessedDependenciesAnnotation.class);               
Collection<TypedDependency> deps = ccProcessed.typedDependencies();                   
for (TypedDependency typedDep : deps) {                                               
    GrammaticalRelation reln = typedDep.reln();                                       
    String type = reln.toString();                                                    
}