Search code examples
javastanford-nlp

Illegal argument exception with Stanford Parser

I tried to parse a sentence using Stanford parser, but I get the exception. The input file, code and exception are specified below.

I think the problem is because the penn tree in input file doesn't handle punctuation. How do I generate a penn tree that handles punctuation too?

Input file

(ROOT
  (S
    (NP (DT A) (NN doctor) (NN investigation) (NN system) (NN (DIS)))
    (VP (VBZ is)
      (NP
        (NP (DT a) (NN part))
        (PP (IN of)
          (NP (DT a) (NN hospital) (NN information) (NN system) (NN (HIS).)))))))

code

            String str="-collapsed -treeFile temp.txt";
            String ar[]=str.split(" ");
            edu.stanford.nlp.trees.EnglishGrammaticalStructure.main(ar);

             try {
                FileOutputStream fw = new FileOutputStream("k.txt");
                PrintStream out = new PrintStream(fw);
                System.setOut(out);



            } catch (Exception e) {
                System.out.print(e);
            }

Exception raised :

Head is null: NN-37
Exception in thread "main" java.lang.IllegalArgumentException: governor or dependent cannot be null
        at edu.stanford.nlp.trees.UnnamedDependency.<init>(UnnamedDependency.java:105)
        at edu.stanford.nlp.trees.TreeGraphNode.dependencies(TreeGraphNode.java:519)
        at edu.stanford.nlp.trees.Tree.dependencies(Tree.java:1090)
        at edu.stanford.nlp.trees.GrammaticalStructure.<init>(GrammaticalStructure.java:71)
        at edu.stanford.nlp.trees.EnglishGrammaticalStructure.<init>(EnglishGrammaticalStructure.java:115)
        at edu.stanford.nlp.trees.EnglishGrammaticalStructure.<init>(EnglishGrammaticalStructure.java:89)
        at edu.stanford.nlp.trees.EnglishGrammaticalStructure.<init>(EnglishGrammaticalStructure.java:61)
        at edu.stanford.nlp.trees.EnglishGrammaticalStructure.<init>(EnglishGrammaticalStructure.java:53)
Solution

  • Unhelpful error message, but this is because the input tree is ill-formed: there is that stray period towards the end. The tree should be a well-formed s-expression.