jackieju / penn_tree

Simple tree builder for penn string like (ROOT(SBAR(...), alternative of nltk.tree

penn_tree

Simple tree builder for penn string like (ROOT(SBAR(...)

An alternative of nltk.tree You can use it to parse output string of stanford sentence structure parser.

API:

1. parse from a penn string:

s = """
(ROOT
  (SBAR
      (WHADJP (WRB how) (JJ many))
          (S
                (NP (RB not) (NN ar) (NNS invoices))
                      (VP (VBP have) (RB not)
                              (VP (VBN been)
                                        (VP
                                                    (ADVP (RB partially))
                                                                (VBN paid)))))))
                                                                
                                                                
"""
tree = Tree.from_tree_string(s)

A tree object will be built.

2. Output string as a tree

print(tree.to_tree_string())

output:

how many not ar invoices have not been partially paid

3. Output stentence content

print(tree.content_string())

output:

ROOT->
  SBAR->
    WHADJP->
      WRB=how
      JJ=many
    S->
      NP->
        RB=not
        NN=ar
        NNS=invoices

      VP->
        VBP=have
        RB=not
        VP->
          VBN=been
          VP->
            ADVP->
              RB=partially

            VBN=paid

Sample:

s = """
(ROOT
  (SBAR
      (WHADJP (WRB how) (JJ many))
          (S
                (NP (RB not) (NN ar) (NNS invoices))
                      (VP (VBP have) (RB not)
                              (VP (VBN been)
                                        (VP
                                                    (ADVP (RB partially))
                                                                (VBN paid)))))))
                                                                
                                                                
"""
print(s)
tree = Tree.from_tree_string(s)
print(tree.content_string())
print(tree.to_tree_string())

output:

(ROOT
  (SBAR
      (WHADJP (WRB how) (JJ many))
          (S
                (NP (RB not) (NN ar) (NNS invoices))
                      (VP (VBP have) (RB not)
                              (VP (VBN been)
                                        (VP
                                                    (ADVP (RB partially))
                                                                (VBN paid)))))))
                                                                
                                                                

how many not ar invoices have not been partially paid
ROOT->
  SBAR->
    WHADJP->
      WRB=how
      JJ=many

    S->
      NP->
        RB=not
        NN=ar
        NNS=invoices

      VP->
        VBP=have
        RB=not
        VP->
          VBN=been
          VP->
            ADVP->
              RB=partially

            VBN=paid