How to convert from Tree type to String type in Python by nltk?

Question

for subtree3 in tree.subtrees():
  if subtree3.label() == 'CLAUSE':
    print(subtree3)
    print subtree3.leaves()

Using this code I able to extract the leaves of the tree. Which are: [('talking', 'VBG'), ('constantly', 'RB')] for a certain example. That is perfectly correct. Now I want this Tree elements to convert into string or in list for some further processing. How can I do that?

What I tried

for subtree3 in tree.subtrees():
  if subtree3.label() == 'CLAUSE':
    print(subtree3)
    print subtree3.leaves()
    fo.write(subtree3.leaves())
fo.close()

But it throws an error :

Traceback (most recent call last):
  File "C:\Python27\Association_verb_adverb.py", line 35, in <module>
    fo.write(subtree3.leaves())
TypeError: expected a character buffer object

I just want to store the leaves in a text file.

What is your input and what do you want as your output? Can you give an example? Also there can be multiple layers in your top-most subtree so depending on what output you require the way you traverse the tree to print it out would be different — alvas, Nov 24 '15 at 19:02

lguiel · Accepted Answer · 2016-11-11T05:02:02.450

It depends on your version of NLTK and Python. I think you're referencing the Tree class in the nltk.tree module. If so, read on.

In your code, it's true that:

subtree3.leaves() returns a "list of tuple" object and,
fo is a Python File IO object, the fo.write only receives a str type as a parameters

you can simply print the tree leaves with fo.write(str(subtree3.leaves())), thus:

for subtree3 in tree.subtrees():
    if subtree3.label() == 'CLAUSE':
        print(subtree3)
        print subtree3.leaves()
        fo.write(str(subtree3.leaves()))
fo.flush()
fo.close()

and don't forget to flush() the buffer.

@Iguiel That is exactly what I want. Thank you a lot. – Salah Nov 25 '15 at 12:35 — Salah, Nov 25 '15 at 12:35

score 3 · Answer 2 · edited May 23 '17 at 12:01

3

Possibly the question is more of trying to write a list of tuples to files instead of traversing the NLTK Tree object. See NLTK: How do I traverse a noun phrase to return list of strings? and Unpacking a list / tuple of pairs into two lists / tuples

To output a list of tuples of 2 strings, I find it useful to use this idiom:

fout = open('outputfile', 'w')

listoftuples = [('talking', 'VBG'), ('constantly', 'RB')]
words, tags = zip(*listoftuples)

fout.write(' '.join(words) + '\t' + ' '.join(tags) + '\n')

But the zip(*list) code might not work if there are multiple levels in your subtrees.

edited May 23 '17 at 12:01

Community

1
1

answered Nov 25 '15 at 09:19

alvas

115,346
109
446
738

I'm glad the answer helps =) – alvas Nov 25 '15 at 13:23
@alvas f we have multiple levels in our subtrees, how could we write the tree into a txt file? Because, as you said, it's not possible to use `zip(*list)` in this case – joasa Aug 09 '17 at 12:55

How to convert from Tree type to String type in Python by nltk?

2 Answers2