transforming wordnet txt into lists in python nltk

Question

I am running the following function:

import nltk
from nltk.corpus import wordnet as wn

def noun_names(list): 
    for synset in list:
        for lemma in synset.lemmas():
            print lemma.name()

noun_names(list(wn.all_synsets(wn.NOUN)))

and it returns a long list of all the names of nouns in wordnet:

eg

epoch
Caliphate
Christian_era
Common_era
day
year_of_grace
Y2K
generation
anniversary

How do I take this output, which is neither a string or a list, and turn it into a list? Thanks so much.

Answer 1

Instead of printing to the stdout with your:

print lemma.name()

Why not append it to a list and return the list?

def noun_names(list):
    names = []
    for synset in list:
        for lemma in synset.lemmas():
            names.append(lemma.name())
    return names

names = noun_names(list(wn.all_synsets(wn.NOUN)))

Answer 2

It's not returning anything. Your function is printing , not returning.

You need to return a list. As a side note, you should rename your function parameter from list . You'll clobber something unintentionally with that.

One option is to modify your function slightly to append to a list and then return that:

def noun_names(word_list):
    lemma_list = []
    for synset in word_list:
        for lemma in synset.lemmas():
            lemma_list.append(lemma.name())
    return lemma_list

Another option is to change the above into a list comprehension:

def noun_names(word_list): 
    return [lemma.name() for synset in word_list for lemma in synset.lemmas()]

---

Both of these function return a list with the same information. Notice that I removed the list() function wrapper around wn.all_synsets(wn.NOUN) since the function returns a list .

lemma_list1 = noun_names(wn.all_synsets(wn.NOUN))
lemma_list2 = noun_names_1(wn.all_synsets(wn.NOUN))
print len(lemma_list1), len(lemma_list2), len(lemma_list1) == len(lemma_list2), lemma_list1 == lemma_list2

That final print statement outputs:

146347 146347 True True

This shows that both lists have the same number of elements ( 146347 each and the first True ) and the lists themselves are equal. A more appropriate test in the code is:

assert len(lemma_list1) == len(lemma_list2)
assert lemma_list1 == lemma_list2

If the lists aren't the same length or equal the assert statement will throw an exception.

Answer 3

If you just need the list of lemma, check out OMW (Open Multilingual WordNet) http://compling.hss.ntu.edu.sg/omw/

$ wget http://compling.hss.ntu.edu.sg/omw/wns/eng.zip
$ unzip eng.zip
$ cut -f3 eng/wn-data-eng.tab | (read;cat)