Creates a dictionary object with the 3 most common prepositions for every verb in the NLTK Brown Corpus. Export ConditionalFrequencyDist Object as a JSON

prepositions.py

import nltk
from nltk.corpus import brown
import json

prepchoices = nltk.ConditionalFreqDist((v[0], p[0]) 
    for (v, p) in nltk.bigrams(brown.tagged_words(tagset="universal")) 
        if v[1] == "VERB" and p[1] == "ADP") 

pc = dict()
for word in prepchoices.conditions():
    # get three most common prepositions
    word_dist = prepchoices[word].most_common(3)

    # convert [('with', 5), ('to', 4), ('about', 3)]
    # to ['with', 'to', 'about']
    word_dist_clean = [x[0] for x in word_dist]

    pc[word] = word_dist_clean

with open('prepositions.json', 'w') as fp:
    json.dump(pc, fp)