从元组列表中搜索索引

Question

I have a sentence = "hi there, my car number is H 11231, and my card number is 11122" . I tokenized the sentence then POS tagged the tokenized sentence. I want to grab the car number, I created a loop to check if the index is at a number lets say (11231). Then check 1 tuple before or after if it has the tag NNP (which stands for 1 Letter)

import nltk

sentence = 'hi there, my car number is H 11231, and my card number is 11122'

tokenizedSent = nltk.word_tokenize(sentence)
tagged = nltk.pos_tag(tokenizedSent)
print(tagged)
output = []

print(tagged)
for i in tagged: 
    if i[1] == 'CD':
        output.append(i[0])
    elif i[1] == 'NNP':
        output.append(i[0])

The sentence has two numbers which are 11231 and 11122. However, only one of them is the car number which is the one before is tagged by NNP

Answer 1

Your solution traverse the collection of tag and only takes CD and NNP tags.

You want only numbers.

so first thing to do : get the numbers :

IndexTag = namedtuple('IndexTag', ['tag', 'index'])
numbers = []
for i, tag in enumerate(tagged):
    if tag[1] == 'number':
        numbers.append(IndexTag(tag, i))

now that you have your numbers you can check that the "previous tag is NNP" :

car_ids = []
for number in numbers:
    if number.index > 0 and tagged[number.index - 1][1] == 'NNP':
       car_ids.append(number)