使用正则表达式来约束元组列表

Question

Given a list of tuples of words and their part-of-speech from a sentence:

[('We', 'PRP'),
 ('took', 'VBD'),
 ('advantage', 'NN'),
 ('of', 'IN'),
 ('the', 'DT'),
 ('half', 'JJ'),
 ('price', 'NN'),
 ('sushi', 'NN'),
 ('deal', 'NN'),
 ('on', 'IN'),
 ('saturday', 'NN')]

I would like to extract terms that have certain PoS sequences using a regexp. This would be something like ('JJ')*('NN')+ so I have a list of [('advantage', 'half price sushi deal', 'saturday')] . What is the most appropriate way of carrying out such a task, bearing in mind I will be doing this for hundreds of sentences?

Thank you!

Answer 1

I think this might be something that will do the trick:

a = [('We', 'PRP'),
 ('took', 'VBD'),
 ('advantage', 'NN'),
 ('of', 'IN'),
 ('the', 'DT'),
 ('half', 'JJ'),
 ('price', 'NN'),
 ('sushi', 'NN'),
 ('deal', 'NN'),
 ('on', 'IN'),
 ('saturday', 'NN')]

b = iter(a[1:])

my_list = []
inner_list = []
accepted = ['JJ', 'NN']

for item in a:
    word = item[0]
    check = item[1]
    try:
        against = next(b)
        if check in accepted:
            if against[1] not in accepted:
                inner_list.append(word)
                my_list.append(inner_list)
                inner_list = []
            else:
                inner_list.append(word)
    except StopIteration:
        if check in accepted:
             inner_list.append(word)
             my_list.append(inner_list)
final = [' '.join(item) for item in my_list]