Python conditional joining of *consecutive* strings that don't end in punctuation with those that do

Question

I have a list of words,

list1 = ['hello', 'how', 'are', 'you?', 'i', 'am', 'fine', 'thanks.', 'great!']

Which I would like to join to be,

list2 = ['hello how are you?', 'i am fine thanks.', 'great!']

Is there an simple pythonic way to do this? I have considered doing an itertools.groupby join but the problem is that all of the elements of my group don't have the same criteria (I can't just query if they all have punctuation). Basically, whether or not element x gets included is a function of potentially element x+n where n can be large. This complicates the problem.

Answer 1

Don't use groupby(); you'd get separate groups for those words with and without punctuation, which you then have to re-combine.

Use a generator function:

import string

def sentence_groups(l, punctuation=tuple(string.punctuation)):
    group = []
    for w in l:
        group.append(w)
        if w.endswith(punctuation):
            yield group
            group = []
    if group:
        yield group

The generator collects words from the input list until one ends with punctuation, at which point that whole group is yielded, after which the group is cleared for a new group.

When iteration ends and there are still words in the group, that last group is yielded too (even though they don't have punctuation at the end).

Use this together whith str.join() to produce your output:

>>> list1 = ['hello', 'how', 'are', 'you?', 'i', 'am', 'fine', 'thanks.', 'great!']
>>> [' '.join(group) for group in sentence_groups(list1)]
['hello how are you?', 'i am fine thanks.', 'great!']

I used all punctuation in the string.punctuation string; this is quite broad:

>>> string.punctuation
'!"#$%&\'()*+,-./:;<=>?@[\\]^_`{|}~'

Pass in a tuple of specific punctuation characters as the second argument if you wanted to narrow that down, or hardcode your own definition.

Answer 2

A humble solution:

import string

words = ['hello', 'how', 'are', 'you?', 'i', 'am', 'fine', 'thanks.', 'great!']
sents = []

range_flag = 0
for index, word in enumerate(words):
  if word[-1] in string.punctuation:
    sents.append(words[range_flag+1:index+1])
    print(range_flag, index)
    range_flag = index

print([" ".join(s) for s in sents])

(0, 3)
(3, 7)
(7, 8)
['how are you?', 'i am fine thanks.', 'great!']