Quick implementation of character n-grams for word
To generate bigrams:
In [8]: b='student'
In [9]: [b[i:i+2] for i in range(len(b)-1)]
Out[9]: ['st', 'tu', 'ud', 'de', 'en', 'nt']
To generalize to a different n
:
In [10]: n=4
In [11]: [b[i:i+n] for i in range(len(b)-n+1)]
Out[11]: ['stud', 'tude', 'uden', 'dent']
Try zip
:
>>> def word2ngrams(text, n=3, exact=True):
... """ Convert text into character ngrams. """
... return ["".join(j) for j in zip(*[text[i:] for i in range(n)])]
...
>>> word2ngrams('foobarbarblacksheep')
['foo', 'oob', 'oba', 'bar', 'arb', 'rba', 'bar', 'arb', 'rbl', 'bla', 'lac', 'ack', 'cks', 'ksh', 'she', 'hee', 'eep']
but do note that it's slower:
import string, random, time
def zip_ngrams(text, n=3, exact=True):
return ["".join(j) for j in zip(*[text[i:] for i in range(n)])]
def nozip_ngrams(text, n=3):
return [text[i:i+n] for i in range(len(text)-n+1)]
# Generate 10000 random strings of length 100.
words = [''.join(random.choice(string.ascii_uppercase) for j in range(100)) for i in range(10000)]
start = time.time()
x = [zip_ngrams(w) for w in words]
print time.time() - start
start = time.time()
y = [nozip_ngrams(w) for w in words]
print time.time() - start
print x==y
[out]:
0.314492940903
0.197558879852
True
Although late, NLTK has an inbuilt function that implements ngrams
# python 3
from nltk import ngrams
["".join(k1) for k1 in list(ngrams("hello world",n=3))]
['hel', 'ell', 'llo', 'lo ', 'o w', ' wo', 'wor', 'orl', 'rld']