#s:(str)        - a string we break down into ngrams,
#w:(int)        - width of ngrams,
#stride:(int)   - how many indexes to move along the input
#discard:(bool) - whether or not to discard any remaining ngrams when they are less than the specified width w.
def ngrams(s, w=1, stride=None, discard=False):
  #example usage #000:
  #ngrams("foobarbazk", 3, 3)
  #expected output:
  #  ['foo', 'bar', 'baz', 'k']

  #example usage #001:
  #ngrams("foobarbazk", 3, 2)
  #expected output:
  #  ['foo', 'oba', 'arb', 'baz', 'zk']

  #example usage #002:
  #ngrams("foobarbazk", 3, 3, True)
  #expected output:
  #  ['foo', 'bar', 'baz']

  if stride==None:
    stride=w
  results = []
  s=s.lower()
  i = 0
  if w < len(s):
    results.append(s[0:w])
  else:
    return [s]

  i=i+stride

  while i < len(s):
    result = s[i:i+w]
    if len(result) == w or discard==False:
      results.append(s[i:i+w]) #in a strict language this would probably be an overflow
    i = i + stride

  return results