Say, 1 dog, 2 dogs – it’s easy in English, but not so simple in Russian.

Whatever language you are working with, try to find a general rule.With Russian, it turned out begin quite simple – for numbers ending with [1], [2, 3, 4] and [5, 6, 7, 8, 9] there are generally 3 forms. Except “teens” – [11, 12, ..., 19] that are in the third form no matter what.

If the words are unknown to you, you may want to factors things that word forms. But it is separate algorithm. This one is about sending a number and getting a word form.

def find_base(base_word, f):
  """
    base_word - first form of the objects' name
    f - form number (0 to 2)
  """
  l = {
    'фраза': ['фраза', 'фразы', 'фраз'],
    'тема': ['тема', 'темы', 'тем'],
    # many more
  }
  return l.get(base_word, ['']*3)[f]


def find_word(obj, base_word):
  """
    obj - quantity of objects
    base_word - first form of the objects' name
  """

  obj_int = int(obj)  # to determine 'teens'
  obj = str(obj).strip()  # to get last digit

  if obj.endswith('1'):
    if obj_int == 1 or obj_int > 20: k = 0
    else: k = 2
  elif obj.endswith('2') or obj.endswith('3') or obj.endswith('4'):
    if obj_int < 10 or obj_int > 20: k = 1
    else: k = 2
  else: k = 2
  return find_base(base_word, k)

It is not as short as possible, but it’s easily readable. I aim for that.

Basically, this code:

>> find_word(11, 'фраза')

Return

>> фраз

This thing I used for a statistics module for Lingomost.