t m one | Python Fiddle

__author__ = 'Dorota'

import collections, re

def find_all_digrams(corps, set_of_nouns, set_of_genitives, dict_of_digrams):
    for line in range(len(corps)) :
        for nr_elementu in range(len(corps[line])):
            word_last = corps[line][nr_elementu]
            if nr_elementu > 0 :
                word_first = corps[line][nr_elementu-1]
                if word_last in set_of_nouns and word_first in set_of_nouns:
                    key = word_first + " " + word_last
                    common_bonus(key, dict_of_digrams)
                    is_genitive_bonus(key, set_of_genitives, dict_of_digrams)
                    print line
                    print dict_of_digrams

def common_bonus(digram, dict_of_digrams):
    if dict_of_digrams.has_key(digram):
        dict_of_digrams[digram] += 1
    else:
        dict_of_digrams[digram] = 1

def is_genitive_bonus(digram, set_of_genitives, dict_of_digrams):
    pair = digram.split(" ")
    if pair[0] in set_of_genitives and pair[1] in set_of_genitives:
        dict_of_digrams[digram] += 3

corps = open('korpus_tm1.txt', 'r')
all_words_in_corps = map(lambda l: l.split(" "), corps.readlines())

list_of_nouns = open('lista_rzeczownikow.txt').read().split('\n')
set_of_nouns = set(list_of_nouns)

list_of_genitives = open('lista_dopelniaczy.txt').read().split('\n')
set_of_genitives = set(list_of_genitives)

dict_of_digrams = {}

funkcje_digramy.find_all_digrams(all_words_in_corps, set_of_nouns, set_of_genitives, dict_of_digrams)

Browser Version Not Supported

Due to Python Fiddle's reliance on advanced JavaScript techniques, older browsers might have problems running it correctly. Please download the latest version of your favourite browser.
Chrome 10+ Firefox 4+ Safari 5+
IE 10+

Let me try anyway!

Python Fiddle

Python Cloud IDE