word count

wordcount.py

import json
import re
filename = './swedish.txt'

print('Start reading')

dictionary = {}
with open(filename, encoding="utf-8") as file:
    while line := file.readline():
        line = line.rstrip().split(' ')
        for word in line:
            word = word.strip('.').strip('-').strip("\\").strip(',').strip('?')
            if word[:1].isdigit() or word == "" or re.match(r'^[a-zA-Z]+\)', word):
                continue
            print('.') 
            if word not in dictionary:
                dictionary[word] = 1
            else:
                dictionary[word] += 1

sorted_dictionary = dict(sorted(dictionary.items()))

with open('result.json', 'w', encoding="utf-8") as fp:
    json.dump(sorted_dictionary, fp, ensure_ascii=False)