modular-bot-framework-for-telegram/modules/troll-spelling-corrector/index.py

import re

''' Grammar reference
all_possible_match_strings = [
    "силка",  "силки",
    "силки",  "силок",
    "силці",  "силкам",
    "силку",  "силки",
    "силкою", "силками",
    "силці",  "силках",
    "силко",  "силки",

    "лінк",   "лінка",  "лінки",
    "лінку",  "лінки",  "лінок",
    "лінку",  "лінці",  "лінкам",
    "лінк",   "лінку",  "лінки",
    "лінком", "лінкою", "лінками",
    "лінку",  "лінці",  "лінках",
    "лінке",  "лінко",  "лінки"
]
'''

EXTRACT_PADDING = 11
EXTRACT_INCLUDE_WHOLE_WORDS = True

corrections = [       # IDs
    "посилання",      # 0
    "посиланню",      # 1
    "посиланням",     # 2
    "на посиланні",   # 3
    "посилань",       # 4
    "посиланнями",    # 5
    "посиланнях"      # 6
]

replacements = [
    ["силка", "силки", "силку", "силко",
     "лінк", "лінка", "лінки", "лінку", "лінке", "лінко"],
    ["силці",
     "лінку", "лінці"],
    ["силкам", "силкою",
     "лінкам", "лінком", "лінкою"],
    ["на силці",
     "на лінку", "на лінці"],
    ["силок",
     "лінок"],
    ["силками",
     "лінками"],
    ["силках",
     "лінках"]
]

#unique_match_strings = set(all_possible_match_strings)

#ua_alphabet = "абвгґдеєжзиіїйклмнопрстуфхцчшщьюя"
#regex_matchers = [re.compile(fr"((?<=[^{ua_alphabet}])|(?<=\b)|(?<=^)|(?<= )){i}((?=[^{ua_alphabet}])|(?=\b)|(?=$)|(?= ))", re.DEBUG)
#        for i in unique_match_strings]

ua_alphabet = "абвгґдеєжзиіїйклмнопрстуфхцчшщьюя"

for i, group in enumerate(replacements):
    for j, match_word in enumerate(group):
        #replacements[i][j] = re.compile(fr"((?<=[^{ua_alphabet}])|(?<=\b)|(?<=^)|(?<= )){i}((?=[^{ua_alphabet}])|(?=\b)|(?=$)|(?= ))")
        replacements[i][j] = [match_word, re.compile(fr"((?<=[^{ua_alphabet}])|(?<=\b)|(?<=^)|(?<= )){match_word}((?=[^{ua_alphabet}])|(?=\b)|(?=$)|(?= ))")]
        #print(replacements[i][j])

#_ = [print(fr"(?<=[^абвгґдеєжзиіїйклмнопрстуфхцчшщьюя]){i}(?=[^абвгґдеєжзиіїйклмнопрстуфхцчшщьюя])") for i in unique_match_strings]

def process(message, path):
    lowercase_message = message.text.lower()

    for correct_word_id, group in enumerate(replacements):
        for match_word, matcher in group:
            result = matcher.search(lowercase_message)

            if result:
                l = len(message.text)

                mistake_start = result.start()
                mistake_end = result.end()

                print(mistake_start, mistake_end)

                original_text_before = message.text[max(mistake_start-EXTRACT_PADDING,0):mistake_start]
                original_text_after = message.text[mistake_end:min(mistake_end+EXTRACT_PADDING,l)]
                original_text_mistake = message.text[mistake_start:mistake_end]
                
                if EXTRACT_INCLUDE_WHOLE_WORDS:
                    while 0 <= mistake_start - EXTRACT_PADDING - 1 < l and \
                            message.text[mistake_start-EXTRACT_PADDING-1].isalnum():
                        mistake_start -= 1
                        original_text_before = message.text[max(mistake_start-EXTRACT_PADDING,0):result.start()]

                    while 0 <= mistake_end + EXTRACT_PADDING < l and \
                            message.text[mistake_end+EXTRACT_PADDING].isalnum():
                        mistake_end += 1
                        original_text_after = message.text[result.end():min(mistake_end+EXTRACT_PADDING,l)]


                if len(message.text[:mistake_start]) > EXTRACT_PADDING:
                    original_text_before_continue = "..."
                else:
                    original_text_before_continue = ""

                if len(message.text[mistake_end:]) > EXTRACT_PADDING:
                    original_text_after_continue = "..."
                else:
                    original_text_after_continue = ""

                original_extract = original_text_before_continue + original_text_before \
                        + original_text_mistake + original_text_after + original_text_after_continue

                correct_word = corrections[correct_word_id]
                if original_text_mistake == match_word.capitalize():
                    correct_word = corrections[correct_word_id].capitalize()
                elif original_text_mistake == match_word.upper():
                    correct_word = corrections[correct_word_id].upper()

                fixed_extract = original_text_before_continue + original_text_before \
                        + correct_word + original_text_after + original_text_after_continue

                return f'"{original_extract}" -> "{fixed_extract}" 🌚', None
    else:
        return "", None