preprocess/utils/common/extracters.py

import re
from constants.constants import *
from preprocess.utils.common.utils import find_full_word, remove_full_words, detect_language_simple

def remove_brand_from_name(name, brand):
    pattern = r'\b' + brand + r'\b'
    name_wo_brand = re.sub(pattern, '', name).strip()
    if (len(name_wo_brand) < 3):
        return name

    return name_wo_brand


def insert_brand_in_name(name, brand):
    pattern = r'\b' + brand + r'\b'
    if not re.search(pattern, name):
        return brand + " " + name

    return name


def extract_spark(text, remove=False):
    # Handle specific cases when 'champagne' or 'шампань' is not drink type, but region
    if 'коньяк' in text:
        return None, text

    types_variations = TYPES_VARIATIONS_DICT['шампанское']
    type_str = find_full_word(text, types_variations)
    if type_str:
        if remove:
            text = text.replace(type_str, ' ')
        return 'шампанское', text

    return None, text


def extract_color(text):
    colors_variations = COLORS_VARIATIONS_DICT
    for color in colors_variations.keys():
        found_color_str = find_full_word(text, colors_variations[color])
        if found_color_str:
            return color, found_color_str

    return None, None


def extract_sour(text):
    sours_variations = SOURS_VARIATIONS_DICT
    for sour in sours_variations.keys():
        found_sour_str = find_full_word(text, sours_variations[sour])
        if found_sour_str:
            return sour, found_sour_str

    return None, None

def extract_stars(text):
    match = re.search(r'((одна|две|три|четыре|пять|шесть|семь|восьмь|девять|\d{1,2})\s*(звезды|звезд|звёзд|зв))', text, re.IGNORECASE)
    if match:
        return match.group(0), match.group(0)

    return None, None


def extract_color_and_sour(text, remove=False):
    #special_name = find_full_word(text, SPECIFIC_NAMES)
    #if special_name:
    #    text = text.replace(special_name, '###SPECIAL_NAME###')

    found_color, found_color_str = extract_color(text)
    found_sour, found_sour_str = extract_sour(text)

    #Additional cases for not wine products (put these attributes to colod and sour for simplicity)
    if not found_color:
        found_color, found_color_str = extract_stars(text)

    if remove:
        if found_color_str and (detect_language_simple(found_color_str) == 'ru') and not found_color_str == 'розе':
            text = remove_full_words(text, [found_color_str])

        if found_sour_str:
            text = remove_full_words(text, [found_sour_str])

    #if special_name:
    #    text = text.replace('###SPECIAL_NAME###', special_name)

    return found_color, found_sour, text

def extract_other_attributes(text, remove=False):
    other = []

    for attr in OTHER_ATTRIBUTES_VARIATIONS_DICT.keys():
        attr_word = find_full_word(text, [attr])
        if not attr_word:
            attr_word = find_full_word(text, OTHER_ATTRIBUTES_VARIATIONS_DICT[attr])

        if attr_word:
            text = remove_full_words(text, [attr_word])
            other.append(attr_word)

    return other, text


def extract_years(text, remove=False):
    """
    Извлекает сочетание числа и слова, указывающего возраст (например: '50 лет', '21 years').
    # Добавить поддержку выражений вида "не менее|for at least 3 лет"
    """
    # Регулярное выражение ищет числа и слова 'лет' или 'years' с учетом регистра
    # ron piet 20 y.o 40%
    #match = re.search(r'\b(?<!\d)((\d{1,2})\s*(лет|years old|years|y.o))\b', text, re.IGNORECASE)
    #match = re.search(r'\b(?<!\d)(\d{1,2})\s*(лет|years old|years|y.o|ans)\b', text, re.IGNORECASE)
    #match = re.search(r'\b(?<!\d)((\d{1,2})(\s*|\-)(лет|года|yo|year old|years old|years|ans))\b', text, re.IGNORECASE)
    match = re.search(r'\b(?<!\d)((\d{1,2})(\s*|\-)(лет|ти летний|летний|года|yo|y o|year old|year|years old|years|ans))\b', text, re.IGNORECASE)
    if not match:
        match = re.search(r'\b((трех|четырех|пяти|шести|семи|восьми|девяти)летний)\b', text, re.IGNORECASE)
        if not match:
            match = re.search(r'выдержка (\d{1,2}\s*(лет|л))', text, re.IGNORECASE)

    if match:
        if remove:
            text = text.replace(match.group(0), ' ')
            text = text.replace('выдержка', ' ').replace('aging', ' ').replace('ageing', ' ')
        return match.group(2), text

    return None, text


def extract_production_year(text, remove=False):
    """
    Извлекает год производства (четырехзначное число в диапазоне 1900–2099) из строки.
    Например: '2019'.
    """
    #pattern = r'\b(19\d{2}|20\d{2})[гГ]*\b'
    #pattern = r'\b(19\d{2}|20\d{2})\s*[гГ]*\.*(?:\b|$)'
    #pattern = r'\b(19\d{2}|20\d{2})\s*[гГ]*(\/y)*\.*(?:\b|$)'
    pattern = r'\b(19\d{2}|20\d{2})\s*(г/у|г)*\.*(?:\b|$)'
    #r'\b' + str(production_year) + r'\s*[гГ]*\.*(?:\b|$)', ' ', text
    match = re.search(pattern, text)
    if match:
        if remove:
            text = re.sub(pattern, ' ', text).replace("г/у", " ").replace("год урожая", " ")
        return match.group(1), text

    return None, text


def extract_alcohol_content(text, remove=False):
    """
    Извлекает содержание алкоголя из строки.
    Например: '40% или 10-12%'.
    """
    #    pattern = r'(\d{1,2}(?:[.,]\d+)?\s*%)'
    pattern = r'\b(\d{1,2}(?:[.,]\d+)?(?:\s*\-\s*\d{1,2}(?:[.,]\d+)?)?\s*%)'
    match = re.search(pattern, text)
    if match:
        if remove:
            text = re.sub(pattern, ' ', text)
        # Заменяем запятую на точку для единообразия (если нужно)
        return match.group(1).replace(' ', '').replace(',', '.'), text

    return None, text


def is_volume(value):
    """
    Проверяет, является ли значение валидным объемом (<= 10 литров).
    """
    try:
        volume = float(value)
        return volume if volume <= 10 else None
    except ValueError:
        return None


def extract_volume_or_number(text, remove = False):
    """
    Извлекает объем в литрах или число с плавающей точкой из строки.
    Например: '0,75л', '0.5', или '1,5 л'.
    Добавить: обработку мл, например '750 мл'
    """
    # Попытка найти объем с буквой 'л' или без пробела перед ней
    pattern = r'((\d+(?:[\.,]\d+)?\s*[лl]|(?:\d+(?:[\.,]\d+)?[лl])))(\s*[/хx*]\s*(\d{1,2}))*'
    match_with_l = re.search(pattern, text)
    if match_with_l:
        vol = is_volume(match_with_l.group(1).replace(',', '.').replace('л', '').replace('l', '').strip())
        if vol:
            if remove:
                text = re.sub(pattern, ' ', text)
            return vol, text

    # Если не найдено число с буквой л, то ищем просто число с плавающей точкой (и возможно '*' после него)
    pattern = r'(?<!№)\b(\d{1,2}(?:[\.,]\d+))(\s*[/хx*]\s*(\d{1,3}))*(?!\%)\b(?!\s*(№|-er|er|\d{3,}))'
    match_number = re.search(pattern, text)
    if match_number:
        #print("match_number: " + str(match_number.group(1)) + " for [" + text + "]")
        #return is_volume(match_number.group(1).replace(',', '.')), match_number.group(1)
        vol = is_volume(match_number.group(1).replace(',', '.').strip())
        if vol:
            if remove:
                text = re.sub(pattern, ' ', text)
            return vol, text

    return None, text


def extract_gb(text, remove=False):
    if not isinstance(text, str) or not text:
        return None, text

    gb = find_full_word(text, GBS)
    if gb:
        text = text.replace(gb, ' ')

        gb2 = find_full_word(text, GBS)
        if gb2:
            text = text.replace(gb2, ' ')


    return gb, text


def extract_type_by_brand_name(text):
    for k, v in TYPES_FROM_BRAND_DICT.items():
        if find_full_word(text, [k]):
            return v

    return None


def extract_specific_brands_from_name(name):
    return find_full_word(name, SPECIFIC_BRANDS)


# Finds specific brands in names that makes collisions with parsing routine.
# For example, brand 'Voskevaz 1932' makes parsing algorithm to think that '1932' is year of the wine
def replace_specific_brand_and_name(text):
    special_brand = find_full_word(text, SPECIFIC_BRANDS)
    if special_brand:
        text = text.replace(special_brand, '###special_brand###')

    special_name = find_full_word(text, SPECIFIC_NAMES)
    if special_name:
        text = text.replace(special_name, '###special_name###')

    return text, special_brand, special_name


def restore_specific_brand_and_name(text, special_brand=None, special_name=None):
    if special_name:
        text = text.replace('###special_name###', special_name)

    if special_brand:
        text = text.replace('###special_brand###', special_brand)

    return text


def extract_attributes_from_name(text):
    gb, text = extract_gb(text, True)
    alcohol, text = extract_alcohol_content(text, True)
    years, text = extract_years(text, True)
    production_year, text = extract_production_year(text, True)
    volume_or_number, text = extract_volume_or_number(text, True)
    color, sour, text = extract_color_and_sour(text, True)
    other, text = extract_other_attributes(text, True)

    return text, alcohol, volume_or_number, years, production_year, gb, color, sour, other