search_dob

# Search the date of birth 
def search_dob(text): 
 
    # Define the pattern to get dates (20-07-2018, 20 07 2018, 20-07 2018, 20 07-2018) 
    # reg_dob = '(\d{2}[\/ -]\d{2}[\/ -]\d{2,4})' 
    # reg_dob = '(((0[1-9]|([1-2][0-9])|3[01])[-./  \s]?((0[13578])|(10|12)))|((0[1-9]|([1-2][0-9])|3[0])[-./ \s]?((0[469])|11))|((0[1-9]|[1-2][0-9])[-./ \s]?(02)))[-./ \s]?[\d]{2,4}' 
    reg_dob = '((((0[\ ]*[1-9][\ ]*|([1-2][\ ]*[0-9][\ ]*)|3[\ ]*[01][\ ]*)[-.—/  \s]*((0[\ ]*[13578][\ ]*)|(1[\ ]*0[\ ]*|1[\ ]*2[\ ]*)))|((0[\ ]*[1-9][\ ]*|([1-2][\ ]*[0-9][\ ]*)|3[\ ]*[0][\ ]*)[-./ —\s]*((0[\ ]*[469][\ ]*)|1[\ ]*1[\ ]*))|((0[\ ]*[1-9][\ ]*|[1-2][\ ]*[0-9][\ ]*)[-./ —\s]*(0[\ ]*2[\ ]*)))[-./ —\s]*([\d][\ ]*){2,4})'

    # Find all strings that match 
    dates = re.finditer(reg_dob, text) 
    temp = [date.group() for date in dates] 
    dates = temp 
    # for date in dates: 
    #     print(date.group()) 

    # print(dates)
    
    # We iterate through the results and convert it into datetime format 
    for i in range(len(dates)): 
        date = re.sub('[-— ]', '', dates[i])
        # If invalid string format was read, we set its value to zero
        reg_year = '[0-9]+'
        year = re.match(reg_year, date[4:])
        
        year = int(year[0])
        print(year)
        
        if year < 1900 or year > 2000:
            dates[i] = 0
            continue
        # print(dates[i])
        dates[i] = datetime.strptime(date, '%d%m%Y')
    # print(dates)
    # We remove those invalid read strings
    dates = [date for date in dates if not isinstance(date, int)]
    
    # We take the smallest dates in a KTP, beacuse it must be his/her birth date 
    date = min(dates)
 
    return datetime.strftime(date, '%d/%m/%Y')