Modify PyUnit_ Time makes it support the time text of 'xx~xx months'

Because you need to use the function of converting time text into standard time , After investigating a variety of third-party packages, I chose pyunit_time, But there are still many unsatisfactory places , So copy out the whole bag , Modify yourself to meet your needs .

Found today pyunit_time I won't support it ‘xx~xx month ’ Statement of the , for example “6~9 month ” Will only recognize 9 month , however “6 month ~9 month ” Can recognize “6 month ” and “9 month ”, So it will “6~9 month ” Turn into “6 month ~9 month ” That's all right. , In fact, it is essentially a problem of omitting time units .

Here is the original code （ The file is /pyunit_time/filters.py）：

def get_time_key(string) -> list:
    """ Extract the time keyword in the string according to the string   The proportion ：  The night before national day 8 Half past six   return ： [' National Day ', ' The night before 8 Half past six '] :param string:  About the time string of dictation  :return:  Time keywords  """
    keys, start, end = [], -1, -1
    match = pattern.finditer(string)

    for key in match:
        start = key.start()
        if start == end:
            keys[-1] += key.group()
        else:
            keys.append(key.group())
        end = key.end()
    return keys

It uses regular matching ,pattern Call the regular expression prepared before , image “x year ”、‘x month ’ etc. , Individual numbers don't match any , So it will be ignored .

After modification ：

def get_time_key(string) -> list:
    """ Extract the time keyword in the string according to the string   The proportion ：  The night before national day 8 Half past six   return ： [' National Day ', ' The night before 8 Half past six '] :param string:  About the time string of dictation  :return:  Time keywords  """
    if '~' in string:  # lyc Modify the point , Support “xx~xx month ” Statement of the 
        begin, *middle, finish = string.split('~')
        if begin and finish and begin[-1].isdigit() and not finish[-1].isdigit():
            string = string.replace('~', finish[-1] + '~')

    keys, start, end = [], -1, -1
    match = pattern.finditer(string)

    for key in match:
        start = key.start()
        if start == end:
            keys[-1] += key.group()
        else:
            keys.append(key.group())
        end = key.end()
    return keys

Add the last time unit to... Before the original processing “~” After the previous separate number , Completed the conversion , Retest and find that you can recognize “6~9 month ” 了 .