A couple functions which use Pandas to work with a year which runs from Sept to Aug. Each function only wraps a single line of Pandas calls, so their value is limited. But, remembering how to accomplish the things they do is valuable.

pd_service_year.py

import pandas

def get_service_year(date):
    ''' 
    Return service year for given date.

    The service year runs from Sept to Aug and Sept-Dec are in the service year of
    the following calendar year. In other words, the date 2020-09 will return 2021.

    Accepts date as an ISO formatted string (for example YYYY-MM), a datetime.date
    object, a datetime.datetime object, or any of Pandas' date objects.
    '''
    return pandas.Period(date, freq='M').asfreq('A-AUG').year

def get_months_of_service_year(year):
    '''
    Return a list of months in given service year as a pandas.PeriodIndex.

    The service year runs from Sept to Aug and Sept-Dec are in the service year of
    the following calendar year. Therefore, the year 2021 would return a list of
    dates from 2020-09 to 2021-08.

    Accepts year as an integer, an ISO formatted string, a datetime.date object,
    a datetime.datetime object, or any of Pandas' date objects.

    As a pandas.PeriodIndex is returned, one can test if a date is `in` the year.
    For example:

        >>> yr = get_months_of_service_year(2021)
        >>> '2020-11' in yr
        True
        >>> '2020-08' in yr
        False
        >>> datetime.date(2021, 2, 1) in yr
        True

    For a native Python list, do:

        >>> y.astype(str).to_list()
        ['2020-09', '2020-10', ... '2021-08']
        >>> [datetime.date(p.year, p.month, 1) for p in y]
        [datetime.date(2020, 9, 1), ... datetime.date(2021, 8, 1)]
    '''
    return pandas.period_range(
        end=pandas.Period(year, freq='A-AUG'),
        periods=12,
        freq='M'
    )