"TypeError: data is already tz-aware UTC" in get_calendar()

[acowlikeobject]

from trading_calendars import get_calendar
get_calendar('CME')

Raises TypeError: data is already tz-aware UTC, unable to set specified tz: UTC.

This is a result of this issue in pandas where 'UTC' and pytz.UTC are not considered equal (I'm seeing this in pandas version 0.24.0).

I believe the fix would be to change the return in trading_calendars.trading_calendar.scheduled_special_times():

return pd.Series(
        index=pd.DatetimeIndex(days, tz='UTC'),
        data=days_at_time(days, time, tz=tz),
)

to:

return pd.Series(
        index=pd.DatetimeIndex(days, tz=pytz.UTC),
        data=days_at_time(days, time, tz=tz),
)

(Basically, replace 'UTC' with pytz.UTC when creating the index.)

[acowlikeobject]

For now, monkey patching works:

import pytz
import trading_calendars as tc

def scheduled_special_times(calendar, start, end, time, tz):
    days = calendar.holidays(start, end)
    return pd.Series(
        index=pd.DatetimeIndex(days, tz=pytz.UTC),
        data=tc.trading_calendar.days_at_time(days, time, tz=tz),
)
tc.trading_calendar.scheduled_special_times = scheduled_special_times
tc.get_calendar('CME')