TResample报错TypeError: Only valid with DatetimeIndex

🐛 Bug Description

使用TResample将1min数据转为30min K线时报错

To Reproduce

from qlib.data.ops import TResample, Feature
from qlib.data import D
fea = "TResample($open, '30min', 'first')"
D.features(instruments=['BTCUSDT'], fields=[fea], start_time='2022-01-01 9:00:00', end_time='2022-01-01 20:00:00', freq='1min')

报错

TypeError: Only valid with DatetimeIndex, TimedeltaIndex or PeriodIndex, but got an instance of 'RangeIndex'

原因是self.feature.load返回的series的索引是RangeIndex，并不能直接放到pandas的resample里使用，resample只接受Datetime的Index

class TResample(ElemOperator):
    ... ...
    def _load_internal(self, instrument, start_index, end_index, *args):
        series = self.feature.load(instrument, start_index, end_index, *args) 

        if series.empty:
            return series
        else:
            if self.func == "sum":
                return getattr(series.resample(self.freq), self.func)(min_count=1)
            else:
                return getattr(series.resample(self.freq), self.func)()

望修复，谢谢

Start at TResample._load_internal and inspect the index returned by self.feature.load before the pandas resample call. Re-run the supplied D.features example with the 1-minute BTCUSDT data and confirm that 30-minute K-line resampling completes without the DatetimeIndex TypeError.