transport.py 文件源码

def hourly_resample(df, bse=0, minutes=60):
    """
    Args:
        df:
            pandas dataframe containing time series needing resampling
        bse (int):
            base time to set; optional; default is zero (on the hour);
        minutes (int):
            sampling recurrence interval in minutes; optional; default is 60 (hourly samples)
    Returns:
        A Pandas DataFrame that has been resampled to every hour, at the minute defined by the base (bse)
    Description:
        see http://pandas.pydata.org/pandas-docs/dev/generated/pandas.DataFrame.resample.html for more info
        This function uses pandas powerful time-series manipulation to upsample to every minute, then downsample to every hour,
        on the hour.
        This function will need adjustment if you do not want it to return hourly samples, or iusgsGisf you are sampling more frequently than
        once per minute.
        see http://pandas.pydata.org/pandas-docs/stable/timeseries.html#offset-aliases
    """
    if int(str(pd.__version__).split('.')[0]) == 0 and int(str(pd.__version__).split('.')[1]) < 18: # pandas versioning
        df = df.resample('1Min')
    else:
        # you can make this smaller to accomodate for a higher sampling frequency
        df = df.resample('1Min').first()  

    # http://pandas.pydata.org/pandas-docs/dev/generated/pandas.Series.interpolate.html
    df = df.interpolate(method='time', limit=90)

    if int(str(pd.__version__).split('.')[0]) == 0 and int(str(pd.__version__).split('.')[1]) < 18: # pandas versioning
        df = df.resample(str(minutes) + 'Min', closed='left', label='left', base=bse)
    else:
        # modify '60Min' to change the resulting frequency
        df = df.resample(str(minutes) + 'Min', closed='left', label='left', base=bse).first()  
    return df