proximity.py 文件源码

def member_to_badge_proximity(fileobject, time_bins_size='1min', tz='US/Eastern'):
    """Creates a member-to-badge proximity DataFrame from a proximity data file.

    Parameters
    ----------
    fileobject : file or iterable list of str
        The proximity data, as an iterable of JSON strings.

    time_bins_size : str
        The size of the time bins used for resampling.  Defaults to '1min'.

    tz : str
        The time zone used for localization of dates.  Defaults to 'US/Eastern'.

    Returns
    -------
    pd.DataFrame :
        The member-to-badge proximity data.
    """

    def readfile(fileobject):
        for line in fileobject:
            data = json.loads(line)['data']

            for (observed_id, distance) in data['rssi_distances'].items():
                yield (
                    data['timestamp'],
                    str(data['member']),
                    int(observed_id),
                    float(distance['rssi']),
                    float(distance['count']),
                )

    df = pd.DataFrame(
            readfile(fileobject),
            columns=('timestamp', 'member', 'observed_id', 'rssi', 'count')
    )

    # Convert timestamp to datetime for convenience, and localize to UTC
    df['datetime'] = pd.to_datetime(df['timestamp'], unit='s', utc=True) \
            .dt.tz_localize('UTC').dt.tz_convert(tz)
    del df['timestamp']

    # Group per time bins, member and observed_id,
    # and take the first value, arbitrarily
    df = df.groupby([
        pd.TimeGrouper(time_bins_size, key='datetime'),
        'member',
        'observed_id'
    ]).first()

    # Sort the data
    df.sort_index(inplace=True)

    return df