Home

DESCRIPTION

This module is designed to apply a user defined Python function (UDF) to aggregate a time series into a single output raster map.

The user defined function must be named udf_time_series_to_raster_map to be used in this module. The argument of this function is a dictionary that contains the name of the STRDS, the start- and end-time arrays of type pandas.DatetimeIndex and the three dimensional numpy.ndarray. The dictionary is created using the following function:

def create_udf_ts_tile_object(identifier, cell_array, start_time, end_time=None):
"""Create a time series object for a user defined function
:param identifier: The identifier of the time series, in GRASS GIS is would be
                   the STRDS name
:param cell_array: A three dimensional numpy ndarray. For each time stamp a two dimensional
                   (x,y) slice of cell values is provided.
                   - First (0) dimension is time, time stamps are located in separate arrays
                   - Second (1) dimension is y
                   - Third (2) dimension is x
:param start_time: A pandas.DatetimeIndex object that includes the start-time-stamps
                   of the first dimension of the cell array
:param end_time: A pandas.DatetimeIndex object that includes the end-time-stamps
                   of the first dimension of the cell array

The provided data will be put into a dictionary that has the following layout:

.. code: Python
{
    "identifier":identifier,
    "cell_array":cell_array,
    "start_time":start_time,
    "end_time":end_time
}

:return: A dictionary that contains the time series tile
"""

data_object = {}
data_object["identifier"] = identifier
data_object["cell_array"] = cell_array
data_object["start_time"] = start_time
data_object["end_time"] = end_time

return data_object

    
      
      
    
  

The user defined function must return a numpy.ndarray with the same size as a single (x,y) slice in the input numpy.ndarray. This array will be stored in the output raster map layer.

EXAMPLES

Compute the sum of all (x,y) slices in the time series cube along the time axis (0):

import numpy as np
def udf_time_series_to_raster_map(t):
    return np.sum(t["cell_array"], axis=0)

Compute the sum of all slices in the time series cube along the time axis (0) and divide the sum by the number of days in which the time series is valid:

import numpy as np
def udf_time_series_to_raster_map(t):
    if t["end_time"] is not None:
        td = t["end_time"][-1] - t["start_time"][0]
    else:
        td = t["start_time"][-1] - t["start_time"][0]
return np.sum(t["cell_array"], axis=0)/td.days

    
      
      
    
  

AUTHOR

Sören Gebbert