skika.data.wind_sim_generator#

Functions

AQI_map(val)

Classes

PollutionEmission(x, y, r, initial_strength, ...)

Helper class to model emitted pollution.

PollutionSource(x, y, strength, expansion, ...)

Helper class to simulate a source of pollution.

WindSimGenerator([concept, produce_image, ...])

Generator simulating wood smoke pollution sensors.
class skika.data.wind_sim_generator.PollutionEmission(x, y, r, initial_strength, expansion)#

Helper class to model emitted pollution.

Parameters

  • x (int) – X Position of source

  • y (int) – Y Position of source

  • r (int) – Initial radius

  • initial_strength (int) – Amount of pollution emitted initially

  • expansion (int) – How quickly pollution expands

class skika.data.wind_sim_generator.PollutionSource(x, y, strength, expansion, random_state)#

Helper class to simulate a source of pollution.

Parameters

  • x (int) – X Position of source

  • y (int) – Y Position of source

  • strength (int) – Amount of pollution emitted

  • expansion (int) – How quickly pollution expands

  • random_state (int or randomState) – Seed for randomisation

class skika.data.wind_sim_generator.WindSimGenerator(concept=0, produce_image=False, num_sensors=8, sensor_pattern='circle', sample_random_state=None, x_trail=1)#

Generator simulating wood smoke pollution sensors.

This generator is modeled after experiments placing sensors over a town to capture wood pollution readings.

Pollution sources are distributed and generate an emission every few timesteps. Emission movement depends on wind speed and direction, as well as internal dissipation.

Feature sensors pick up readings to generate features, while a central target sensor records a target reading. This is transformed into a classification target.

Parameters

  • concept (int (Default: 0)) – The initial concept

  • produce_image (bool (Default: False)) – Whether or not to produce an image for visualisation

  • num_sensors (int (Default: 8)) – The number of sensors to place

  • sensor_pattern ('circle', 'grid' (Default: circle)) – The pattern to place sensors

  • sample_random_state (int, randomState) – randomState to generate simulation

  • x_trail (int (Default: 1)) – The number of feature x lags to output

Examples

>>> # Imports
>>> from skika.data.synthetic.wind_sim_generator import WindSimGenerator
>>> # Setting up the stream
>>> stream = WindSimGenerator(num_sensors = 16, sensor_pattern = 'grid')
>>> stream.prepare_for_use()
>>> # Retrieving one sample.
>>> # First sample may be all 0 depending on pollution transmission.
>>> stream.next_sample()
(array([[  0.        , 134.78461727,  71.68383298,   0.        ,
      0.        , 104.90590932, 266.91631925,   0.        ,
      0.        ,  51.31476312,  56.51469396,   0.        ,
      0.        ,   0.        ,  11.0821078 ,   0.        ,
      0.        , 153.24292292,  75.4566663 ,   0.        ,
      0.        , 110.42727297, 296.14962106,   0.        ,
      0.        ,  59.22154118,  68.89489498,   0.        ,
      0.        ,   0.        ,  20.02603747,   0.        ,
    199.49156743]]), array([4]))
>>> # Simulate for some time steps.
>>> for i in range(100):
>>>     _ = stream.next_sample()
>>> stream.next_sample()
(array([[  0.        , 119.28319232,  73.21440717,   0.        ,
      0.        , 105.7541374 , 254.93216371,   0.        ,
      0.        ,  51.13375057,  54.91003902,   0.        ,
      0.        ,   0.        ,  11.45964019,   0.        ,
      0.        , 136.27998587,  86.155855  ,   0.        ,
      0.        , 120.36105135, 297.60813426,   0.        ,
      0.        ,  68.54515689,  76.13336341,   0.        ,
      0.        ,   0.        ,  20.32270414,   0.        ,
    190.38410787]]), array([4]))
>>> stream.n_remaining_samples()
-1
>>> stream.has_more_samples()
True
>>> # The concept can be changed at any time.
>>> # This will change wind speed and direction
>>> # as well as the location of pollution sources.
>>> stream.set_concept(6)
>>> stream.next_sample()
(array([[  0.        ,   0.        ,   6.94576111,   0.        ,
        96.21104249,  78.56179139,  68.90448328,   0.        ,
        184.91552406, 119.25640506,  71.48302302,  24.73990969,
        0.        ,   0.        ,   0.        ,   0.        ,
        0.        ,   0.        ,   0.        ,   0.        ,
        101.27478157,  82.69662251,  72.53103503,   0.        ,
        215.61056166, 131.17070743,  78.0006768 ,  21.47142214,
        0.        ,   0.        ,   0.        ,   0.        ,
        227.11417744]]), array([4]))
get_direction_from_concept(concept)#

Get the wind direction from a set concept.

set_concept(concept_seed, difficulty=3)#
Set windspeed, direction and sources.

Randomized per concept, but given the same concept_seed will set to the same values.

Parameters

  • concept_seed (int) – Seed for the concept

  • difficulty (int (Default: 3)) – How difficult a concept is

set_wind(concept=0, direc=None, strength=None)#

Set wind parameters

Parameters

  • concept (int) – The ID of the concept being set

  • direc (int) – The direction of wind in degrees

  • strength (int) – The strength of the wind in meters a second.

update()#

Update simulation

To save time when checking if sensor is within an emission, we exclude check for emissions wholely contained in anouther emission.