added asserts, model descriptions, MEM.R done

Author: surabhinath

measures/Kolmogorov complexity/calculate_Kolmogorov_complexity.py

@@ -3,6 +3,10 @@
 import sys
 from pybdm import BDM
 
+"""
+This module implements the Kolmogorov complexity measure. Refer to Section 2.2 (4) for details.
+"""
+
 def calculate_Kolmogorov_complexity(pattern, grid_size):
     """
     This function calculates the approximate Kolmogorov Complexity of a pattern using Block Decomposition Method.
@@ -24,19 +28,22 @@ def calculate_Kolmogorov_complexity(pattern, grid_size):
     # calculate density for some example patterns:
     #    1) Full black pattern
     #    2) White grid with one central black CellType
-    #    3) Checkarboard
+    #    3) checkerboard
     #    4) Random pattern
 
     all_black = np.ones(grid_size * grid_size, dtype=int)
+    assert calculate_Kolmogorov_complexity(all_black, grid_size) == 25.176631293734488
     print(calculate_Kolmogorov_complexity(all_black, grid_size))
 
     grid_with_one_centre = np.zeros(grid_size * grid_size, dtype=int)
     grid_with_one_centre[(grid_size * grid_size) //2] = 1
+    assert calculate_Kolmogorov_complexity(grid_with_one_centre, grid_size) == 48.354642249885316
     print(calculate_Kolmogorov_complexity(grid_with_one_centre, grid_size))
 
-    checkarboard = np.zeros(grid_size * grid_size, dtype=int)
-    checkarboard[1::2] = 1
-    print(calculate_Kolmogorov_complexity(checkarboard, grid_size))
+    checkerboard = np.zeros(grid_size * grid_size, dtype=int)
+    checkerboard[1::2] = 1
+    assert calculate_Kolmogorov_complexity(checkerboard, grid_size) == 33.43569202047461
+    print(calculate_Kolmogorov_complexity(checkerboard, grid_size))
 
     random = np.random.choice([0, 1], size=(grid_size, grid_size))
     print(calculate_Kolmogorov_complexity(random, grid_size))

measures/asymmetry/calculate_asymmetry.py

@@ -1,6 +1,10 @@
 author = 'Surabhi S Nath'
 import numpy as np
 
+'''
+This module implememts the asymmetry measure. Refer to Section 2.2 (6) and AII for details.
+'''
+
 def get_coords(grid):
 	"""
 	This function returns the indices and count of black pixels in the pattern
@@ -67,19 +71,22 @@ def calculate_asymmetry(grid, grid_size):
     # calculate density for some example patterns:
     #    1) Full black pattern
     #    2) White grid with one central black CellType
-    #    3) Checkarboard
+    #    3) checkerboard
     #    4) Random pattern
-    
+   
     all_black = np.ones(grid_size * grid_size)
+    assert calculate_asymmetry(all_black, grid_size) == (0.0, 0.0)
     print(calculate_asymmetry(all_black, grid_size))
 
     grid_with_one_centre = np.zeros(grid_size * grid_size)
     grid_with_one_centre[(grid_size * grid_size) //2] = 1
+    assert calculate_asymmetry(grid_with_one_centre, grid_size) == (0.0, 0.0)
     print(calculate_asymmetry(grid_with_one_centre, grid_size))
 
-    checkarboard = np.zeros(grid_size * grid_size)
-    checkarboard[1::2] = 1
-    print(calculate_asymmetry(checkarboard, grid_size))
+    checkerboard = np.zeros(grid_size * grid_size)
+    checkerboard[1::2] = 1
+    assert calculate_asymmetry(checkerboard, grid_size) == (0.0, 0.0)
+    print(calculate_asymmetry(checkerboard, grid_size))
 
     random = np.random.choice([0, 1], size=(grid_size, grid_size))
     print(calculate_asymmetry(random, grid_size))

measures/density/calculate_density.py

@@ -1,6 +1,10 @@
 author = 'Surabhi S Nath'
 import numpy as np
 
+"""
+This module implements the density measure. Refer to Section 2.2 (1) and AII for details.
+"""
+
 def calculate_density(grid, grid_size):
     """
     This function caclulates the density of black pixels in a pattern
@@ -19,19 +23,22 @@ def calculate_density(grid, grid_size):
     # calculate density for some example patterns:
     #    1) Full black pattern
     #    2) White grid with one central black CellType
-    #    3) Checkarboard
+    #    3) checkerboard
     #    4) Random pattern
 
     all_black = np.ones(grid_size * grid_size)
+    assert calculate_density(all_black, grid_size) == 1.0
     print(calculate_density(all_black, grid_size))
 
     grid_with_one_centre = np.zeros(grid_size * grid_size)
     grid_with_one_centre[(grid_size * grid_size) //2] = 1
+    assert calculate_density(grid_with_one_centre, grid_size) == 0.0044444444444444444
     print(calculate_density(grid_with_one_centre, grid_size))
 
-    checkarboard = np.zeros(grid_size * grid_size)
-    checkarboard[1::2] = 1
-    print(calculate_density(checkarboard, grid_size))
+    checkerboard = np.zeros(grid_size * grid_size)
+    checkerboard[1::2] = 1
+    assert calculate_density(checkerboard, grid_size) == 0.49777777777777776
+    print(calculate_density(checkerboard, grid_size))
 
     random = np.random.choice([0, 1], size=(grid_size, grid_size))
     print(calculate_density(random, grid_size))

measures/entropy/calculate_entropy.py

@@ -2,6 +2,10 @@
 import numpy as np
 import math
 
+"""
+This module implements the entropy measure. Refer to Section 2.2 (2i) and AII for details.
+"""
+
 def calculate_entropy(gridorflatgrid, grid_size):
     """
     This function calculates the singlescale entropy of the pattern
@@ -33,19 +37,22 @@ def calculate_entropy(gridorflatgrid, grid_size):
     # calculate density for some example patterns:
     #    1) Full black pattern
     #    2) White grid with one central black CellType
-    #    3) Checkarboard
+    #    3) checkerboard
     #    4) Random pattern
 
     all_black = np.ones(grid_size * grid_size, dtype=int)
+    assert calculate_entropy(all_black, grid_size) == 0
     print(calculate_entropy(all_black, grid_size))
 
     grid_with_one_centre = np.zeros(grid_size * grid_size, dtype=int)
     grid_with_one_centre[(grid_size * grid_size) //2] = 1
+    assert calculate_entropy(grid_with_one_centre, grid_size) == 0.041125624368577904
     print(calculate_entropy(grid_with_one_centre, grid_size))
 
-    checkarboard = np.zeros(grid_size * grid_size, dtype=int)
-    checkarboard[1::2] = 1
-    print(calculate_entropy(checkarboard, grid_size))
+    checkerboard = np.zeros(grid_size * grid_size, dtype=int)
+    checkerboard[1::2] = 1
+    assert calculate_entropy(checkerboard, grid_size) == 0.99998575111318
+    print(calculate_entropy(checkerboard, grid_size))
 
     random = np.random.choice([0, 1], size=(grid_size, grid_size))
     print(calculate_entropy(random, grid_size))

measures/entropy_of_means/calculate_entropy_of_means.py

@@ -1,11 +1,14 @@
 """
+This module implements the entropy of means measure based on:
+https://www.geeksforgeeks.org/program-to-count-number-of-connected-components-in-an-undirected-graph/
 Python implementation of the matrix information measurement examples from the
 StackExchange answer written by WilliamAHuber for
 "Measuring entropy/ information/ patterns of a 2d binary matrix"
 http://stats.stackexchange.com/a/17556/43909
-
 Copyright 2014 Cosmo Harrigan
 This program is free software, distributed under the terms of the GNU LGPL v3.0
+
+Refer to Section 2.2 (2iii) and AII for details.
 """
 
 author = 'Cosmo Harrigan'
@@ -35,7 +38,6 @@ def calculate_entropy_of_means(flatgrid, grid_size):
         matrices.append(output_matrix)
 
     prof = profile(matrices)
-    print(prof)
     return np.mean(prof)
 
 if __name__ == "__main__":
@@ -44,19 +46,22 @@ def calculate_entropy_of_means(flatgrid, grid_size):
     # calculate density for some example patterns:
     #    1) Full black pattern
     #    2) White grid with one central black CellType
-    #    3) Checkarboard
+    #    3) checkerboard
     #    4) Random pattern
 
     all_black = np.ones(grid_size * grid_size, dtype=int)
+    assert calculate_entropy_of_means(all_black, grid_size) == 0.0
     print(calculate_entropy_of_means(all_black, grid_size))
 
     grid_with_one_centre = np.zeros(grid_size * grid_size, dtype=int)
     grid_with_one_centre[(grid_size * grid_size) //2] = 1
+    assert calculate_entropy_of_means(grid_with_one_centre, grid_size) == 0.25
     print(calculate_entropy_of_means(grid_with_one_centre, grid_size))
 
-    checkarboard = np.zeros(grid_size * grid_size, dtype=int)
-    checkarboard[1::2] = 1
-    print(calculate_entropy_of_means(checkarboard, grid_size))
+    checkerboard = np.zeros(grid_size * grid_size, dtype=int)
+    checkerboard[1::2] = 1
+    assert calculate_entropy_of_means(checkerboard, grid_size) == 0.5
+    print(calculate_entropy_of_means(checkerboard, grid_size))
 
     random = np.random.choice([0, 1], size=(grid_size, grid_size))
     print(calculate_entropy_of_means(random, grid_size))

measures/entropy_of_means/calculate_profile.py

@@ -53,6 +53,4 @@ def profile(matrices):
 
     @input matrices The set of scaled filtered matrices
     """
-    print("hi")
-    print(matrices[1])
     return [matrix_entropy(scale) for scale in matrices]

measures/intricacy/calculate_intricacy.py

@@ -2,7 +2,10 @@
 import numpy as np
 import sys
 
-# Code based on GeeksforGeeks: https://www.geeksforgeeks.org/program-to-count-number-of-connected-components-in-an-undirected-graph/
+"""
+This module implements the intricacy measure using https://www.geeksforgeeks.org/program-to-count-number-of-connected-components-in-an-undirected-graph/.
+Refer to Section 2.2 (5) and AII for details.
+"""
 
 class Graph:
     """
@@ -126,19 +129,22 @@ def calculate_intricacy(grid, grid_size):
     # calculate density for some example patterns:
     #    1) Full black pattern
     #    2) White grid with one central black CellType
-    #    3) Checkarboard
+    #    3) checkerboard
     #    4) Random pattern
 
     all_black = np.ones(grid_size * grid_size)
+    assert calculate_intricacy(all_black, grid_size) == (1, 1)
     print(calculate_intricacy(all_black, grid_size))
 
     grid_with_one_centre = np.zeros(grid_size * grid_size)
     grid_with_one_centre[(grid_size * grid_size) //2] = 1
+    assert calculate_intricacy(grid_with_one_centre, grid_size) == (2, 2)
     print(calculate_intricacy(grid_with_one_centre, grid_size))
 
-    checkarboard = np.zeros(grid_size * grid_size)
-    checkarboard[1::2] = 1
-    print(calculate_intricacy(checkarboard, grid_size))
+    checkerboard = np.zeros(grid_size * grid_size)
+    checkerboard[1::2] = 1
+    assert calculate_intricacy(checkerboard, grid_size) == (225, 2)
+    print(calculate_intricacy(checkerboard, grid_size))
 
     random = np.random.choice([0, 1], size=(grid_size, grid_size))
     print(calculate_intricacy(random, grid_size))

measures/local spatial complexity/calculate_local_spatial_complexity.py

@@ -1,6 +1,11 @@
 author = 'Surabhi S Nath'
 import numpy as np
 
+"""
+This module implements the local spatial complexity measure based on Javaheri Javid, M. A. (2019). Aesthetic Automata: Synthesis and Simulation of Aesthetic Behaviour in Cellular Automata (Doctoral dissertation, Goldsmiths, University of London).
+Refer to Section 2.2 (3) and AII for details.
+"""
+
 def get_tuples(col1, col2, dirn, grid, grid_size):
 	"""
 	This function calculates the numerator of the conditional and joint distributions and the denominator of the conditional distribution
@@ -132,22 +137,22 @@ def calculate_local_spatial_complexity(pattern, grid_size):
 	# calculate density for some example patterns:
 	#    1) Full black pattern
 	#    2) White grid with one central black CellType
-	#    3) Checkarboard
+	#    3) checkerboard
 	#    4) Random pattern
 
 	all_black = np.ones(grid_size * grid_size, dtype=int)
+	assert calculate_local_spatial_complexity(all_black, grid_size) == (0.0, 0.0)
 	print(calculate_local_spatial_complexity(all_black, grid_size))
 
 	grid_with_one_centre = np.zeros(grid_size * grid_size, dtype=int)
 	grid_with_one_centre[(grid_size * grid_size) //2] = 1
+	assert calculate_local_spatial_complexity(grid_with_one_centre, grid_size) == (0.04331646911530629, 0.0)
 	print(calculate_local_spatial_complexity(grid_with_one_centre, grid_size))
 
-	checkarboard = np.zeros(grid_size * grid_size, dtype=int)
-	checkarboard[1::2] = 1
-	print(calculate_local_spatial_complexity(checkarboard, grid_size))
+	checkerboard = np.zeros(grid_size * grid_size, dtype=int)
+	checkerboard[1::2] = 1
+	assert calculate_local_spatial_complexity(checkerboard, grid_size) == (0.0, 0.0)
+	print(calculate_local_spatial_complexity(checkerboard, grid_size))
 
 	random = np.random.choice([0, 1], size=(grid_size, grid_size))
-	print(calculate_local_spatial_complexity(random, grid_size))
-
-# Ref:
-# Javaheri Javid, M. A. (2019). Aesthetic Automata: Synthesis and Simulation of Aesthetic Behaviour in Cellular Automata (Doctoral dissertation, Goldsmiths, University of London).
+	print(calculate_local_spatial_complexity(random, grid_size))

measures/mean_entropy/calculate_mean_entropy.py

@@ -5,6 +5,10 @@
 sys.path.insert(1, '../entropy/')
 from calculate_entropy import *
 
+"""
+This module implements the mean entropy measure. Refer to Section 2.2 (2ii) and AII for details.
+"""
+
 def calculate_mean_entropy(gridorflatgrid, grid_size):
     """
     This function calculates the mean entropy of a pattern across all scales
@@ -36,19 +40,22 @@ def calculate_mean_entropy(gridorflatgrid, grid_size):
     # calculate density for some example patterns:
     #    1) Full black pattern
     #    2) White grid with one central black CellType
-    #    3) Checkarboard
+    #    3) checkerboard
     #    4) Random pattern
 
     all_black = np.ones(grid_size * grid_size, dtype=int)
+    assert calculate_mean_entropy(all_black, grid_size) == 0.0
     print(calculate_mean_entropy(all_black, grid_size))
 
     grid_with_one_centre = np.zeros(grid_size * grid_size, dtype=int)
     grid_with_one_centre[(grid_size * grid_size) //2] = 1
+    assert calculate_mean_entropy(grid_with_one_centre, grid_size) == 0.0563395650797406
     print(calculate_mean_entropy(grid_with_one_centre, grid_size))
 
-    checkarboard = np.zeros(grid_size * grid_size, dtype=int)
-    checkarboard[1::2] = 1
-    print(calculate_mean_entropy(checkarboard, grid_size))
+    checkerboard = np.zeros(grid_size * grid_size, dtype=int)
+    checkerboard[1::2] = 1
+    assert calculate_mean_entropy(checkerboard, grid_size) == 0.9326281610774253
+    print(calculate_mean_entropy(checkerboard, grid_size))
 
     random = np.random.choice([0, 1], size=(grid_size, grid_size))
     print(calculate_mean_entropy(random, grid_size))

measures/quadtree/calculate_quadtree.py

@@ -1,6 +1,10 @@
 author = 'Surabhi S Nath'
 import numpy as np
 
+"""
+This module implements the quadtree measure. Refer to AII for details.
+"""
+
 def all_same(grid):
     """
     Checks if all pixels of the subgrid are the same state
@@ -60,19 +64,22 @@ def calculate_quadtree(grid, grid_size):
     # calculate density for some example patterns:
     #    1) Full black pattern
     #    2) White grid with one central black CellType
-    #    3) Checkarboard
+    #    3) checkerboard
     #    4) Random pattern
 
     all_black = np.ones(grid_size * grid_size, dtype=int)
+    assert calculate_quadtree(all_black, grid_size) == 0 
     print(calculate_quadtree(all_black, grid_size))
 
     grid_with_one_centre = np.zeros(grid_size * grid_size, dtype=int)
     grid_with_one_centre[(grid_size * grid_size) //2] = 1
+    assert calculate_quadtree(grid_with_one_centre, grid_size) == 4 
     print(calculate_quadtree(grid_with_one_centre, grid_size))
 
-    checkarboard = np.zeros(grid_size * grid_size, dtype=int)
-    checkarboard[1::2] = 1
-    print(calculate_quadtree(checkarboard, grid_size))
+    checkerboard = np.zeros(grid_size * grid_size, dtype=int)
+    checkerboard[1::2] = 1
+    assert calculate_quadtree(checkerboard, grid_size) == 84
+    print(calculate_quadtree(checkerboard, grid_size))
 
     random = np.random.choice([0, 1], size=(grid_size, grid_size))
     print(calculate_quadtree(random, grid_size))