Powerlawsv2 (#23)

* fixed unit tests for clusterd graphs

* fixed bug

* format with black

---------

Co-authored-by: Nicholas Landry <nicholas.landry.91@gmail.com>

tests/test_generative.py

@@ -12,3 +12,40 @@ def test_erdos_renyi():
 
     A = erdos_renyi(10, 0.3, seed=0)
     assert A.sum() == 16
+
+
+def test_unipartite_clustering():
+    n = 21
+    clique_numbers = np.arange(1, 20)
+    clique_number = 5
+    clique_size = n // clique_number  # the number of nodes per clique
+    clique_membership = 1  # the number of cliques per node
+    my_p_dist = delta_dist(clique_size)
+    my_g_dist = delta_dist(clique_membership)
+
+    k1 = [clique_membership] * n
+    k2 = [clique_size] * clique_number
+
+    # A = clustered_unipartite(clique_number, n, my_p_dist, my_g_dist)
+    A = clustered_network(k1, k2)
+
+    G = nx.from_numpy_array(A)
+    # Calculate the Laplacian matrix
+    L = nx.laplacian_matrix(G).toarray()
+    # Calculate the eigenvalues of the Laplacian matrix
+    eigenvalues = np.linalg.eigvals(L)
+    num_zeros = np.count_nonzero(
+        np.isclose(eigenvalues, 0)
+    )  # this should be equal to the clique number
+    assert num_zeros == clique_number
+
+
+def test_truncated_power_law_configuration_model():
+    n = 500
+    x_min = 1
+    x_max = n
+    r = 2
+    A = truncated_power_law_configuration(n, x_min, x_max, r)
+    G = nx.from_numpy_array(A)
+    max_degree = max(dict(G.degree()).values())
+    assert max_degree <= x_max