import networkx as nx
from networkx.algorithms.approximation import average_clustering
# This approximation has to be be exact in regular graphs
# with no triangles or with all possible triangles.
def test_petersen():
# Actual coefficient is 0
G = nx.petersen_graph()
assert average_clustering(G, trials=int(len(G) / 2)) == nx.average_clustering(G)
def test_petersen_seed():
# Actual coefficient is 0
G = nx.petersen_graph()
assert average_clustering(
G, trials=int(len(G) / 2), seed=1
) == nx.average_clustering(G)
def test_tetrahedral():
# Actual coefficient is 1
G = nx.tetrahedral_graph()
assert average_clustering(G, trials=int(len(G) / 2)) == nx.average_clustering(G)
def test_dodecahedral():
# Actual coefficient is 0
G = nx.dodecahedral_graph()
assert average_clustering(G, trials=int(len(G) / 2)) == nx.average_clustering(G)
def test_empty():
G = nx.empty_graph(5)
assert average_clustering(G, trials=int(len(G) / 2)) == 0
def test_complete():
G = nx.complete_graph(5)
assert average_clustering(G, trials=int(len(G) / 2)) == 1
G = nx.complete_graph(7)
assert average_clustering(G, trials=int(len(G) / 2)) == 1