def test_unsupervised_vs_auc():
def loader():
return pg.load_datasets_multiple_communities(["graph9"])
algorithms = pg.create_variations(pg.create_many_filters(), pg.create_many_variation_types())
time_scores = pg.benchmark_scores(pg.benchmark(algorithms, loader(), pg.Time))
assert sum(time_scores) > 0
measures = {"AUC": lambda ground_truth, exlude: pg.MultiSupervised(pg.AUC, ground_truth, exlude),
"NDCG": lambda ground_truth, exlude: pg.MultiSupervised(pg.NDCG, ground_truth, exlude),
"Density": lambda graph: pg.MultiUnsupervised(pg.Density, graph),
"Conductance": lambda graph: pg.MultiUnsupervised(pg.Conductance(autofix=True).as_unsupervised_method(), graph),
"Modularity": lambda graph: pg.MultiUnsupervised(pg.Modularity(max_positive_samples=5).as_unsupervised_method(), graph),
"CCcos": lambda graph: pg.ClusteringCoefficient(graph, similarity="cos", max_positive_samples=5),
"CCdot": lambda graph: pg.ClusteringCoefficient(graph, similarity="dot", max_positive_samples=5),
"LinkAUCcos": lambda graph: pg.LinkAssessment(graph, similarity="cos", max_positive_samples=5),
"LinkAUCdot": lambda graph: pg.LinkAssessment(graph, similarity="dot", max_positive_samples=5),
"HopAUCcos": lambda graph: pg.LinkAssessment(graph, similarity="cos", hops=2, max_positive_samples=5),
"HopAUCdot": lambda graph: pg.LinkAssessment(graph, similarity="dot", hops=2, max_positive_samples=5),
}
scores = {}#measure: pg.benchmark_scores(pg.benchmark(algorithms, loader(), measures[measure])) for measure in measures}
for measure in measures: # do this as a for loop, because pytest becomes a little slow above list comprehension
scores[measure] = pg.benchmark_scores(pg.benchmark(algorithms, loader(), measures[measure]))
supervised = {"AUC", "NDCG"}
evaluations = dict()
for measure in measures:
evaluations[measure] = abs(pg.SpearmanCorrelation(scores["AUC"])(scores[measure]))
#for measure in measures:
# print(measure, evaluations[measure])
assert max([evaluations[measure] for measure in measures if measure not in supervised]) == evaluations["LinkAUCdot"]
def test_correlation_compliance():
graph = next(pg.load_datasets_graph(["graph5"]))
# TODO: Make spearman and pearson correlation support tensorflow
alg1 = pg.PageRank(alpha=0.5)
alg2 = pg.PageRank(alpha=0.99)
pearson_ordinals = pg.PearsonCorrelation(pg.Ordinals(alg1)(graph))(
pg.Ordinals(alg2)(graph))
spearman = pg.SpearmanCorrelation(alg1(graph))(alg2(graph))
assert pearson_ordinals == spearman
def test_venuerank():
graph = next(pg.load_datasets_graph(["bigraph"]))
for _ in supported_backends():
venuerank = pg.PageRank(alpha=0.85,
max_iters=10000,
converge_to_eigenvectors=True,
tol=1.E-12)
venuerank_result = venuerank.rank(graph)
small_restart = pg.PageRank(alpha=0.99, max_iters=10000, tol=1.E-12)
small_restart_result = small_restart.rank(graph)
#assert venuerank.convergence.iteration < small_restart.convergence.iteration / 2
corr = pg.SpearmanCorrelation(pg.Ordinals()(venuerank_result))(
pg.Ordinals()(small_restart_result))
assert corr > 0.99
def test_rank_order_convergence():
graph = next(pg.load_datasets_graph(["graph9"]))
algorithm1 = pg.Ordinals(pg.PageRank(0.85, tol=1.E-12, max_iters=1000))
algorithm2 = pg.Ordinals(
pg.PageRank(0.85, convergence=pg.RankOrderConvergenceManager(0.85)))
algorithm3 = pg.Ordinals(
pg.PageRank(0.85,
convergence=pg.RankOrderConvergenceManager(
0.85, 0.99, "fraction_of_walks")))
for _ in supported_backends():
ranks1 = algorithm1.rank(graph, {"A": 1})
ranks2 = algorithm2.rank(graph, {"A": 1})
ranks3 = algorithm3.rank(graph, {"A": 1})
assert pg.SpearmanCorrelation(ranks1)(ranks2) > 0.95
assert pg.SpearmanCorrelation(ranks1)(ranks3) > 0.95
assert pg.SpearmanCorrelation(ranks3)(ranks2) > 0.95
assert "17 iterations" in str(algorithm3.ranker.convergence)
with pytest.raises(Exception):
algorithm = pg.Ordinals(
pg.PageRank(0.85,
convergence=pg.RankOrderConvergenceManager(
0.85, 0.99, "unknown")))
algorithm.rank(graph, {"A": 1})
import pygrank as pg
loader = list(pg.load_datasets_multiple_communities(["bigraph", "cora", "citeseer"]))
algorithms = pg.create_variations(pg.create_demo_filters(), pg.create_many_variation_types())
algorithms = pg.create_variations(algorithms, pg.Normalize) # add normalization to all algorithms
print("Algorithms", len(algorithms))
measures = {"AUC": lambda ground_truth, exlude: pg.MultiSupervised(pg.AUC, ground_truth, exlude),
"NDCG": lambda ground_truth, exlude: pg.MultiSupervised(pg.NDCG, ground_truth, exlude),
"Density": lambda graph: pg.MultiUnsupervised(pg.Density, graph),
"Modularity": lambda graph: pg.MultiUnsupervised(pg.Modularity, graph),
"LinkCC": lambda graph: pg.ClusteringCoefficient(graph, similarity="dot"),
"LinkAUCcos": lambda graph: pg.LinkAssessment(graph, similarity="cos"),
"HopAUCdot": lambda graph: pg.LinkAssessment(graph, similarity="dot", hops=2),
}
scores = {measure: pg.benchmark_scores(pg.benchmark(algorithms, loader, measures[measure])) for measure in measures}
evaluations_vs_auc = dict()
evaluations_vs_ndcg = dict()
for measure in measures:
evaluations_vs_auc[measure] = abs(pg.SpearmanCorrelation(scores["AUC"])(scores[measure]))
evaluations_vs_ndcg[measure] = abs(pg.SpearmanCorrelation(scores["NDCG"])(scores[measure]))
pg.benchmark_print([("Measure", "AUC corr", "NDCG corr")]
+ [(measure, evaluations_vs_auc[measure], evaluations_vs_ndcg[measure]) for measure in measures])