def clustering(G, nodes=None, weight=None):
# FIXME(weibin): clustering now only correct in directed graph.
# FIXME: nodes and weight not support.
pg = G.project_to_simple()
ctx = graphscope.clustering(pg)
return ctx.to_dataframe({"node": "v.id", "result": "r"})
def test_run_app_on_directed_graph(
p2p_project_directed_graph,
sssp_result,
pagerank_result,
hits_result,
bfs_result,
clustering_result,
dc_result,
ev_result,
katz_result,
):
# sssp
ctx1 = sssp(p2p_project_directed_graph, src=6)
r1 = (ctx1.to_dataframe({
"node": "v.id",
"r": "r"
}).sort_values(by=["node"]).to_numpy(dtype=float))
r1[r1 == 1.7976931348623157e308] = float(
"inf") # replace limit::max with inf
assert np.allclose(r1, sssp_result["directed"])
ctx2 = sssp(p2p_project_directed_graph, 6)
r2 = (ctx2.to_dataframe({
"node": "v.id",
"r": "r"
}).sort_values(by=["node"]).to_numpy(dtype=float))
r2[r2 == 1.7976931348623157e308] = float(
"inf") # replace limit::max with inf
assert np.allclose(r2, sssp_result["directed"])
assert np.allclose(
ctx2.to_dataframe({
"node": "v.id",
"r": "r"
},
vertex_range={
"begin": 1,
"end": 4
}).sort_values(by=["node"]).to_numpy(),
[[1.0, 260.0], [2.0, 229.0], [3.0, 310.0]],
)
assert np.allclose(
sorted(ctx1.to_numpy("r", vertex_range={
"begin": 1,
"end": 4
})),
sorted([260.0, 229.0, 310.0]),
)
r3 = sssp(p2p_project_directed_graph, 100000000)
assert r3 is not None
# pagerank
ctx_pr = pagerank(p2p_project_directed_graph, delta=0.85, max_round=10)
ret_pr = (ctx_pr.to_dataframe({
"node": "v.id",
"r": "r"
}).sort_values(by=["node"]).to_numpy(dtype=float))
assert np.allclose(ret_pr, pagerank_result["directed"])
# hits
ctx_hits = hits(p2p_project_directed_graph, tolerance=0.001)
ret_hub = (ctx_hits.to_dataframe({
"node": "v.id",
"hub": "r.hub"
}).sort_values(by=["node"]).to_numpy(dtype=float))
ret_auth = (ctx_hits.to_dataframe({
"node": "v.id",
"auth": "r.auth"
}).sort_values(by=["node"]).to_numpy(dtype=float))
assert np.allclose(ret_hub, hits_result["hub"])
assert np.allclose(ret_auth, hits_result["auth"])
# bfs
ctx4 = bfs(p2p_project_directed_graph, src=6)
r4 = (ctx4.to_dataframe({
"node": "v.id",
"r": "r"
}).sort_values(by=["node"]).to_numpy(dtype=int))
assert np.all(r4 == bfs_result["directed"])
ctx5 = bfs(p2p_project_directed_graph, 6)
r5 = (ctx5.to_dataframe({
"node": "v.id",
"r": "r"
}).sort_values(by=["node"]).to_numpy(dtype=int))
assert np.all(r5 == bfs_result["directed"])
assert np.all(
ctx5.to_dataframe(
{
"node": "v.id",
"r": "r"
}, vertex_range={
"begin": 1,
"end": 4
}).sort_values(by=["node"]).to_numpy() == [[1, 5], [2, 5], [3, 6]])
assert np.all(
sorted(ctx5.to_numpy("r", vertex_range={
"begin": 1,
"end": 4
})) == [5, 5, 6])
# simple_path
assert is_simple_path(p2p_project_directed_graph, [1, 10])
with pytest.raises(InvalidArgumentError,
match="Louvain not support directed graph."):
louvain(p2p_project_directed_graph)
# clustering
ctx_clustering = clustering(p2p_project_directed_graph)
ret_clustering = (ctx_clustering.to_dataframe({
"node": "v.id",
"r": "r"
}).sort_values(by=["node"]).to_numpy(dtype=float))
assert np.allclose(ret_clustering, clustering_result["directed"])
# degree_centrality
ctx_dc = degree_centrality(p2p_project_directed_graph)
ret_dc = (ctx_dc.to_dataframe({
"node": "v.id",
"r": "r"
}).sort_values(by=["node"]).to_numpy(dtype=float))
assert np.allclose(ret_dc, dc_result["directed"])
# eigenvector_centrality
ctx_ev = eigenvector_centrality(p2p_project_directed_graph)
ret_ev = (ctx_ev.to_dataframe({
"node": "v.id",
"r": "r"
}).sort_values(by=["node"]).to_numpy(dtype=float))
assert np.allclose(ret_ev, ev_result["directed"])
# katz_centrality
ctx_katz = katz_centrality(p2p_project_directed_graph)
def _clustering(G):
return graphscope.clustering(G)