def test_load_modern_graph(modern_graph_data_dir):
image, ci = get_gs_image_on_ci_env()
if ci:
sess = graphscope.session(
show_log=True,
num_workers=1,
k8s_gs_image=image,
)
else:
sess = graphscope.session(
show_log=True,
num_workers=1,
)
graph = load_modern_graph(sess, modern_graph_data_dir)
interactive = sess.gremlin(graph)
queries = [
"g.V().has('name','marko').count()",
"g.V().has('person','name','marko').count()",
"g.V().has('person','name','marko').outE('created').count()",
"g.V().has('person','name','marko').outE('created').inV().count()",
"g.V().has('person','name','marko').out('created').count()",
"g.V().has('person','name','marko').out('created').values('name').count()",
]
for q in queries:
result = interactive.execute(q).all().result()[0]
assert result == 1
def test_demo_distribute(data_dir, modern_graph_data_dir):
image, ci = get_gs_image_on_ci_env()
if ci:
sess = graphscope.session(
show_log=True,
num_workers=2,
k8s_gs_image=image,
)
else:
sess = graphscope.session(
show_log=True,
num_workers=2,
)
graph = load_ldbc(sess, data_dir)
# Interactive engine
interactive = sess.gremlin(graph)
sub_graph = interactive.subgraph( # noqa: F841
'g.V().hasLabel("person").outE("knows")')
person_count = (interactive.execute(
'g.V().hasLabel("person").outE("knows").bothV().dedup().count()').all(
).result()[0])
knows_count = (interactive.execute(
'g.V().hasLabel("person").outE("knows").count()').all().result()[0])
interactive2 = sess.gremlin(sub_graph)
sub_person_count = interactive2.execute("g.V().count()").all().result()[0]
sub_knows_count = interactive2.execute("g.E().count()").all().result()[0]
assert person_count == sub_person_count
assert knows_count == sub_knows_count
# Analytical engine
# project the projected graph to simple graph.
simple_g = sub_graph.project_to_simple(v_label="person", e_label="knows")
pr_result = graphscope.pagerank(simple_g, delta=0.8)
tc_result = graphscope.triangles(simple_g)
# add the PageRank and triangle-counting results as new columns to the property graph
# FIXME: Add column to sub_graph
sub_graph.add_column(pr_result, {"Ranking": "r"})
sub_graph.add_column(tc_result, {"TC": "r"})
# test subgraph on modern graph
mgraph = load_modern_graph(sess, modern_graph_data_dir)
# Interactive engine
minteractive = sess.gremlin(mgraph)
msub_graph = minteractive.subgraph( # noqa: F841
'g.V().hasLabel("person").outE("knows")')
person_count = (minteractive.execute(
'g.V().hasLabel("person").outE("knows").bothV().dedup().count()').all(
).result()[0])
msub_interactive = sess.gremlin(msub_graph)
sub_person_count = msub_interactive.execute(
"g.V().count()").all().result()[0]
assert person_count == sub_person_count
# GNN engine
sess.close()
def test_query_modern_graph(modern_graph_data_dir):
gs_image, gie_manager_image = get_gs_image_on_ci_env()
sess = graphscope.session(
num_workers=1,
k8s_gs_image=gs_image,
k8s_gie_graph_manager_image=gie_manager_image,
k8s_coordinator_cpu=0.5,
k8s_coordinator_mem="2500Mi",
k8s_vineyard_cpu=0.1,
k8s_vineyard_mem="512Mi",
k8s_engine_cpu=0.1,
k8s_engine_mem="1500Mi",
k8s_vineyard_shared_mem="2Gi",
k8s_volumes=get_k8s_volumes(),
)
graph = load_modern_graph(sess, modern_graph_data_dir)
interactive = sess.gremlin(graph)
queries = [
"g.V().has('name','marko').count()",
"g.V().has('person','name','marko').count()",
"g.V().has('person','name','marko').outE('created').count()",
"g.V().has('person','name','marko').outE('created').inV().count()",
"g.V().has('person','name','marko').out('created').count()",
"g.V().has('person','name','marko').out('created').values('name').count()",
]
for q in queries:
result = interactive.execute(q).all().result()[0]
assert result == 1
def test_traversal_modern_graph(modern_graph_data_dir):
from gremlin_python.process.traversal import Order
from gremlin_python.process.traversal import P
gs_image, gie_manager_image = get_gs_image_on_ci_env()
sess = graphscope.session(
show_log=True,
num_workers=1,
k8s_gs_image=gs_image,
k8s_gie_graph_manager_image=gie_manager_image,
k8s_coordinator_cpu=0.5,
k8s_coordinator_mem="2500Mi",
k8s_vineyard_cpu=0.1,
k8s_vineyard_mem="512Mi",
k8s_engine_cpu=0.1,
k8s_engine_mem="1500Mi",
k8s_vineyard_shared_mem="2Gi",
k8s_volumes=get_k8s_volumes(),
)
graph = load_modern_graph(sess, modern_graph_data_dir)
interactive = sess.gremlin(graph)
g = interactive.traversal_source()
assert g.V().has("name", "marko").count().toList()[0] == 1
assert g.V().has("person", "name", "marko").count().toList()[0] == 1
assert g.V().has("person", "name",
"marko").outE("created").count().toList()[0] == 1
assert (g.V().has("person", "name",
"marko").outE("created").inV().count().toList()[0] == 1)
assert g.V().has("person", "name",
"marko").out("created").count().toList()[0] == 1
assert (g.V().has(
"person", "name",
"marko").out("created").values("name").count().toList()[0] == 1)
assert (g.V().hasLabel("person").has("age", P.gt(30)).order().by(
"age", Order.desc).count().toList()[0] == 2)
def test_add_vertices_edges(gs_session_distributed, modern_graph_data_dir):
graph = load_modern_graph(gs_session_distributed, modern_graph_data_dir)
graph = graph.add_vertices(
Loader(os.path.join(modern_graph_data_dir, "person.csv"),
delimiter="|"),
"person2",
["name", ("age", "int")],
"id",
)
assert "person2" in graph.schema.vertex_labels
graph = graph.add_edges(
Loader(
os.path.join(modern_graph_data_dir, "knows.csv"),
delimiter="|",
),
"knows2",
["weight"],
src_label="person2",
dst_label="person2",
)
assert "knows2" in graph.schema.edge_labels
interactive = gs_session_distributed.gremlin(graph)
g = interactive.traversal_source()
assert g.V().count().toList()[0] == 10
assert g.E().count().toList()[0] == 8
def test_query_modern_graph(gs_session, modern_graph_data_dir):
graph = load_modern_graph(gs_session, modern_graph_data_dir)
interactive = gs_session.gremlin(graph)
queries = [
"g.V().has('name','marko').count()",
"g.V().has('person','name','marko').count()",
"g.V().has('person','name','marko').outE('created').count()",
"g.V().has('person','name','marko').outE('created').inV().count()",
"g.V().has('person','name','marko').out('created').count()",
"g.V().has('person','name','marko').out('created').values('name').count()",
]
for q in queries:
result = interactive.execute(q).all().result()[0]
assert result == 1
def test_demo_distribute(gs_session_distributed, data_dir,
modern_graph_data_dir):
graph = load_ldbc(gs_session_distributed, data_dir)
# Interactive engine
interactive = gs_session_distributed.gremlin(graph)
sub_graph = interactive.subgraph( # noqa: F841
'g.V().hasLabel("person").outE("knows")')
person_count = (interactive.execute(
'g.V().hasLabel("person").outE("knows").bothV().dedup().count()').all(
).result()[0])
knows_count = (interactive.execute(
'g.V().hasLabel("person").outE("knows").count()').all().result()[0])
interactive2 = gs_session_distributed.gremlin(sub_graph)
sub_person_count = interactive2.execute("g.V().count()").all().result()[0]
sub_knows_count = interactive2.execute("g.E().count()").all().result()[0]
assert person_count == sub_person_count
assert knows_count == sub_knows_count
# Analytical engine
# project the projected graph to simple graph.
simple_g = sub_graph.project_to_simple(v_label="person", e_label="knows")
pr_result = graphscope.pagerank(simple_g, delta=0.8)
tc_result = graphscope.triangles(simple_g)
# add the PageRank and triangle-counting results as new columns to the property graph
# FIXME: Add column to sub_graph
sub_graph.add_column(pr_result, {"Ranking": "r"})
sub_graph.add_column(tc_result, {"TC": "r"})
# test subgraph on modern graph
mgraph = load_modern_graph(gs_session_distributed, modern_graph_data_dir)
# Interactive engine
minteractive = gs_session_distributed.gremlin(mgraph)
msub_graph = minteractive.subgraph( # noqa: F841
'g.V().hasLabel("person").outE("knows")')
person_count = (minteractive.execute(
'g.V().hasLabel("person").outE("knows").bothV().dedup().count()').all(
).result()[0])
msub_interactive = gs_session_distributed.gremlin(msub_graph)
sub_person_count = msub_interactive.execute(
"g.V().count()").all().result()[0]
assert person_count == sub_person_count
def test_traversal_modern_graph(gs_session, modern_graph_data_dir):
from gremlin_python.process.traversal import Order
from gremlin_python.process.traversal import P
gs_image, gie_manager_image = get_gs_image_on_ci_env()
graph = load_modern_graph(gs_session, modern_graph_data_dir)
interactive = gs_session.gremlin(graph)
g = interactive.traversal_source()
assert g.V().has("name", "marko").count().toList()[0] == 1
assert g.V().has("person", "name", "marko").count().toList()[0] == 1
assert g.V().has("person", "name",
"marko").outE("created").count().toList()[0] == 1
assert (g.V().has("person", "name",
"marko").outE("created").inV().count().toList()[0] == 1)
assert g.V().has("person", "name",
"marko").out("created").count().toList()[0] == 1
assert (g.V().has(
"person", "name",
"marko").out("created").values("name").count().toList()[0] == 1)
assert (g.V().hasLabel("person").has("age", P.gt(30)).order().by(
"age", Order.desc).count().toList()[0] == 2)
def arrow_modern_graph(graphscope_session):
graph = load_modern_graph(graphscope_session,
prefix="{}/modern_graph".format(test_repo_dir))
yield graph
graph.unload()
def test_demo_distribute(data_dir, modern_graph_data_dir):
gs_image, gie_manager_image = get_gs_image_on_ci_env()
sess = graphscope.session(
num_workers=1,
k8s_gs_image=gs_image,
k8s_gie_graph_manager_image=gie_manager_image,
k8s_coordinator_cpu=0.5,
k8s_coordinator_mem="2500Mi",
k8s_vineyard_cpu=0.1,
k8s_vineyard_mem="512Mi",
k8s_engine_cpu=0.1,
k8s_engine_mem="1500Mi",
k8s_etcd_cpu=2,
k8s_vineyard_shared_mem="2Gi",
k8s_volumes=get_k8s_volumes(),
)
graph = load_ldbc(sess, data_dir)
# Interactive engine
interactive = sess.gremlin(graph)
sub_graph = interactive.subgraph( # noqa: F841
'g.V().hasLabel("person").outE("knows")'
)
person_count = (
interactive.execute(
'g.V().hasLabel("person").outE("knows").bothV().dedup().count()'
)
.all()
.result()[0]
)
knows_count = (
interactive.execute('g.V().hasLabel("person").outE("knows").count()')
.all()
.result()[0]
)
interactive2 = sess.gremlin(sub_graph)
sub_person_count = interactive2.execute("g.V().count()").all().result()[0]
sub_knows_count = interactive2.execute("g.E().count()").all().result()[0]
assert person_count == sub_person_count
assert knows_count == sub_knows_count
# Analytical engine
# project the projected graph to simple graph.
simple_g = sub_graph.project_to_simple(v_label="person", e_label="knows")
pr_result = graphscope.pagerank(simple_g, delta=0.8)
tc_result = graphscope.triangles(simple_g)
# add the PageRank and triangle-counting results as new columns to the property graph
# FIXME: Add column to sub_graph
sub_graph.add_column(pr_result, {"Ranking": "r"})
sub_graph.add_column(tc_result, {"TC": "r"})
# test subgraph on modern graph
mgraph = load_modern_graph(sess, modern_graph_data_dir)
# Interactive engine
minteractive = sess.gremlin(mgraph)
msub_graph = minteractive.subgraph( # noqa: F841
'g.V().hasLabel("person").outE("knows")'
)
person_count = (
minteractive.execute(
'g.V().hasLabel("person").outE("knows").bothV().dedup().count()'
)
.all()
.result()[0]
)
msub_interactive = sess.gremlin(msub_graph)
sub_person_count = msub_interactive.execute("g.V().count()").all().result()[0]
assert person_count == sub_person_count
# GNN engine
sess.close()
