def test_nodeset_create_composite_index(self, graph, clear_graph):
labels = ['TestNode']
properties = ['some_key', 'other_key']
ns = NodeSet(labels, merge_keys=properties)
ns.create_index(graph)
result = list(graph.run("CALL db.indexes()"))
for row in result:
# the result of the db.indexes() procedure is different for Neo4j 3.5 and 4
# this should also be synced with differences in py2neo versions
if 'tokenNames' in row:
assert row['tokenNames'] == labels and row['properties'] == properties \
or row['tokenNames'] == labels and row['properties'] == properties
elif 'labelsOrTypes' in row:
assert row['labelsOrTypes'] == labels and row['properties'] == properties \
or row['labelsOrTypes'] == labels and row['properties'] == properties
def test_nodeset__estimate_type_of_property_values():
ns = NodeSet(['Test'], merge_keys=['uuid'])
for i in range(10):
ns.add_node({'uuid': i, 'key': 'value', 'foo': 20.5, 'changing': 4})
# change type for one property to make sure that str is set for inconsistent types
ns.add_node({'uuid': 1, 'key': 'value', 'foo': 20.5, 'changing': 30.4})
types = ns._estimate_type_of_property_values()
assert types['uuid'] == int
assert types['key'] == str
assert types['foo'] == float
assert types['changing'] == str
def test_read_from_files(self, root_dir, clear_graph, graph):
files_path = os.path.join(root_dir, "test", "files")
json_file_path = os.path.join(files_path, 'nodes_csv_json.json')
csv_file_path = os.path.join(files_path, 'nodes_csv_json.csv')
assert os.path.exists(json_file_path)
assert os.path.exists(csv_file_path)
ns = NodeSet.from_csv_json_set(csv_file_path, json_file_path)
assert ns.labels == ['Test']
assert ns.merge_keys == ['test_id']
ns.merge(graph)
assert list(graph.run("MATCH (n:Test) RETURN count(n)"))[0][0] == 2
def test_nodeset_recreate_existing_single_index(self, graph, clear_graph):
"""
The output/error when you try to recreate an existing index is different in Neo4j 3.5 and 4.
Create an index a few times to make sure this error is handled.
"""
labels = ['TestNode']
properties = ['some_key']
ns = NodeSet(labels, merge_keys=properties)
ns.create_index(graph)
ns.create_index(graph)
ns.create_index(graph)
def test_nodeset_add_unique():
ns = NodeSet(['Test', 'Foo'], merge_keys=['name'])
for i in range(10):
ns.add_unique({'name': 'Peter'})
assert len(ns.nodes) == 1
def test_str():
ns = NodeSet(['Test', 'Foo'], merge_keys=['uuid'])
assert str(ns) == "<NodeSet (['Test', 'Foo']; ['uuid'])>"
def test_nodeset_merge_preserve_keeps_append_props(self, graph,
clear_graph):
"""
Merge a nodeset 3 times and check number of nodes.
"""
ns = NodeSet(['Test'],
merge_keys=['uuid'],
append_props=['key'],
preserve=['key'])
for i in range(100):
ns.add_node({'uuid': i, 'key': 'value'})
ns.merge(graph)
assert list(
graph.run("MATCH (n:Test) where 'value' IN n.key RETURN count(n)")
)[0][0] == 100
append_ns = NodeSet(['Test'],
merge_keys=['uuid'],
append_props=['key'],
preserve=['key'])
for i in range(100):
append_ns.add_node({'uuid': i, 'key': 'other_value'})
append_ns.merge(graph)
assert list(
graph.run("MATCH (n:Test) where 'value' IN n.key RETURN count(n)")
)[0][0] == 100
assert list(
graph.run(
"MATCH (n:Test) where 'other_value' IN n.key RETURN count(n)")
)[0][0] == 0
def nodeset_multiple_labels_multiple_merge_keys():
ns = NodeSet(['Test', 'Foo', 'Bar'], merge_keys=['uuid', 'other'])
for i in range(1000):
ns.add_node({'uuid': i, 'other': i + 10})
return ns
def dataset(cls) -> NodeSet:
"""
:return: Return a :class:`~graphio.NodeSet` instance for this ModelNode.
"""
return NodeSet(cls.__labels__, merge_keys=cls.__merge_keys__)
def small_nodeset() -> NodeSet:
ns = NodeSet(['Test'], merge_keys=['uuid'])
for i in range(100):
ns.add_node({'uuid': i, 'key': 'value'})
return ns
def test_update_nodes(self):
ns = NodeSet(['Test'], ['name'], indexed=True)
ns.add_node({'name': 'Peter', 'age': 50})
ns.update_node({'name': 'Peter', 'age': 60, 'city': 'Munich'})
assert len(ns.nodes) == 1
assert ns.nodes[0] == {'name': 'Peter', 'age': 60, 'city': 'Munich'}
def test_error_on_non_indexed(self):
ns = NodeSet(['Test'], ['name'])
with pytest.raises(TypeError):
ns.update_node({'name': 'Peter', 'age': 60, 'city': 'Munich'})
def create_nodes_test(graph, clear_graph):
ns1 = NodeSet(['Test'], merge_keys=['uuid'])
ns2 = NodeSet(['Foo'], merge_keys=['uuid'])
for i in range(100):
ns1.add_node({'uuid': i})
ns2.add_node({'uuid': i})
ns1.create(graph)
ns2.create(graph)
return ns1, ns2
def test_nodeset_merge_key_id():
ns = NodeSet(['Test'], ['name', 'foo'])
merge_key_id = ns._merge_key_id({'name': 'Peter', 'foo': 'bar'})
assert merge_key_id == ('Peter', 'bar')
def test_nodeset_merge_preserve(self, graph, clear_graph):
"""
Merge a nodeset 3 times and check number of nodes.
"""
ns = NodeSet(['Test'], merge_keys=['uuid'])
for i in range(100):
ns.add_node({'uuid': i, 'key': 'value'})
ns.merge(graph)
do_not_overwrite_ns = NodeSet(['Test'],
merge_keys=['uuid'],
preserve=['key'])
for i in range(100):
do_not_overwrite_ns.add_node({'uuid': i, 'key': 'other_value'})
do_not_overwrite_ns.merge(graph)
assert list(
graph.run("MATCH (n:Test) where n.key = 'value' RETURN count(n)")
)[0][0] == 100
assert list(
graph.run(
"MATCH (n:Test) where n.key = 'other_value' RETURN count(n)")
)[0][0] == 0
def test_create_instance_add_nodes(self, labels, merge_keys, data):
ns = NodeSet(labels, merge_keys)
for i in data:
ns.add_node(i)
def test_relationshipset_csv_merge(self, graph, clear_graph,
neo4j_import_dir):
# create the nodes required here
ns1 = NodeSet(['Test', 'Other'], merge_keys=['uuid', 'numerical'])
ns2 = NodeSet(['Foo', 'SomeLabel'], merge_keys=['uuid', 'value'])
for i in range(20):
ns1.add_node({'uuid': i, 'numerical': 1})
ns2.add_node({'uuid': i, 'value': 'foo'})
ns1.create_index(graph)
ns1.create(graph)
ns2.create_index(graph)
ns2.create(graph)
rs = RelationshipSet('TEST', ['Test', 'Other'], ['Foo', 'SomeLabel'],
['uuid', 'numerical'], ['uuid', 'value'])
rs.uuid = 'peter'
rs.create_index(graph)
for i in range(10):
rs.add_relationship({
'uuid': i,
'numerical': 1
}, {
'uuid': i,
'value': 'foo'
}, {
'value': i,
'other_value': 'peter'
})
# add a few relationships with different props
for i in range(10, 20):
rs.add_relationship({
'uuid': i,
'numerical': 1
}, {
'uuid': i,
'value': 'foo'
}, {
'second_value': i,
'other_second_value': 'peter'
})
path = rs.to_csv(neo4j_import_dir)
# note: this is a hack to copy files into a running Docker container from Python
# needed to run the tests without too many changes locally and in GitHub Actions
copy_to_all_docker_containers(path, '/var/lib/neo4j/import')
query = rs.csv_query('MERGE')
# run query a few times to check for duplications
graph.run(query)
graph.run(query)
graph.run(query)
result = graph.run(
"MATCH (source:Test:Other)-[r:TEST]->(target:Foo:SomeLabel) RETURN r"
).data()
assert len(result) == len(rs.relationships)