def test_csv_to_neo_load():
"""
load csv to neo4j test
"""
n = NeoTransformer()
n.save_from_csv("tests/resources/x1n.csv", "tests/resources/x1e.csv")
class KGraph:
def __init__(self):
self.ports = ports = {'bolt_port': 7687}
self.xform = NeoTransformer(self.ports)
def transform(self):
self.xform.save_node({
"id": "AWESOME:23534",
"name": "Really Great",
"category": "bioentity"
})
def test_load(clean_slate, query):
t = NeoTransformer(query[0],
uri=DEFAULT_NEO4J_URL,
username=DEFAULT_NEO4J_USERNAME,
password=DEFAULT_NEO4J_PASSWORD)
t.save()
nr = t.http_driver.query("MATCH (n) RETURN count(n)")
[node_counts] = [x for x in nr][0]
assert node_counts == query[1]
er = t.http_driver.query("MATCH ()-[p]->() RETURN count(p)")
[edge_counts] = [x for x in er][0]
assert edge_counts == query[2]
def load_and_merge(yaml_file: str) -> Transformer:
"""Load and merge sources defined in the config YAML.
Args:
yaml_file: A string pointing to a KGX compatible config YAML.
Returns:
kgx.Transformer: The merged transformer that contains the merged graph.
"""
config = parse_load_config(yaml_file)
transformers: List = []
# read all the sources defined in the YAML
for key in config['target']:
target = config['target'][key]
logging.info("Loading {}".format(key))
if target['type'] in get_file_types():
# loading from a file
transformer = get_transformer(target['type'])()
for f in target['filename']:
transformer.parse(f, input_format='tsv')
transformers.append(transformer)
elif target['type'] == 'neo4j':
transformer = NeoTransformer(None, target['uri'],
target['username'],
target['password'])
transformer.load()
transformers.append(transformer)
else:
logging.error("type {} not yet supported".format(target['type']))
# merge all subgraphs into a single graph
merged_transformer = Transformer()
merged_transformer.merge_graphs([x.graph for x in transformers])
merged_transformer.report()
# write the merged graph
if 'destination' in config:
destination = config['destination']
if destination['type'] in ['csv', 'tsv', 'ttl', 'json', 'tar']:
destination_transformer = get_transformer(destination['type'])(
merged_transformer.graph)
destination_transformer.save(destination['filename'],
extension=destination['type'])
elif destination['type'] == 'neo4j':
destination_transformer = NeoTransformer(
merged_transformer.graph,
uri=destination['uri'],
username=destination['username'],
password=destination['password'])
destination_transformer.save_with_unwind()
else:
logging.error(
"type {} not yet supported for KGX load-and-merge operation.".
format(destination['type']))
return merged_transformer
def test_neo_to_graph_upload():
""" loads a neo4j graph from a json file
"""
jt = JsonTransformer()
jt.parse('resources/robodb2.json')
nt = NeoTransformer(jt.graph, host='localhost', port='7474', username='******', password='******')
nt.save_with_unwind()
nt.neo4j_report()
def test_neo_to_graph_transform():
"""
load from neo4j and transform to nx graph
"""
nt = NeoTransformer(host='localhost', port='7474', username='******', password='******')
nt.load()
nt.report()
t = PandasTransformer(nt.graph)
t.save(os.path.join(target_dir, "neo_graph.csv"))
def test_csv_to_neo_load():
"""
load csv to neo4j test
"""
pt = PandasTransformer()
pt.parse(os.path.join(resource_dir, "cm_nodes.csv"))
pt.parse(os.path.join(resource_dir, "cm_edges.csv"))
nt = NeoTransformer(pt.graph, host='localhost', port='7474', username='******', password='******')
nt.save_with_unwind()
nt.neo4j_report()
def test_neo_to_graph_transform():
"""
load from neo4j and transform to nx graph
"""
n = NeoTransformer()
n.load()
n.report()
t = PandasTransformer(n)
t.save("target/neo_graph.csv")
def test_get_edges(clean_slate, query):
g = get_graph('kgx-unit-test')[3]
t = NeoTransformer(g,
uri=DEFAULT_NEO4J_URL,
username=DEFAULT_NEO4J_USERNAME,
password=DEFAULT_NEO4J_PASSWORD)
t.save()
for k, v in query[1].items():
t.set_edge_filter(k, v)
edges = t.get_edges()
edge_list = [x[1] for x in edges]
assert len(edges) == query[2]
def test_neo_to_graph_transform():
"""
load from neo4j and transform to nx graph
"""
nt = NeoTransformer(uri=DEFAULT_NEO4J_URL,
username=DEFAULT_NEO4J_USERNAME,
password=DEFAULT_NEO4J_PASSWORD)
nt.load()
nt.report()
t = PandasTransformer(nt.graph)
t.save(os.path.join(target_dir, "neo_graph.csv"))
def test_neo_to_graph_upload():
""" loads a neo4j graph from a json file
"""
jt = JsonTransformer()
jt.parse('resources/robodb2.json')
nt = NeoTransformer(jt.graph,
uri=DEFAULT_NEO4J_URL,
username=DEFAULT_NEO4J_USERNAME,
password=DEFAULT_NEO4J_PASSWORD)
nt.save()
nt.neo4j_report()
def test_graph_to_neo_load():
"""
load nx graph to neo4j test
"""
t = PandasTransformer()
t.parse("tests/resources/x1n.csv")
t.parse("tests/resources/x1e.csv")
t.report()
n = NeoTransformer(t)
n.save()
n.neo4j_report()
def test_get_nodes(clean_slate, query):
g = query[0]
t = NeoTransformer(g,
uri=DEFAULT_NEO4J_URL,
username=DEFAULT_NEO4J_USERNAME,
password=DEFAULT_NEO4J_PASSWORD)
t.save()
for k, v in query[1].items():
t.set_node_filter(k, v)
nodes = t.get_nodes()
assert len(nodes) == query[2]
node_ids = [x['id'] for x in nodes]
for x in query[3]:
assert x in node_ids
def test_csv_to_neo_load():
"""
load csv to neo4j test
"""
pt = PandasTransformer()
pt.parse(os.path.join(resource_dir, "cm_nodes.csv"))
pt.parse(os.path.join(resource_dir, "cm_edges.csv"))
nt = NeoTransformer(pt.graph,
uri=DEFAULT_NEO4J_URL,
username=DEFAULT_NEO4J_USERNAME,
password=DEFAULT_NEO4J_PASSWORD)
nt.save()
nt.neo4j_report()
def test_csv_to_neo_load():
"""
load csv to neo4j test
"""
pt = PandasTransformer()
pt.parse("resources/x1n.csv")
pt.parse("resources/x1e.csv")
nt = NeoTransformer(pt.graph,
host='http://localhost',
port='7474',
username='******',
password='******')
nt.save_with_unwind()
nt.neo4j_report()
def test_neo_to_graph_transform():
"""
load from neo4j and transform to nx graph
"""
return
nt = NeoTransformer(host='localhost',
port='7474',
username='',
password='')
nt.load()
nt.report()
t = PandasTransformer(nt.graph)
t.save("target/neo_graph.csv")
def test_csv_to_neo_load():
"""
load csv to neo4j test
"""
return
pt = PandasTransformer()
pt.parse("resources/nodes.csv")
pt.parse("resources/edges.csv")
nt = NeoTransformer(pt.graph,
host='localhost',
port='7474',
username='',
password='')
nt.save_with_unwind()
nt.neo4j_report()
def test_save_merge(clean_slate):
g = get_graph('kgx-unit-test')[2]
t = NeoTransformer(g,
uri=DEFAULT_NEO4J_URL,
username=DEFAULT_NEO4J_USERNAME,
password=DEFAULT_NEO4J_PASSWORD)
t.save()
t.graph.add_node('B',
id='B',
publications=['PMID:1', 'PMID:2'],
category=['biolink:NamedThing'])
t.graph.add_node('C', id='C', source='kgx-unit-test')
t.graph.add_edge('A',
'B',
subject='A',
object='B',
edge_label='biolink:related_to',
test_prop='VAL123')
assert t.graph.number_of_nodes() == 3
t.save()
nr = t.http_driver.query("MATCH (n) RETURN n")
for node in nr:
data = node[0]['data']
if data['id'] == 'B':
assert 'category' in data and data['category'] == [
'biolink:NamedThing'
]
assert 'publications' in data and data['publications'] == [
'PMID:1', 'PMID:2'
]
er = t.http_driver.query("MATCH ()-[p]-() RETURN p",
data_contents=True,
returns=(Node, Relationship, Node))
for edge in er:
data = edge[0].properties
# assert data['id'] == 'A-biolink:related_to-B'
assert data['subject'] == 'A'
assert data['object'] == 'B'
assert data['edge_label'] == 'biolink:related_to'
assert data['test_prop'] == 'VAL123'
parser.add_argument('--node_filter',
action='append',
help='A filter that can be applied to nodes')
parser.add_argument('--edge_filter',
action='append',
help='A filter that can be applied to edges')
parser.add_argument('--uri',
help='URI/URL for Neo4j (including port)',
default='localhost:7474')
parser.add_argument('--username', help='username', default='neo4j')
parser.add_argument('--password', help='password', default='demo')
args = parser.parse_args()
# Initialize NeoTransformer
nt = NeoTransformer(None,
uri=args.uri,
username=args.username,
password=args.password)
if args.node_filter:
for f in args.node_filter:
k, v = f.split('=')
nt.set_node_filter(k, set(v))
if args.edge_filter:
for f in args.edge_filter:
k, v = f.split('=')
nt.set_edge_filter(k, set(v))
# Read from Neo4j with the given filter constraints (if any)
nt.load()
nt.report()
parser.add_argument('--edges', help='file with edges in TSV format')
parser.add_argument('--uri',
help='URI/URL for Neo4j (including port)',
default='localhost:7474')
parser.add_argument('--username', help='username', default='neo4j')
parser.add_argument('--password', help='password', default='demo')
args = parser.parse_args()
if args.nodes is None and args.edges is None:
usage()
exit()
# Initialize PandasTransformer
t = PandasTransformer()
# Load nodes and edges into graph
if args.nodes:
t.parse(args.nodes, input_format='tsv')
if args.edges:
t.parse(args.edges, input_format='tsv')
# Initialize NeoTransformer
n = NeoTransformer(t.graph,
uri=args.uri,
username=args.username,
password=args.password)
# Save graph into Neo4j
n.save()
n.neo4j_report()
def load_and_merge(yaml_file: str) -> nx.MultiDiGraph:
"""Load and merge sources defined in the config YAML.
Args:
yaml_file: A string pointing to a KGX compatible config YAML.
Returns:
networkx.MultiDiGraph: The merged graph.
"""
gm = GraphMerge()
config = parse_load_config(yaml_file)
transformers: List = []
# make sure all files exist before we start load
for key in config['target']:
target = config['target'][key]
logging.info("Checking that file exist for {}".format(key))
if target['type'] in get_file_types():
for f in target['filename']:
if not os.path.exists(f) or not os.path.isfile(f):
raise FileNotFoundError(
"File {} for transform {} in yaml file {} "
"doesn't exist! Dying.", f, key, yaml_file)
# read all the sources defined in the YAML
for key in config['target']:
target = config['target'][key]
logging.info("Loading {}".format(key))
if target['type'] in get_file_types():
# loading from a file
transformer = get_transformer(target['type'])()
for f in target['filename']:
transformer.parse(f, input_format='tsv')
transformer.graph.name = key
transformers.append(transformer)
elif target['type'] == 'neo4j':
transformer = NeoTransformer(None, target['uri'],
target['username'],
target['password'])
transformer.load()
transformers.append(transformer)
transformer.graph.name = key
else:
logging.error("type {} not yet supported".format(target['type']))
stats_filename = f"{key}_stats.yaml"
generate_graph_stats(transformer.graph, key, stats_filename)
# merge all subgraphs into a single graph
merged_graph = gm.merge_all_graphs([x.graph for x in transformers])
merged_graph.name = 'merged_graph'
generate_graph_stats(merged_graph, merged_graph.name,
f"merged_graph_stats.yaml")
# write the merged graph
if 'destination' in config:
for _, destination in config['destination'].items():
if destination['type'] == 'neo4j':
destination_transformer = NeoTransformer(
merged_graph,
uri=destination['uri'],
username=destination['username'],
password=destination['password'])
destination_transformer.save_with_unwind()
elif destination['type'] in get_file_types():
destination_transformer = get_transformer(
destination['type'])(merged_graph)
destination_transformer.save(destination['filename'],
extension=destination['type'])
else:
logging.error(
"type {} not yet supported for KGX load-and-merge operation."
.format(destination['type']))
return merged_graph
default='7474')
parser.add_argument('--username',
help='username (default: neo4j)',
default='neo4j')
parser.add_argument('--password',
help='password (default: demo)',
default='demo')
args = parser.parse_args()
if args.nodes is None or args.edges is None:
usage()
exit()
print(args)
# Initialize PandasTransformer
t = PandasTransformer()
# Load nodes and edges into graph
t.parse(args.nodes, error_bad_lines=False)
t.parse(args.edges, error_bad_lines=False)
# Initialize NeoTransformer
# TODO: eliminate bolt
n = NeoTransformer(t.graph, args.host, {'http': args.http_port}, args.username,
args.password)
# Save graph into Neo4j
n.save_with_unwind()
n.neo4j_report()
def test_create_constraint_query(category):
sanitized_category = NeoTransformer.sanitize_category([category])
q = NeoTransformer.create_constraint_query(sanitized_category)
assert q == f"CREATE CONSTRAINT ON (n:{sanitized_category}) ASSERT n.id IS UNIQUE"
def test_sanitize_category():
categories = ['biolink:Gene', 'biolink:GeneOrGeneProduct']
s = NeoTransformer.sanitize_category(categories)
assert s == ['`biolink:Gene`', '`biolink:GeneOrGeneProduct`']
action='append',
help='A filter that can be applied to node and/or edges')
parser.add_argument('--host',
help='host to connect with Neo4j',
default='localhost')
parser.add_argument('--bolt_port',
help='Bolt port to connect with Neo4j',
default='7687')
parser.add_argument('--username',
help='username (default: neo4j)',
default='neo4j')
parser.add_argument('--password',
help='password (default: demo)',
default='demo')
args = parser.parse_args()
# Initialize NeoTransformer
n = NeoTransformer(None, args.host, {'bolt': args.bolt_port}, args.username,
args.password)
if args.filter is not None:
if len(args.filter) > 0:
for filter in args.filter:
k, v = filter.split('=')
# Set filters
n.set_filter(k, v)
# Read from Neo4j with the given filter constraints (if any)
n.load()
n.report()
def __init__(self):
self.ports = ports = {'bolt_port': 7687}
self.xform = NeoTransformer(self.ports)
def load_and_merge(yaml_file: str) -> nx.MultiDiGraph:
"""Load and merge sources defined in the config YAML.
Args:
yaml_file: A string pointing to a KGX compatible config YAML.
Returns:
networkx.MultiDiGraph: The merged graph.
"""
config = parse_load_config(yaml_file)
transformers: List = []
# make sure all files exist before we start load
for key in config['target']:
target = config['target'][key]
logging.info("Checking that file exist for {}".format(key))
if target['type'] in get_file_types():
for f in target['filename']:
if not os.path.exists(f) or not os.path.isfile(f):
raise FileNotFoundError(
"File {} for transform {} in yaml file {} "
"doesn't exist! Dying.", f, key, yaml_file)
# read all the sources defined in the YAML
for key in config['target']:
target = config['target'][key]
logging.info("Loading {}".format(key))
if target['type'] in get_file_types():
# loading from a file
try:
transformer = get_transformer(target['type'])()
if target['type'] in {'tsv', 'neo4j'}:
if 'filters' in target:
apply_filters(target, transformer)
for f in target['filename']:
transformer.parse(f, input_format='tsv')
transformer.graph.name = key
if 'operations' in target:
apply_operations(target, transformer)
transformers.append(transformer)
except:
logging.error("Failed loading {}".format(f))
elif target['type'] == 'neo4j':
transformer = NeoTransformer(None, target['uri'],
target['username'],
target['password'])
if 'filters' in target:
apply_filters(target, transformer)
transformer.load()
if 'operations' in target:
apply_operations(target, transformer)
transformers.append(transformer)
transformer.graph.name = key
else:
logging.error("type {} not yet supported".format(target['type']))
stats_filename = f"{key}_stats.yaml"
generate_graph_stats(transformer.graph, key, stats_filename)
# merge all subgraphs into a single graph
merged_graph = merge_all_graphs([x.graph for x in transformers])
merged_graph.name = 'merged_graph'
generate_graph_stats(merged_graph, merged_graph.name,
"merged_graph_stats.yaml", ['provided_by'],
['provided_by'])
# write the merged graph
if 'destination' in config:
for _, destination in config['destination'].items():
if destination['type'] == 'neo4j':
destination_transformer = NeoTransformer(
merged_graph,
uri=destination['uri'],
username=destination['username'],
password=destination['password'])
destination_transformer.save()
elif destination['type'] in get_file_types():
destination_transformer = get_transformer(
destination['type'])(merged_graph)
mode = 'w:gz' if destination['type'] in {'tsv'} else None
if destination['type'] in {'nt', 'nt.gz', 'ttl'}:
destination_transformer.set_property_types(PROPERTY_TYPES)
destination_transformer.save(destination['filename'],
output_format=destination['type'],
mode=mode)
else:
logging.error(
"type {} not yet supported for KGX load-and-merge operation."
.format(destination['type']))
return merged_graph
