from import_utils import GraphImporter
parser = argparse.ArgumentParser(description='DBLP Importer')
parser.add_argument('--db', default="http://localhost:7474/db/data/")
parser.add_argument('--data_dir', '-d', default='/vagrant/_data', help='data directory')
parser.add_argument('--undirected', action='store_true', help='create undirected graph')
parser.add_argument('--sets', action='store_true', help='create set edges')
parser.add_argument('--clear', action='store_true', help='clear the graph')
parser.add_argument('--commitEvery', type=int, default=100, help='commit every x steps')
args = parser.parse_args()
importer = GraphImporter(args.db, args.commitEvery)
if args.clear:
importer.delete_all()
def fix_string(v):
v = v.replace(u'"',u"'")
v = v.replace(u'\\',u'\\\\')
return v
class DBLPImporter(xml.sax.ContentHandler):
def __init__(self):
self.c = 0
self.last = None
self.attr = None
self.attrs = set(['author','editor','title','booktitle','pages','year','address','journal','volume','number','month','url','ee','cdrom','cite','publisher','note','crossref','isbn','series','school','chapter'])
self.nodes = set(['article','inproceedings','proceedings','book','incollection','phdthesis','mastersthesis','www'])
parser.add_argument('--data_file',
'-d',
default='./data/marclab_476_4hops.json',
help='data directory')
# parser.add_argument('--data_dir', '-d', default='/vagrant_data/kegg/',
# help='data directory')
parser.add_argument('--clear', action='store_true', help='clear the graph')
parser.add_argument('--commitEvery',
type=int,
default=100,
help='commit every x steps')
args = parser.parse_args()
importer = GraphImporter(args.db, args.commitEvery)
if args.clear or True:
importer.delete_all()
with open(args.data_file) as f:
doc = json.load(f)
set_ids = dict()
current_set_id = 0
nodes = doc["nodes"]
for node in nodes:
if node["Label"] is not None:
if node["Label"] in set_ids:
setId = set_ids[node["Label"]]
else:
