def file2graph(self, kg: KG, filePath, format='nt'):
g = rdflib.Graph()
g.parse(filePath, format=format)
relation_triples, attribute_triples = set(), set()
# 循环遍历图谱中每一个三元组
for subj, pred, obj in g:
# 空图谱检查
if (subj, pred, obj) not in g:
raise Exception("It better be!")
if isinstance(obj, rdflib.term.URIRef):
triple = (str(subj).split('/')[-1], str(pred).split('/')[-1],
str(obj).split('/')[-1])
# triple = (str(subj).split('/')[-1], str(pred).split('/')[-1], str(obj).split('/')[-1])
relation_triples.add(triple)
elif isinstance(obj, rdflib.term.Literal):
triple = (str(subj).split('/')[-1], str(pred).split('/')[-1],
str(obj).split('/')[-1])
# triple = (str(subj).split('/')[-1], str(pred).split('/')[-1], str(obj).split('/')[-1])
attribute_triples.add(triple)
kg.add_relations(relation_triples)
kg.add_attributes(attribute_triples)
def __init__(self):
self.__score = 0
self.__metaPaths = []
self.__graph = KG()
self.__Aconstraint = set()
class MeatGraph(object):
def __init__(self):
self.__score = 0
self.__metaPaths = []
self.__graph = KG()
self.__Aconstraint = set()
def __lt__(self, other):
return self.score < other.score
def __gt__(self, other):
return self.score > other.score
def __ge__(self, other):
return self.score > other.score
def __le__(self, other):
return self.score <= other.score
def __hash__(self):
return hash(self.__score)
def __eq__(self, other):
if isinstance(other, self.__class__):
return self.__score == other.score
else:
return False
@property
def metaPaths(self):
return self.__metaPaths
@property
def graph(self):
return self.__graph
def addMetaPath(self, metaPath):
if isinstance(metaPath, MetaPath):
self.__metaPaths.append(metaPath)
self.score = self.merge(self.__metaPaths)
elif isinstance(metaPath, MeatGraph):
for path in metaPath.metaPaths:
self.__metaPaths.append(path)
self.score = self.merge(self.__metaPaths)
else:
print('It is not a MeatGraph/MetaPath object')
def merge(self, paths):
allP = set()
for p in paths:
allP |= set(path2Triples(p))
if len(p) > 1:
for t in set(path2Triples(p)):
self.__graph.add_relations({
(t.head.Class, t.relation.Class, t.tail.Class)
})
pathProb = 1
for T in allP:
pathProb *= T.prob
return pathProb
@property
def score(self):
return self.__score
@score.setter
def score(self, score):
self.__score = score
@property
def Aconstraint(self):
return self.__Aconstraint
# cst: (entity, attribution, value)
def addAconstraint(self, cst: tuple):
self.__Aconstraint.add(cst)
# r1 = ReadRDFFile()
# kg = KG()
# # r1.file2graph(kg, "D:/ws/aa/KeywordSearch/dataSet/CTIKG/CyberSecurity.nt")
# r1.file2graph(kg, "D:/ws/aa/KeywordSearch/dataSet/CTIKG/CTIKG.nt")
#
#
# # Key = NTNav(kg, ["WannaCry","Attack Pattern", "platform"])
# print("entities_set" + str(sys.getsizeof(kg.entities_set)))
# print("relations_set" + str(sys.getsizeof(kg.relations_set)))
# print("rev_hr_dict" + str(sys.getsizeof(kg.rev_hr_dict)))
# print("rev_rt_dict" + str(sys.getsizeof(kg.rev_rt_dict)))
# Key = NTNav(kg, ["WannaCry","Attack Pattern"])
# # Key = NTNav(kg, ["BlackEnergy","system","Vulnerability"])
# # Key = NTNav(kg, ["BlackEnergy","ApplicablePlatform","Weakness"])
# Key.search()
kg = KG()
r = ReadRDFFile()
r.file2graph(kg, "D:/ws/aa/KeywordSearch/dataSet/Yago4/yago-wd-class.nt")
r.file2graph(kg, "D:/ws/aa/KeywordSearch/dataSet/Yago4/yago-wd-full-types.nt")
r.file2graph(kg, "D:/ws/aa/KeywordSearch/dataSet/Yago4/yago-wd-facts.nt")
# print("entities_set" + str(sys.getsizeof(kg.entities_set)))
# print("relations_set" + str(sys.getsizeof(kg.relations_set)))
# print("rev_hr_dict" + str(sys.getsizeof(kg.rev_hr_dict)))
# print("rev_rt_dict" + str(sys.getsizeof(kg.rev_rt_dict)))
Key = NTNav(kg,["Amit Singhal","Google"])
Key.search()