我试图在内存中保留一个带有参考数据的大模型,并用推理规则附加其他模型,当新的三元组出现时,我需要匹配规则来触发,查找参考数据和所有新的三元组以结束第二个模型。
像这样:
#Reference data:
:alice
foaf:name "Alice" ;
schema:identifier "1" ;
.
#New incoming triples:
:claire
foaf:name "Claire" ;
schema:identifier "3" ;
:friend_id "1" ;
.
当 Claire 到达时,应该触发一个规则,通过 id 在参考数据中找到 Alice 并生成额外的三元组
:alice foaf:knows :claire
这应该与正在填充的模型中的 Claire 记录一起结束,而不是 Alice 所在的那个。
我对此的最新尝试:
private void createUnionModel() {
Model referred = RDFDataMgr.loadModel("referred.ttl");
Reasoner reasoner = new GenericRuleReasoner(Rule.rulesFromURL("align.sse"));
Model aligned = ModelFactory.createDefaultModel();
Model union = ModelFactory.createUnion(aligned, referred);
InfModel infModel = ModelFactory.createInfModel(reasoner, union);
RDFDataMgr.read(infModel, "new_triples.ttl");
printModel(referred, "referred");
printModel(aligned, "aligned");
printModel(union, "union");
printModel(infModel, "inference");
}
表明推理规则生成的三元组不会进入aligned
模型:
'alignment rules loaded'
'alignment rules loaded'
----------------- referred
:alice a foaf:Person ;
schema:identifier "1" ;
foaf:name "Alice" .
:bob a foaf:Person ;
schema:identifier "2" ;
foaf:name "Bob" .
----------------- aligned
<http://example.org/claire>
a <http://xmlns.com/foaf/0.1/Person> ;
<http://example.org/friend_id> "1" ;
<http://schema.org/identifier> "3" ;
<http://xmlns.com/foaf/0.1/name>
"Claire" .
----------------- union
:claire a foaf:Person ;
:friend_id "1" ;
schema:identifier "3" ;
foaf:name "Claire" .
:alice a foaf:Person ;
schema:identifier "1" ;
foaf:name "Alice" .
:bob a foaf:Person ;
schema:identifier "2" ;
foaf:name "Bob" .
----------------- inference
'alignment rules loaded'
'rule matched: ' <http://example.org/claire> ' knows ' <http://example.org/alice>
:claire a foaf:Person ;
schema:identifier "3" ;
foaf:knows :alice ;
foaf:name "Claire" .
:alice a foaf:Person ;
schema:identifier "1" ;
foaf:name "Alice" .
:bob a foaf:Person ;
schema:identifier "2" ;
foaf:name "Bob" .
我试图以不同的方式组合模型,即首先构建推理模型,然后将其与参考相结合,但没有帮助。
如何将推理规则生成的三元组与查找数据分开保存在模型中?
有没有办法在ttl
文件而不是 Java 代码中定义它?