UIMA Ruta 错误锦集（一）：MATCHEDTEXT用法

MATCHEDTEXT用来提取一个variable里的text。

如：提取一个句子后，在selection view里查看句子中各个成分的content时，我们会看到：
patternValue1是对这部分内容的描述，而非text本身。而patternValue2-4是text本身。
造成此情况的原因是，在create时，patternValue1后面是variable，而余下三个则被赋值为对应variable包含的text。
代码如下：

DECLARE PredicateRef;
STRING test1,test2,test3, test4;
STRING M1,M2,M3,M4;
STRING Geo1,Geo2,Geo3,Geo4;
p1:Predicate{AND(OR(p1.prSemantics=="desc_obj_state",p1.prSemantics=="desc_obj_augmentation", p1.prSemantics=="desc_obj_reduction"), OR(p1.prStructure=="[epiv ipv]", p1.prStructure=="ipv"))} Lemma.value=="by"?{->MARK(PredicateRef, 1, 2)}; 


Sentence {->MATCHEDTEXT(ExtractText)}->{
    //Comparing a to b, indicator1 predicates v1 in Geo1(la position may change), predicates2 v2 in Geo2, predicate3 v3 in Geo3 and predicate4 v4 in Geo4” 
    CompareTo 
    i1:Indicator Where?{->MATCHEDTEXT(Geo1)} PredicateRef{->MATCHEDTEXT(test1)} v1:Measure{->MATCHEDTEXT(M1)} Where?{->MATCHEDTEXT(Geo1)}
    //ANY+? in case we have () in between
    ANY*? Connectors 
    Where?{->MATCHEDTEXT(Geo2)} PredicateRef?{->MATCHEDTEXT(test2)} VariAdv?{->MATCHEDTEXT(test2)} v2:Measure{->MATCHEDTEXT(M2)} Where?{->MATCHEDTEXT(Geo2)}
    //3
    ANY*? Connectors 
    Where?{->MATCHEDTEXT(Geo3)} PredicateRef?{->MATCHEDTEXT(test3)} VariAdv?{->MATCHEDTEXT(test3)} v3:Measure{->MATCHEDTEXT(M3)} Where?{->MATCHEDTEXT(Geo3)}
    //4
    ANY*? Connectors 
    Where?{->MATCHEDTEXT(Geo4)} PredicateRef?{->MATCHEDTEXT(test4)} VariAdv?{->MATCHEDTEXT(test4)} v4:Measure{->MATCHEDTEXT(M4)} Where?{->MATCHEDTEXT(Geo4)}
    ANY*? Connectors 
    {->CREATE(MergeCompareValueWithReference, 1, 50, "patternIndicator1"=i1.kpiToken, "patternPredicate1"=test1, "patternValue1"=v1, "entityGeo1"=Geo1, "patternPredicate2"=test2, "patternValue2"=M2, "entityGeo2"=Geo2, "patternPredicate3"=test3, "patternValue3"=M3, "entityGeo3"=Geo3, "patternPredicate4"=test4, "patternValue4"=M4, "entityGeo4"=Geo4, "match"=ExtractText)};
    //"patternIndicator2"=i2.kpiToken, "patternPredicate2"=PredicateRef.prToken, "patternPredicate2"=VariAdv.prToken, "patternValue2"=v2, "match"=ExtractText)};
};

最开始定义M1,M2,M3,M4用来储存text。

在代码中用v1:Measure{->MATCHEDTEXT(M1)}提取变量v1的text。在最后create时候，若"patternValue1"=v1, 则selection view中出现描述。若"patternValue1"=M1, 则selction view中展现句子中v1代表的具体text。