How to access feature values that are not primitive types in Ruta script? - uima

Question

I can access features that are defined as primitive types using Ruta script.
For example, posTag is a string feature of TokenAnnotation. The following script works.
STRING str1, str2;
TokenAnnotation{->GETFEATURE("posTag", str1), LOG("posTag=" + str1)};
However, I can't read feature that is defined as another Annotation type.
TokenAnnotation inherits a feature called lemma, which is type Lemma.
Lemma has it's own features. "key" is one of them.
How do I access "key" feature value of a Lemma through a given TokenAnnotation in Ruta script?
I tried Type variable. but, I don't know what I can do with it after assigning lemma feature to a type variable.
It would be great if someone can show me some examples of Type variable usage.
The following is my type descriptions, and cas.
Thanks in advance.
<typeDescription>
<name>uima.tt.TokenLikeAnnotation</name>
<description>Base type for token annotation types</description>
<supertypeName>uima.tt.LexicalAnnotation</supertypeName>
<features>
<featureDescription>
<name>lemma</name>
<description>The best probable entry containing all morphological information for the token</description>
<rangeTypeName>uima.tt.Lemma</rangeTypeName>
</featureDescription>
<featureDescription>
<name>lemmaEntries</name>
<description>List of lemma entries containing all morphological information for the token</description>
<rangeTypeName>uima.cas.FSArray</rangeTypeName>
</featureDescription>
<featureDescription>
<name>dictionaryMatch</name>
<description>A flag indicating whether or not the token matches a dictionary entry</description>
<rangeTypeName>uima.cas.Boolean</rangeTypeName>
</featureDescription>
</features>
</typeDescription>
<typeDescription>
<name>uima.tt.TokenAnnotation</name>
<description>General token annotation type. It is also the base type for the special token types</description>
<supertypeName>uima.tt.TokenLikeAnnotation</supertypeName>
<features>
<featureDescription>
<name>posTag</name>
<description>Part-of-Speech tag</description>
<rangeTypeName>uima.cas.String</rangeTypeName>
</featureDescription>
</features>
</typeDescription>
<typeDescription>
<name>uima.tt.KeyStringEntry</name>
<description>Base type for types defining key/value feature (e.g. uima.tt.Lemma type)</description>
<supertypeName>uima.cas.TOP</supertypeName>
<features>
<featureDescription>
<name>key</name>
<description>A key/value feature (e.g. lemma string in uima.tt.Lemma type)</description>
<rangeTypeName>uima.cas.String</rangeTypeName>
</featureDescription>
</features>
</typeDescription>
<typeDescription>
<name>uima.tt.Lemma</name>
<description>Morphological information retrieved from a lexical dictionary entry</description>
<supertypeName>uima.tt.KeyStringEntry</supertypeName>
<features>
<featureDescription>
<name>partOfSpeech</name>
<description>An integral encoding representing the part-of-speech for the lemma</description>
<rangeTypeName>uima.cas.Integer</rangeTypeName>
</featureDescription>
<featureDescription>
<name>frost_ExtendedPOS</name>
<description>An integer representing additional information related to the part-of-speech</description>
<rangeTypeName>uima.cas.Integer</rangeTypeName>
</featureDescription>
<featureDescription>
<name>isStopword</name>
<description/>
<rangeTypeName>uima.cas.Boolean</rangeTypeName>
</featureDescription>
</features>
</typeDescription>
<cas:NULL xmi:id="0"/>
<tcas:DocumentAnnotation xmi:id="8" sofa="1" begin="0" end="70" language="en"/>
<tcas:DocumentAnnotation xmi:id="21" sofa="14" begin="0" end="22" language="en"/>
<ontology:Column xmi:id="27" sofa="14" begin="0" end="11"/>
<ontology:Column xmi:id="31" sofa="14" begin="12" end="22"/>
<ontology:Table xmi:id="35" sofa="14" begin="0" end="22"/>
<uimatypes:TitlecaseAlphabetic xmi:id="42" sofa="14" begin="0" end="8" lemma="74" lemmaEntries="74" dictionaryMatch="true" posTag="NN"/>
<uimatypes:TitlecaseAlphabetic xmi:id="58" sofa="14" begin="12" end="17" lemma="90" lemmaEntries="90" dictionaryMatch="true" posTag="NN"/>
<uimatypes:TitlecaseAlphabetic xmi:id="66" sofa="14" begin="18" end="22" lemma="98" lemmaEntries="98" dictionaryMatch="true" posTag="NN"/>
<uimatypes:UppercaseAlphabetic xmi:id="50" sofa="14" begin="9" end="11" lemma="82" lemmaEntries="82" dictionaryMatch="true" posTag="NN"/>
<tt:SentenceAnnotation xmi:id="106" sofa="14" begin="0" end="22" sentenceNumber="1"/>
<tt:ParagraphAnnotation xmi:id="111" sofa="14" begin="0" end="22" paragraphNumber="1"/>
<type:CW xmi:id="116" sofa="14" begin="0" end="8"/>
<type:CW xmi:id="132" sofa="14" begin="12" end="17"/>
<type:CW xmi:id="140" sofa="14" begin="18" end="22"/>
<type:SPACE xmi:id="120" sofa="14" begin="8" end="9"/>
<type:SPACE xmi:id="128" sofa="14" begin="11" end="12"/>
<type:SPACE xmi:id="136" sofa="14" begin="17" end="18"/>
<type:CAP xmi:id="124" sofa="14" begin="9" end="11"/>
<type:RutaBasic xmi:id="144" sofa="14" begin="0" end="8"/>
<type:RutaBasic xmi:id="149" sofa="14" begin="8" end="9"/>
<type:RutaBasic xmi:id="154" sofa="14" begin="9" end="11"/>
<type:RutaBasic xmi:id="159" sofa="14" begin="11" end="12"/>
<type:RutaBasic xmi:id="164" sofa="14" begin="12" end="17"/>
<type:RutaBasic xmi:id="169" sofa="14" begin="17" end="18"/>
<type:RutaBasic xmi:id="174" sofa="14" begin="18" end="22"/>
<demo:Identifier xmi:id="187" sofa="14" begin="0" end="11"/>
<cas:Sofa xmi:id="1" sofaNum="2" sofaID="global1" mimeType="text" sofaString="<Table><Column>Employee ID</Column><Column>Birth Date</Column></Table>"/>
<cas:Sofa xmi:id="14" sofaNum="3" sofaID="global2" mimeType="text" sofaString="Employee ID Birth Date"/>
<tt:Lemma xmi:id="74" key="employee" partOfSpeech="3" frost_ExtendedPOS="0" isStopword="false"/>
<tt:Lemma xmi:id="90" key="birth" partOfSpeech="3" frost_ExtendedPOS="0" isStopword="false"/>
<tt:Lemma xmi:id="98" key="date" partOfSpeech="3" frost_ExtendedPOS="0" isStopword="false"/>
<tt:Lemma xmi:id="82" key="id" partOfSpeech="3" frost_ExtendedPOS="0" isStopword="false"/>
<cas:View sofa="1" members="8"/>
<cas:View sofa="14" members="21 27 31 35 42 58 66 50 106 111 116 132 140 120 128 136 124 144 149 154 159 164 169 174 187"/>

Answer 1

Yes, you can do that, but the support for TOP feature structure is limitied in UIMA Ruta 2.4.0 (current release). Only annotation types are supported.
This works best with feature expressions in UIMA Ruta. You can simply use a dot to refer to a feature of the matched annotation (can be stacked).
In your example, this could look like:
TokenAnnotation.lemma.key=="birth"{-> T1};
TokenAnnotation{TokenAnnotation.posTag=="NN" -> T1};
... but you probably need to replace uima.cas.TOP with uima.tcas.Annotation in your type system.
A type veriable won't help to solve this, but the new annotation variables could help here, but are not really needed.
DISCLAIMER: I am a developer of UIMA Ruta