[
https://issues.apache.org/jira/browse/UIMA-3591?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13903327#comment-13903327
]
Richard Eckart de Castilho commented on UIMA-3591:
--------------------------------------------------
I'm afraid, this issue cannot be fixed.
uimaFIT correctly coerces single values on multi-valued parameters *if it knows
that the parameter is multi-valued*. This knowledge is obtained from the
resource meta data that is part of a descriptor. The descriptor can come either
from an XML file, or it can have been dynamically generated using uimaFIT. In
this issue, the descriptor is dynamically created using uimaFIT.
For uimaFIT to know that a parameter exists and is multi-valued, a class field
must have been marked using the @ConfigurationParameter annotation. If the
field is an array or collection type, then uimaFIT will correctly mark the
parameter as multi-valued in the descriptor.
In the present case, uimaFIT analyzes the WhitespaceTokenizer class which does
not contain any @ConfigurationParameter annotations. Thus, the descriptor
generated for this class does not contain any parameter declarations.
Consequently, none of the type coercion mechanisms can function, as these rely
on the parameter declarations. uimaFIT will still be able to initialize the
component, but only if the parameters passed in the
createEngineDescription-call match exactly the types expected by the UIMA
component.
> Multi-values parameter does not accept single value when
> @ConfigurationParameter is not present
> -----------------------------------------------------------------------------------------------
>
> Key: UIMA-3591
> URL: https://issues.apache.org/jira/browse/UIMA-3591
> Project: UIMA
> Issue Type: Bug
> Components: uimaFIT
> Affects Versions: 2.0.0uimaFIT
> Reporter: Richard Eckart de Castilho
>
> This is not working
> {noformat}
> AnalysisEngineFactory.createEngineDescription(WhitespaceTokenizer.class,
> "SofaNames", SimpleParserAE.SOFA_NAME_TEXT_ONLY);
> {noformat}
> I got a ClassCastException
> {noformat}
> Caused by: java.lang.ClassCastException: java.lang.String cannot be cast to
> [Ljava.lang.String;
> at
> org.apache.uima.annotator.WhitespaceTokenizer.initialize(WhitespaceTokenizer.java:328)
> at
> org.apache.uima.analysis_engine.impl.PrimitiveAnalysisEngine_impl.initializeAnalysisComponent(PrimitiveAnalysisEngine_impl.java:252)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
