[
https://issues.apache.org/jira/browse/UIMA-3346?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13826231#comment-13826231
]
Andrew MacKinlay edited comment on UIMA-3346 at 11/19/13 10:46 PM:
-------------------------------------------------------------------
I personally would be happy with either approach. There's no recommended way to
do this already, apart from from manually calling e.g.
CollectionReaderDescription.toXML ? (I note that this approach, by default,
dumps the whole type system to the AE descriptor).
was (Author: admackin):
I personally would be happy with either approach. There's no recommended way to
do this already, apart from from manually calling
AnalysisEngineDescriptor.toXML ? (I note that this approach, by default, dumps
the whole type system to the AE descriptor).
> "generate" goal should include type system imports
> --------------------------------------------------
>
> Key: UIMA-3346
> URL: https://issues.apache.org/jira/browse/UIMA-3346
> Project: UIMA
> Issue Type: Improvement
> Components: uimaFIT-Maven-Plugin
> Affects Versions: 2.1.0uimaFIT
> Reporter: Jens Grivolla
> Priority: Minor
>
> The "generate" goal of the uimaFIT maven plugin should include type system
> imports in the generated descriptors. This would make those descriptors
> directly usable in descriptor-based workflows such as CPE or UIMA-AS.
> In our case, we point to the TS descriptor files directly in types.txt, thus
> not relying on any "magic" TS discovery. This could translate directly to a
> corresponding import by name in the XML descriptor. More sophisticated setups
> might of course be more difficult to handle.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
