Hi,
the important part is missing in the documentation. Sorry for that, I
will fix the documentation.
EXEC has another parameter, a list of types, which are internally
reindexed in Ruta. This is necessary for therule inference and
conditions like PARTOF, which would be extremely inefficient otherwise.
If you use the following rule, then you are able to match/use all
annotations, which are added by the HtmlAnnotator
Document{-> EXEC(HtmlAnnotator, {TAG})}; // or only "{A}" in your case
There might be another problem, if you set the parameter "onlyContent"
of the HtmlAnnotator to false. Then, you additionally need to make the
markup annotations visible:
Document{-> RETAINTYPE(MARKUP)};
I hope this helps.
Best,
Peter
On 25.06.2013 17:13, Fouad HAMDI wrote:
> Hi,
>
> I am currently trying the Ruta workbench and I would like to know if it is
> possible to refer to annotations put by an engine from a script.
>
> Here is a simple example: I want to analyze an html document and I would like
> to mark all the links inside the document. I write this script:
>
> ENGINE utils.HtmlAnnotator;
>
> TYPESYSTEM utils.HtmlTypeSystem;
> DECLARE Link;
>
> Document{-> EXEC(HtmlAnnotator)};
>
> org.apache.uima.ruta.type.html.A {-> MARK(Link)};
>
> I test with the following input:
>
> <html>
> <body>
> <ul>
> <li><a href="#">Link 1</a></li>
> <li><a href="#">Link 2</a></li>
> </ul>
>
> <p><a href="#">Link 3</a></li></p>
> </body>
> </html>
>
> In the output, I expected to find 3 "Link" annotations but I obtained none.
> However, the output file contains the "org.apache.uima.ruta.type.html.A"
> annotations as expected.
>
> I didn't find how to use the annotations put by an engine on the document in
> the remaining of my script.
>
> I conclude that the Document object annotated by the engine is not accessible
> from the script after executing this engine.
>
> Is this possible to do such an operation in RUTA scripts ?
>
> Thank you for your help.
>
> Best regards,
> Fouad
