Hi Peter,
Thank you for these information, I think I get the idea behind this notion of
view now, it was not very clear for me before.
Best regards,
Fouad
Le 26 juin 2013 à 18:33, Peter Klügl a écrit :
> Hi,
>
> the TikaAnnotator creates a new view "textView" that contains the
> MarkupAnnotations (I have seen no tika annotations in "_InitialView"). This
> means that you cannot call the TikaAnnotator and match its annotations in the
> same script. You could run two scripts, the first one calls the TikaAnnotator
> and the second one with the rules is then applied on the new view. The view
> on which the rules are applied can be specified in the run configuration. In
> a real application, this would be solved with sofa mappings in an aggregate
> analysis engine.
>
> Best,
>
> Peter
>
>
> On 26.06.2013 10:06, Fouad HAMDI wrote:
>> Hi Peter,
>>
>> Thanks for your answer, some precisions are described below:
>>
>>> Hi,
>>>
>>> On 26.06.2013 09:33, Fouad HAMDI wrote:
>>>> Hi,
>>>>
>>>> I am trying to use the Tika Annotator within RUTA scripts.
>>>>
>>>> Here is my current script:
>>>>
>>>> PACKAGE test;
>>>>
>>>> ENGINE tika.MarkupAnnotator;
>>>>
>>>> TYPESYSTEM tika.MarkupAnnotationTypeSystem;
>>>> DECLARE Link;
>>>>
>>>> Document{-> EXEC(MarkupAnnotator, {MarkupAnnotation})};
>>>>
>>>> MarkupAnnotation { FEATURE("name", "a") -> MARK(Link) };
>>>>
>>>>
>>>> and the file I try to annotate:
>>>>
>>>> <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>
>>>>
>>>> After I execute the script, I noticed two points:
>>>>
>>>> - The annotation browser view does not display the "MarkupAnnotation" tags
>>>>
>>>> - The "MarkupAnnotation" condition is not triggered so that no "Link" tag
>>>> is present in the output
>>>>
>>>> (When I visualize the output file in a text editor, the "MarkupAnnotation"
>>>> tags are here)
>>>
>>> There are MarkupAnnotations in the xmiCAS, but they are not listed in in
>>> the Annotation Browser? Or are they not visualized (highlighted)? In the
>>> first case, I would bet that the CAS Editor was opened with a type
>>> system that does not contain the Tika type system. Here, you could
>>> deselect the usage of prior type systems in the preferences: UIMA Cas
>>> Editor->Cas Editor Ide
>>>
>>
>> The MarkupAnnotations are not listed by the viewer.
>>
>> If I change the option and I select manually my type system (that imports
>> the MarkupAnnotator system), they still are not listed (attachment in this
>> email).
>>
>> <Pièce jointe Mail.png>
>>
>> For information, here is the content of my Test type system that I use to
>> visualize the content in the Annotation Browser:
>>
>> <?xml version="1.0" encoding="UTF-8"?><typeSystemDescription
>> xmlns="http://uima.apache.org/resourceSpecifier";>
>> <name>test.TestTypeSystem</name>
>> <imports>
>> <import location="../BasicTypeSystem.xml"/>
>> <import location="../tika/MarkupAnnotationTypeSystem.xml"/>
>> </imports>
>> <types>
>> <typeDescription>
>> <name>test.Test.Link</name>
>> <description>Type defined in test.Test</description>
>> <supertypeName>uima.tcas.Annotation</supertypeName>
>> </typeDescription>
>> </types>
>> </typeSystemDescription>
>>
>> Best regards,
>> Fouad
>>
>>
>>>
>>>> Therefore, I have two questions (maybe the answer is the same):
>>>>
>>>> - How can I visualize the annotations placed by the external engine ?
>>>
>>> Yes, that should be possible without any problems since that is only a
>>> functionality of the CAS Editor if the annotations have been added by
>>> the engine whether or not you can match with rules on them.
>>>
>>>> - How can I trigger the "MarkupAnnotation" condition within my script ?
>>>
>>> I haven't tried the Tika annotator for a while. I take a look at your
>>> example later this day and try to find the problem.
>>>
>>> Best,
>>>
>>> Peter
>>>
>>>
>>>> Thank you for your help.
>>>>
>>>> Best regards,
>>>> Fouad
>>>>
>>>> FYI: I put the MarkupAnnotation.xml and MarkupAnnotationTypeSystem.xml
>>>> files in my RUTA project and I reference the Tika libraries in my Run
>>>> configuration. I use the HEAD version of the Tika Annotator.
>>>
>>
>
