uima-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Silvestre Losada <silvestre.los...@gmail.com>
Subject Re: Ruta partofneq
Date Fri, 27 Feb 2015 13:40:09 GMT
Hi,

I submited a patch with a solution for the error reported
https://issues.apache.org/jira/browse/UIMA-4261

Best

On 25 February 2015 at 10:41, Silvestre Losada <silvestre.losada@gmail.com>
wrote:

> Done
>
> https://issues.apache.org/jira/browse/UIMA-4261
>
>
>
> On 24 February 2015 at 19:35, Peter Klügl <pkluegl@uni-wuerzburg.de>
> wrote:
>
>> Hi,
>>
>> could you open an issue and attach it there? That would be great.
>>
>> Best,
>>
>> Peter
>>
>> Am 24.02.2015 um 18:13 schrieb Silvestre Losada:
>>
>>  Hi Peter,
>>>
>>> The problem happens if the Annotations are created by external analysis
>>> engine, using something like this
>>>
>>> ENGINE TestAE;
>>> Document{-> EXEC(TestAE,{TestType})};
>>> (TestType{-> UNMARKALL(TestType)}){PARTOFNEQ(TestType)};
>>>
>>> It seems that  ExecAction is removing the Type form RutaBasic.partOf
>>> array
>>> at some point of the execution. After that in PartOfNeqCondition in
>>> method
>>> check next condition is always returning false because the Type was
>>> removed
>>> previously from ruta basic.
>>>
>>>      boolean partOf = beginAnchor.isPartOf(t) || endAnchor.isPartOf(t);
>>>      if (!partOf) {
>>>        return false;
>>>      }
>>>
>>> I have uima ruta test project that reproduces the error I can send to you
>>> in zip file.
>>>
>>> Best.
>>>
>>> On 20 February 2015 at 20:41, Peter Klügl <pkluegl@uni-wuerzburg.de>
>>> wrote:
>>>
>>>  Hi,
>>>>
>>>> hmmm, that's strange. When I apply the rules on the document "A B C D",
>>>> only one T1 annotation remains.
>>>>
>>>> On which document did you test the rules?
>>>>
>>>> Best,
>>>>
>>>> Peter
>>>>
>>>> Am 20.02.2015 um 09:17 schrieb Silvestre Losada:
>>>>
>>>>   HI again,
>>>>
>>>>> Now I'm running into this problem
>>>>> DECLARE T1;
>>>>> "A B C D" -> T1;
>>>>> "B" -> T1;
>>>>> "C D" -> T1;
>>>>> "D" -> T1;
>>>>>
>>>>> (T1{-> UNMARKALL(T1)}){PARTOFNEQ(T1)};
>>>>>
>>>>> The ouput is
>>>>> "A B C D" -> T1;
>>>>> "D" -> T1;
>>>>>
>>>>> I suspect that this is because D is part of "A B C D" and "C D"
>>>>>
>>>>> Im using lastest version in trunk.
>>>>>
>>>>>
>>>>> Kind regards
>>>>>
>>>>> On 14 February 2015 at 09:29, Silvestre Losada <
>>>>> silvestre.losada@gmail.com>
>>>>> wrote:
>>>>>
>>>>>   Thanks Peter,
>>>>>
>>>>>>
>>>>>> It seems to work.
>>>>>>
>>>>>> On 13 February 2015 at 22:18, Peter Klügl <pkluegl@uni-wuerzburg.de>
>>>>>> wrote:
>>>>>>
>>>>>>   This should work just fine and should remove both contained
>>>>>> annotations.
>>>>>>
>>>>>>> I just tested it with the current trunk and the following script:
>>>>>>>
>>>>>>> DECLARE T1;
>>>>>>> "A B" -> T1;
>>>>>>> "B" -> T1;
>>>>>>> "B" -> T1;
>>>>>>>
>>>>>>> (T1{-> UNMARK(T1)}){PARTOFNEQ(T1)};
>>>>>>>
>>>>>>> If applied on the test "A B", only the largest annotation remains.
>>>>>>>
>>>>>>> Can you test it with the current trunk in case I fixed the bug
a few
>>>>>>> minutes ago by accident. If not, can you give me more information
>>>>>>> about
>>>>>>> the
>>>>>>> context of your rule?
>>>>>>>
>>>>>>> Best,
>>>>>>>
>>>>>>> Peter
>>>>>>>
>>>>>>> Am 12.02.2015 um 10:12 schrieb Silvestre Losada:
>>>>>>>
>>>>>>>    I dont know if this is a bug or if it is wokring well. I have
the
>>>>>>>
>>>>>>>  following
>>>>>>>> annotations.
>>>>>>>>
>>>>>>>> AnnotationA
>>>>>>>>         begin:0
>>>>>>>>         ends:8
>>>>>>>>          id:1
>>>>>>>> AnnotationA
>>>>>>>>         begin:4
>>>>>>>>         ends:8
>>>>>>>>          id:2
>>>>>>>> AnnotationA
>>>>>>>>         begin: 4
>>>>>>>>         ends:8
>>>>>>>>          id:3
>>>>>>>>
>>>>>>>> Then if apply the following ruta
>>>>>>>>
>>>>>>>> (AnnotationA{-> UNMARK(AnnotationA)}){PARTOFNEQ(AnnotationA)};
>>>>>>>>
>>>>>>>> The output is
>>>>>>>> AnnotationA
>>>>>>>>         begin:0
>>>>>>>>         ends:8
>>>>>>>>          id:1
>>>>>>>> AnnotationA
>>>>>>>>         begin: 4
>>>>>>>>         ends:8
>>>>>>>>          id:3
>>>>>>>>
>>>>>>>> I expect that annotations with id 2 and 3 will be removed.
Is there
>>>>>>>> any
>>>>>>>> way
>>>>>>>> to remove both
>>>>>>>>
>>>>>>>> Kind regards
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message