uima-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sumit Madan <sumit.ma...@scai.fraunhofer.de>
Subject Re: UIMA Ruta use of features in a block statement
Date Wed, 12 Jul 2017 16:35:52 GMT
Am 06.07.2017 um 18:00 schrieb Peter Klügl:
> Hi,

Hi Peter,

> I cannot reproduce the problem. So you have maybe a small runnable test?

I've created a small runnable test with DKPro (1.8.0) and Ruta (2.6.0). 
I've to say that I'm also unable to reproduce the problem. All block 
constructs are working just fine in the runnable test.

For the documents that we process in our internal pipeline, which is 
mostly identical to the runnable test, the block constructs on Lemma 
type are not working. The pipeline uses our own internal type system. 
I've tried to do some debugging but not able to find the problem. A 
characteristic of our pipeline is that we work on a different view than 
_InitialView. I've also considered this aspect in the runnable test.

As I cannot provide you the source code of the internal pipeline, we've 
created a minimal serialized CAS with debug types. May be you can 
identify the problem in the CAS? (The serialized CAS has been sent to 
your private email address.)

The applied rules:
###
BLOCK(forEACH) Lemma{}{
       Lemma{->MATCHEDTEXT(s), ASSIGN(a,contains(s,"er"))};
       Lemma{a ->Value1};
}
###

Another small issue: DebugRuleApply.element doesn't always match with 
the original rules. For example, for the second rule, it contains the 
value "Lemma{-&gt;MATCHEDTEXT(s, ),ASSIGN(a, contains(s))};", which is 
not identical to the user input.

>
> Best,
>
>
> Peter
>
>
>
> Am 31.05.2017 um 14:18 schrieb Sumit Madan:
>> Am 31.05.2017 um 13:26 schrieb Sumit Madan:
>>> Am 31.05.2017 um 09:59 schrieb Peter Klügl:
>>>> Hi,
>>> Hi Peter,
>>>
>>>> Am 30.05.2017 um 19:35 schrieb Sumit Madan:
>>>>> We were not able to access the annotation within the BLOCK.
>>>>>
>>>>> ###
>>>>> STRING s;
>>>>> BOOLEAN a ;
>>>>>
>>>>> // This is not working for us:
>>>>> BLOCK(forEACH) Lemma{}{
>>>>>       Lemma{->MATCHEDTEXT(s), ASSIGN(a,contains(s,"er"))};
>>>>>       Lemma{a ->Test1};
>>>>> }
>>>>>
>>>>> // This is working:
>>>>> BLOCK(forEACH) Lemma{}{
>>>>>       W{->MATCHEDTEXT(s), ASSIGN(a,contains(s,"er"))};
>>>>>       W{a ->Test2};
>>>>> }
>>>>>
>>>>> // This is also working:
>>>>> BLOCK(forEACH) Lemma{}{
>>>>> Document{->MATCHEDTEXT(s), ASSIGN(a,contains(s,"er"))};
>>>>> Document{a ->Test3};
>>>>> }
>>>>> ###
>>>>>
>>>> All three examples should work. Which ruta version do you use? It looks
>>>> like a bug.
>>> Lisa is testing again and will get back to you.
>> We tested it again and can confirm the following:
>>
>> ###
>> STRING s;
>> BOOLEAN a ;
>>
>> // This is not working:
>> BLOCK(forEACH1) Lemma{} { // Matches a lot
>>       Lemma{->MATCHEDTEXT(s), ASSIGN(a,contains(s,"er"))}; // Matches 0
>>       Lemma{a ->Test1};
>> }
>>
>> // This is not working:
>> BLOCK(forEACH2) Lemma{} { // Matches a lot
>> Lemma{contains(Lemma.ct, "er") -> MARK(Test2)}; // Matches 0
>> }
>>
>> // This is working:
>> Lemma{contains(Lemma.ct, "er") -> MARK(Test3)}; // Matches 1
>>
>> // This is working:
>> BLOCK(forEACH4) Lemma{}{
>>       W{->MATCHEDTEXT(s), ASSIGN(a,contains(s,"er"))}; // Matches 1
>>       W{a ->Test4};
>> }
>>
>> // This is also working:
>> BLOCK(forEACH5) Lemma{}{
>>      Document{->MATCHEDTEXT(s), ASSIGN(a,contains(s,"er"))}; // Matches 1
>>      Document{a ->Test5};
>> }
>> ###
>>
>> Is it possible that in a BLOCK only Ruta types are available?
>>
>> We are using the latest stable version 2.6.0.
>>


-- 
Mit freundlichem Gruß / Best regards

Sumit Madan, M.Sc.

Wissenschaftlicher Mitarbeiter / Research fellow

Fraunhofer-Institute for Algorithms and Scientific Computing (SCAI)
Department of Bioinformatics
Schloss Birlinghoven
D-53754 Sankt Augustin

Room: C3-242
Tel.: +49 2241 14 2997
Email:sumit.madan@scai.fraunhofer.de
Internet:http://www.scai.fraunhofer.de/


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