You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@uima.apache.org by "Chavan, Girish" <ch...@upmc.edu> on 2009/07/20 18:21:59 UTC
Determine which AE created an annotation
Is there a way to determine which Analysis Engine/ component in an
aggregate analysis engine created a particular annotation?
_______________
Girish Chavan, MSIS
Department of Biomedical Informatics (DBMI)
University of Pittsburgh
UPMC Cancer Pavilion, 302D
5150 Centre Avenue
Pittsburgh, PA 15232
Office: 412-623-4084
Email: chavang@upmc.edu
Re: Determine which AE created an annotation
Posted by Thilo Goetz <tw...@gmx.de>.
Chavan, Girish wrote:
> Is there a way to determine which Analysis Engine/ component in an
> aggregate analysis engine created a particular annotation?
Not by default. Some people add a special "source" field
or similar to their annotations and have their annotators
record who created which annotation. However, this is up
to the person who wrote the annotator.
--Thilo
>
>
>
> _______________
>
> Girish Chavan, MSIS
>
> Department of Biomedical Informatics (DBMI)
> University of Pittsburgh
> UPMC Cancer Pavilion, 302D
> 5150 Centre Avenue
> Pittsburgh, PA 15232
>
> Office: 412-623-4084
>
> Email: chavang@upmc.edu
>
>
>
>
Re: Determine which AE created an annotation
Posted by Marshall Schor <ms...@schor.com>.
There is no "built-in" way to do this. But many people write their
annotators to include an additional feature which identifies which
annotator set a particular value.
-Marshall
Chavan, Girish wrote:
> Is there a way to determine which Analysis Engine/ component in an
> aggregate analysis engine created a particular annotation?
>
>
>
> _______________
>
> Girish Chavan, MSIS
>
> Department of Biomedical Informatics (DBMI)
> University of Pittsburgh
> UPMC Cancer Pavilion, 302D
> 5150 Centre Avenue
> Pittsburgh, PA 15232
>
> Office: 412-623-4084
>
> Email: chavang@upmc.edu
>
>
>
>
>
Re: Determine which AE created an annotation
Posted by Thilo Goetz <tw...@gmx.de>.
Burn Lewis wrote:
> Only if your components are designed to add an "author" feature to each
> annotation ... its not part of the standard Annotation.
>
> - Burn.
>
Didn't see your reply there at first...
--Thilo
Re: Determine which AE created an annotation
Posted by Burn Lewis <bu...@gmail.com>.
Only if your components are designed to add an "author" feature to each
annotation ... its not part of the standard Annotation.
- Burn.
Re: drools for writing grammars
Posted by Peter Klügl <pk...@ki.informatik.uni-wuerzburg.de>.
Hallo,
>
> Does textmarker have defined BNF or EBNF grammar, so we could use ANTLR for building AST?
>
> Thanks for any opinion/comment.
>
>
I'm not quite sure what you mean. Please correct me if I don't
understand you correctly.
The wiki (http://tmwiki.informatik.uni-wuerzburg.de) contains two BNFs
(or similar definition): The current/old grammar will be replaced with
the new, more comprehensible grammar next month. We are using ANTLR
ourself for building an AST of the DLTK framework and for parsing the
inference objects. But both grammars aren't really reusable for other
purposes since they contain some java code and are targets of current
developments.
Peter
Re: drools for writing grammars
Posted by Mohamed <mo...@yahoo.com>.
Hi all,
we are still in evaluation phase, so I am wondering, if the described scenario makes any sense to you. I am not much into NLP area, so, expert view would be very helpful.
Also, drools support for adding dialects, makes me think if something can be done with textmarker. Drools people did that for clips as described here:
http://blog.athico.com/search/label/Clips
Does textmarker have defined BNF or EBNF grammar, so we could use ANTLR for building AST?
Thanks for any opinion/comment.
Kind regards,
Mohamed
--- On Tue, 7/21/09, Mohamed <mo...@yahoo.com> wrote:
From: Mohamed <mo...@yahoo.com>
Subject: Re: drools for writing grammars
To: uima-user@incubator.apache.org
Date: Tuesday, July 21, 2009, 3:16 PM
Hi,
we are buiding knoledge management platform with the aim to support both GATE and UIMA type system. Our main concern is not to provide interoperability between two models, but to provide user the oportunity to select any of the two and upload rule files that will work with either UIMA or GATE, depending on their type (package).
Also, the idea is that uploaded rule is compiled on fly and inserted into memory making it
immediately available for processing.
That is why drools seems interesting to me. Also, if it is somehow possible to use its LHS expressions to work with both UIMA and GATE type model, and allow arbitrary Java code on RHS, with great authoring support that it provides out of box, it makes it very suitable for our client, since it is only one rule language to learn (and support).
Using textmarker for uima, and jape for gate leaves us with a problem how to unify two authoring approaches, support both languages, etc.
kind regards,
Mohamed
--- On Tue, 7/21/09, Roberto Franchini <ro...@gmail.com> wrote:
From: Roberto Franchini <ro...@gmail.com>
Subject: Re: drools for writing grammars
To: uima-user@incubator.apache.org
Date: Tuesday, July 21, 2009, 1:25 PM
On Tue, Jul 21, 2009 at 1:00 PM, Peter
Klügl<pk...@ki.informatik.uni-wuerzburg.de> wrote:
[cut]
>
> The implementation of JAPE should be faster than TextMarker, but I don't
> know how fast the bridge is. In contrast to JAPE, TextMarker is still in an
> early project phase. JAPE is well documented and tested, TextMarker not
> (yet).
>
I wrote my own bridge for JAPE since the one in GATE 4.0 was
compatible only with IBM-UIMA.
It's a custom, very custom, ad targeted to our type sytem.
This bridge map to JAPE and back to UIMA a lot of annotations, so
there are a lot of objects to be garbage collected for each document.
For comparison a 3-thread-pipeline with jape-uima-bridge inside
poresses 80/90000 docs per hour. My docs are small: < 5kb.
The same pipeline without jape goes to 140k doc/h.
Cheers,
R.
--
Roberto Franchini
http://www.celi.it
http://www.blogmeter.it
http://www.memesphere.it
Tel +39-011-6600814
jabber:ro.franchini@gmail.com skype:ro.franchini
Re: drools for writing grammars
Posted by Mohamed <mo...@yahoo.com>.
Hi,
we are buiding knoledge management platform with the aim to support both GATE and UIMA type system. Our main concern is not to provide interoperability between two models, but to provide user the oportunity to select any of the two and upload rule files that will work with either UIMA or GATE, depending on their type (package).
Also, the idea is that uploaded rule is compiled on fly and inserted into memory making it
immediately available for processing.
That is why drools seems interesting to me. Also, if it is somehow possible to use its LHS expressions to work with both UIMA and GATE type model, and allow arbitrary Java code on RHS, with great authoring support that it provides out of box, it makes it very suitable for our client, since it is only one rule language to learn (and support).
Using textmarker for uima, and jape for gate leaves us with a problem how to unify two authoring approaches, support both languages, etc.
kind regards,
Mohamed
--- On Tue, 7/21/09, Roberto Franchini <ro...@gmail.com> wrote:
From: Roberto Franchini <ro...@gmail.com>
Subject: Re: drools for writing grammars
To: uima-user@incubator.apache.org
Date: Tuesday, July 21, 2009, 1:25 PM
On Tue, Jul 21, 2009 at 1:00 PM, Peter
Klügl<pk...@ki.informatik.uni-wuerzburg.de> wrote:
[cut]
>
> The implementation of JAPE should be faster than TextMarker, but I don't
> know how fast the bridge is. In contrast to JAPE, TextMarker is still in an
> early project phase. JAPE is well documented and tested, TextMarker not
> (yet).
>
I wrote my own bridge for JAPE since the one in GATE 4.0 was
compatible only with IBM-UIMA.
It's a custom, very custom, ad targeted to our type sytem.
This bridge map to JAPE and back to UIMA a lot of annotations, so
there are a lot of objects to be garbage collected for each document.
For comparison a 3-thread-pipeline with jape-uima-bridge inside
poresses 80/90000 docs per hour. My docs are small: < 5kb.
The same pipeline without jape goes to 140k doc/h.
Cheers,
R.
--
Roberto Franchini
http://www.celi.it
http://www.blogmeter.it
http://www.memesphere.it
Tel +39-011-6600814
jabber:ro.franchini@gmail.com skype:ro.franchini
Re: drools for writing grammars
Posted by Roberto Franchini <ro...@gmail.com>.
On Tue, Jul 21, 2009 at 1:00 PM, Peter
Klügl<pk...@ki.informatik.uni-wuerzburg.de> wrote:
[cut]
>
> The implementation of JAPE should be faster than TextMarker, but I don't
> know how fast the bridge is. In contrast to JAPE, TextMarker is still in an
> early project phase. JAPE is well documented and tested, TextMarker not
> (yet).
>
I wrote my own bridge for JAPE since the one in GATE 4.0 was
compatible only with IBM-UIMA.
It's a custom, very custom, ad targeted to our type sytem.
This bridge map to JAPE and back to UIMA a lot of annotations, so
there are a lot of objects to be garbage collected for each document.
For comparison a 3-thread-pipeline with jape-uima-bridge inside
poresses 80/90000 docs per hour. My docs are small: < 5kb.
The same pipeline without jape goes to 140k doc/h.
Cheers,
R.
--
Roberto Franchini
http://www.celi.it
http://www.blogmeter.it
http://www.memesphere.it
Tel +39-011-6600814
jabber:ro.franchini@gmail.com skype:ro.franchini
Re: drools for writing grammars
Posted by Peter Klügl <pk...@ki.informatik.uni-wuerzburg.de>.
Hallo,
>
>
> Yesterday was too late for my brain :)
> Well. textmarker (http://textmarker.sourceforge.net/) seems very
> interesting to me.
> Is it comparable with jape?
> R
A short comparison (please correct me if I am wrong):
JAPE and TextMarker are rule-based tools that process patterns of
annotations. JAPE is able to directly include Java code and is built on
GATE (a UIMA bridge is available). TextMarker uses an imperative
processing of the rule set, provides scripting elements and an extension
mechanism for third-party condition, actions and functions (for numbers,
strings, booleans and types) and is directly built on UIMA. Analysis
Engines and Type System are created by a build process and arbitrary
Analysis Engine can be executed by rules.
We think that a good development environment is essential for the
usability of a knowledge engineering tool and are currently working on a
context-sensitive auto-completion, a formatter, a test-driven
development approach and implementations of prominent rule learning
methods. I must admit that I don't know if there is an engineering
support for JAPE.
The implementation of JAPE should be faster than TextMarker, but I don't
know how fast the bridge is. In contrast to JAPE, TextMarker is still in
an early project phase. JAPE is well documented and tested, TextMarker
not (yet).
Peter
--
Peter Klügl
pkluegl@uni-wuerzburg.de
Re: drools for writing grammars
Posted by Roberto Franchini <ro...@gmail.com>.
On Mon, Jul 20, 2009 at 11:20 PM, Roberto
Franchini<ro...@gmail.com> wrote:
> On Mon, Jul 20, 2009 at 7:35 PM, Mohamed Riyaz<mo...@yahoo.com> wrote:
>> Hi,
>> I would like to use drools for writing grammars in UIMA, similar to JAPE in gate.
>> While browsing uima list, I found one post with the same idea.
>> Did somebody perform any research and had any result on that subject?
>>
>
> I've got the same idea some time ago, but unfortunally we didn't have
> time to spend on this.
> Another way is to write an ANTLR parser for a custum grammar-syntax.
> Maybe we can open a project some where (google-code, sf...) just to
> write a proof of concept.
> Let me know.
Yesterday was too late for my brain :)
Well. textmarker (http://textmarker.sourceforge.net/) seems very
interesting to me.
Is it comparable with jape?
R.
--
Roberto Franchini
http://www.celi.it
http://www.blogmeter.it
http://www.memesphere.it
Tel +39-011-6600814
jabber:ro.franchini@gmail.com skype:ro.franchini
Re: drools for writing grammars
Posted by Roberto Franchini <ro...@gmail.com>.
On Mon, Jul 20, 2009 at 7:35 PM, Mohamed Riyaz<mo...@yahoo.com> wrote:
> Hi,
> I would like to use drools for writing grammars in UIMA, similar to JAPE in gate.
> While browsing uima list, I found one post with the same idea.
> Did somebody perform any research and had any result on that subject?
>
I've got the same idea some time ago, but unfortunally we didn't have
time to spend on this.
Another way is to write an ANTLR parser for a custum grammar-syntax.
Maybe we can open a project some where (google-code, sf...) just to
write a proof of concept.
Let me know.
R.
--
Roberto Franchini
http://www.celi.it
http://www.blogmeter.it
http://www.memesphere.it
Tel +39-011-6600814
jabber:ro.franchini@gmail.com skype:ro.franchini
drools for writing grammars
Posted by Mohamed Riyaz <mo...@yahoo.com>.
Hi,
I would like to use drools for writing grammars in UIMA, similar to JAPE in gate.
While browsing uima list, I found one post with the same idea.
Did somebody perform any research and had any result on that subject?
Kind regards,
Mohamed
--- On Mon, 7/20/09, Chavan, Girish <ch...@upmc.edu> wrote:
From: Chavan, Girish <ch...@upmc.edu>
Subject: Determine which AE created an annotation
To: uima-user@incubator.apache.org
Date: Monday, July 20, 2009, 4:21 PM
Is there a way to determine which Analysis Engine/ component in an
aggregate analysis engine created a particular annotation?
_______________
Girish Chavan, MSIS
Department of Biomedical Informatics (DBMI)
University of Pittsburgh
UPMC Cancer Pavilion, 302D
5150 Centre Avenue
Pittsburgh, PA 15232
Office: 412-623-4084
Email: chavang@upmc.edu