headers
Thilo Goetz | 2 Oct 13:57
Picon
Picon

CFP open for ApacheCon Europe 2009

If you only have thirty seconds:

The Call for Papers for ApacheCon Europe 2009, to be held in
Amsterdam, from 23rd to 27th March, is now open! Submit your proposals
at http://eu.apachecon.com/c/aceu2009/cfp/ before 24th October.

Remember that early bird prices for ApacheCon US 2008, to be held in
New Orleans, from 3rd to 7th November, will go up this Friday, at
midnight Eastern time!

Sponsorship opportunities for ApacheCon US 2008 and ApacheCon EU 2009
are still available. If you or your company are interested in becoming
a sponsor, please contact Delia Frees at delia@... for
details.

*******

If you want all the details:

ApacheCon Europe 2009 - Leading the Wave of Open Source
Amsterdam, The Netherlands
23rd to 27th March, 2009

Call for Papers Opens for ApacheCon Europe 2009

The Apache Software Foundation (ASF) invites submissions to its
official conference, ApacheCon Europe 2009. To be held 23rd to 27th
March, 2009 at the Mövenpick Hotel Amsterdam City Centre, ApacheCon
serves as a forum for showcasing the ASF's latest developments,
including its projects, membership, and community. ApacheCon offers
(Continue reading)

Permalink | Reply | 0 comments |
headers
Thilo Goetz | 2 Oct 14:03
Picon
Picon

Re: [UIMA] Annotated corpus

Hi Manuel,

how about storing the individual documents in XCAS or XMI
format?  That's probably the simplest and best supported
way, though not very compact.

--Thilo

Manuel Fiorelli wrote:
> Hi,
> I know that the UIMA architecture provides the CAS (common analysis system)
> to share analysis data about a single artifact. Is there a standard way to
> store an annotated corpus, which could be used, for example, to train an AE?
> 
> Regards,
> Manuel Fiorelli
>
Permalink | Reply | 1 comment |
headers
Manuel Fiorelli | 2 Oct 14:26
Picon

Re: [UIMA] Annotated corpus

Thank you for the answer. Since I am new to UIMA, I wanted to verify that
there isn't a better solution.

Manuel Fiorelli
Permalink | Reply | 0 comments |
headers
Manuel Fiorelli | 2 Oct 14:51
Picon

[UIMA] How to remove a feature structure from a CAS

Hello,
I am developing an interactive annotator, which should produce a corpus of
annotated documents for AE training. Since a human annotator can decide to
remove an annotation, I would like to know if there were an API for remove a
feature structure from the CAS after it was created.

Manuel Fiorelli
Permalink | Reply | 3 comments |
headers
Thilo Goetz | 2 Oct 15:01
Picon
Picon

Re: [UIMA] How to remove a feature structure from a CAS

Manuel Fiorelli wrote:
> Hello,
> I am developing an interactive annotator, which should produce a corpus of
> annotated documents for AE training. Since a human annotator can decide to
> remove an annotation, I would like to know if there were an API for remove a
> feature structure from the CAS after it was created.
> 
> Manuel Fiorelli
> 

Sort of.  You can remove the annotation from the index.  The
annotation will continue to exist for the lifetime of the
CAS, but it will no longer be accessible.  If you then save
the CAS to disk, the deleted annotation will not be saved.

If you use the JCas, you can do removeFromIndexes() on the
annotation itself.  If using the base CAS, you need to use
cas.removeFsFromIndexes(annotation) -- of course this will
work with the JCas as well.

--Thilo
Permalink | Reply | 2 comments |
headers
Manuel Fiorelli | 2 Oct 15:11
Picon

Re: [UIMA] How to remove a feature structure from a CAS

Thanks again,
if I understand well, the best we can do is to prevent the feature structure
from being saved. I think I cannot use this strategy since it can leak
memory (until one saves and reload).
A little question again, after I create a FS, is it added to the default
unnamed bag index? (or I have to add it manually)

Manuel Fiorelli
Permalink | Reply | 1 comment |
headers
Thilo Goetz | 2 Oct 16:14
Picon
Picon

Re: [UIMA] How to remove a feature structure from a CAS

Manuel Fiorelli wrote:
> Thanks again,
> if I understand well, the best we can do is to prevent the feature structure
> from being saved. I think I cannot use this strategy since it can leak
> memory (until one saves and reload).

True.  If you have a potentially long-running interactive
application, you need to consider that issue.  What you
can do is once in a while copy your current content into
a clean CAS, using the CasCopier.  Kinda clumsy...

> A little question again, after I create a FS, is it added to the default
> unnamed bag index? (or I have to add it manually)

You have to call addToIndexes().  After that, it's both
on the annotation and the default bag index.

> 
> Manuel Fiorelli
>
Permalink | Reply | 0 comments |
headers
Aaron Kaplan | 2 Oct 21:48
Picon
Favicon

Order of CASes after multiplier

I can't find this documented anywhere, but I observe empirically that
components downstream of a CAS multiplier receive all of the new CASes
generated by that multiplier before receiving the original CAS that was
the input to the multiplier.  Can I rely on that order?

-Aaron
Permalink | Reply | 1 comment |
headers
Eddie Epstein | 3 Oct 00:38
Picon

Re: Order of CASes after multiplier

Hi Aaron,

Yes, for a synchronous UIMA aggregate the parent CAS is released only after
all its children are released.

When deployed asynchronously in UIMA AS, the parent CAS is released from the
CM when hasNext returns false, and is then available to be routed. If the
parent CAS comes into an async aggregate on a process call (that it, it is
not itself generated by a CM inside), it will not be returned until all of
its children CAS have been released; this allows an error on a child CAS to
[optionally] generate an error on its parent CAS.

Eddie

On Thu, Oct 2, 2008 at 3:48 PM, Aaron Kaplan <aaron.kaplan@...>wrote:

> I can't find this documented anywhere, but I observe empirically that
> components downstream of a CAS multiplier receive all of the new CASes
> generated by that multiplier before receiving the original CAS that was
> the input to the multiplier.  Can I rely on that order?
>
> -Aaron
>
Permalink | Reply | 0 comments |
headers
Charles Proefrock | 6 Oct 23:59
Picon

UimaASProcessCasTimeout exception hangs application


When experimenting with the UIMA-AS examples and how the error handling mechanisms work in terms of
timeouts due to AS Aggregates taking too long or going offline, we came across a situation in which the
RunRemoteAsyncAE hangs and never returns after receiving and processing a UimaASProcessCasTimeout
exception.  Ultimately, we simply want the system to recover and try the next CAS or return gracefully
without having to call a hard System.exit(1).

Our tests are based on the Deploy_MeetingDetectorTAE_RemoteRoomNumber.xml example.  All we did was the following:
(1)   added “Thread.sleep(5000);” to RoomNumberAnnotator.java process to simulate  a longer process.
(2)   Executed: startBroker.bat
(3)   Executed: deployAsyncService <DIR>\Deploy_RoomNumberAnnotator.xml
(4)   Executed: runRemoteAsyncAE tcp://localhost:61616 MeetingDetectorTaeQueue –c
<DIR>\FileSystemCollectionReader.xml -t 4 -i

In summary: when setting and triggering the time outs on the RoomNumber remoteAnalysisEngine, the
exception is thrown and caught, each additional CAS is tried, and the system exists gracefully.  If the
UimaAsynchronousEngine within the runRemoteAsyncAE times out, then the system hangs and never
returns. 

Is there a trick that we’re missing? Is this expected?

- Charles

 
_________________________________________________________________
See how Windows Mobile brings your life together—at home, work, or on the go.
http://clk.atdmt.com/MRT/go/msnnkwxp1020093182mrt/direct/01/
Permalink | Reply | 4 comments |

Gmane