Re: Running UIMA job on Hadoop with Behemoth

516 views
Skip to first unread message

Julien Nioche

unread,
Nov 28, 2012, 4:20:31 AM11/28/12
to digita...@googlegroups.com
Hi Jaipal

The tutorial was a bit behind and I have now fixed it. Thanks for pointing this out. See https://github.com/DigitalPebble/behemoth/wiki/behemothcompilation for instructions on how to compile Behemoth.

Thanks

Julien

On Wednesday, 28 November 2012 08:00:00 UTC, Jaipal Reddy wrote:
Hii Omar..

I am following this tutorial https://github.com/DigitalPebble/behemoth/wiki/tutorial  to configure the UIMA on Hadoop.. After launching the mvn , i am unale to find out the Build.xml and also */behemoth/modules file so i can't creat behemoth-.job , Could you help me out here 

On Friday, 21 September 2012 21:27:53 UTC+5:30, Omar wrote:

Hi,

I have a fully functional UIMA job that does simple annotation. I can successfully launch it via my local CAS GUI.

I have been trying to run the UIMA job on Hadoop using Apache Behemoth. I am wondering if  someone has worked on this? The job runs successfully but in the hadoop output directory; there is no output from the UIMA job. I can see in the Hadoop job tracker output that the job completed successfully and it copied its input data to the final output directory.

Can someone point me to what could be going on here?

Thanks

Omar

Jaipal Reddy

unread,
Nov 28, 2012, 4:27:27 AM11/28/12
to digita...@googlegroups.com

Hi  Julien 

I am trying to figure out but the thing is  https://github.com/DigitalPebble/behemoth/wiki/tutorial in this link mentioned that run "ant job" to run this we need build.xml..
If you fixed that please share me..


Thanks
Jaipal

DigitalPebble

unread,
Nov 28, 2012, 4:30:50 AM11/28/12
to digita...@googlegroups.com
The page https://github.com/DigitalPebble/behemoth/wiki/tutorial has been fixed following your comments. The other page I was refering to should help you generate the job files

--
You received this message because you are subscribed to the Google Groups "DigitalPebble" group.
To view this discussion on the web, visit https://groups.google.com/d/msg/digitalpebble/-/4tMISM9aVUoJ.

To post to this group, send an email to digita...@googlegroups.com.
To unsubscribe from this group, send email to digitalpebbl...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/digitalpebble?hl=en-GB.



--
 
Open Source Solutions for Text Engineering
 
http://digitalpebble.blogspot.com
http://www.digitalpebble.com

Jaipal R Pola

unread,
Nov 28, 2012, 5:09:50 AM11/28/12
to digita...@googlegroups.com
Hi  Julien,

Did you configured the UIMA with Hadoop, If yes i need some inputs  from you..

Thanks
Jaipal


Best Regards,
Jaipal R Pola
09494140102

DigitalPebble

unread,
Nov 28, 2012, 5:14:36 AM11/28/12
to digita...@googlegroups.com
Jaipal, 

I assume you mean "did you run the UIMA module with Behemoth" - the answer is obviously yes. Other people on the lists have too. Ask specific questions of you want to get useful answers

On 28 November 2012 10:09, Jaipal R Pola <jaipalr...@gmail.com> wrote:
Did you configured the UIMA with Hadoop, If yes i need some inputs  from you..



Jaipal R Pola

unread,
Nov 28, 2012, 5:18:38 AM11/28/12
to digita...@googlegroups.com
Julien
Ya Behmoth is used to run UIMA on Hadoop only right..



Best Regards,
Jaipal R Pola
09494140102




On 28 November 2012 15:44, DigitalPebble <jul...@digitalpebble.com> wrote:
sk specific questions of

DigitalPebble

unread,
Apr 16, 2013, 5:01:27 AM4/16/13
to digita...@googlegroups.com
Hi


The procedure is very similar for UIMA, first generate a job file for the UIMA module, then copy the pear file to HDFS

./bin/hadoop fs -copyFromLocal /mylocalpath/WhitespaceTokenizer.pear /apps/WhitespaceTokenizer.pear

This time the parameters in behemoth-site.xml to specify are

<property>
  <name>uima.annotations.filter</name>
  <value>org.apache.uima.TokenAnnotation,org.apache.uima.SentenceAnnotation</value>
  <description>Annotations types to consider when serializing to the behemoth format, separated by commas 
  </description>
</property>
<property>
  <name>uima.features.filter</name>
  <value>org.apache.uima.TokenAnnotation:posTag</value>
  <description>Feature names to consider when serializing to the behemoth format, separated by commas 
  </description>
</property>
./bin/hadoop jar ./behemoth-uima*job.jar com.digitalpebble.behemoth.uima.UIMADriver
 /data/behemothcorpus /data/behemothcorpus-2 /apps/WhitespaceTokenizer.pear
You should get the annotations created by your UIMA pipeline when specifying them with the params above

Julien

On 16 April 2013 09:14, <vigne...@gmail.com> wrote:

Hello,

I'm facing the exact issue.

The behemoth output simply contains the files that I gave as input. The job generated by behemoth is obviously running without any errors on hadoop. Has anyone figured out how to get the annotations?

--Vignesh

On Friday, 21 September 2012 21:27:53 UTC+5:30, Omar wrote:

Hi,

I have a fully functional UIMA job that does simple annotation. I can successfully launch it via my local CAS GUI.

I have been trying to run the UIMA job on Hadoop using Apache Behemoth. I am wondering if  someone has worked on this? The job runs successfully but in the hadoop output directory; there is no output from the UIMA job. I can see in the Hadoop job tracker output that the job completed successfully and it copied its input data to the final output directory.

Can someone point me to what could be going on here?

Thanks

Omar

--
You received this message because you are subscribed to the Google Groups "DigitalPebble" group.
To unsubscribe from this group and stop receiving emails from it, send an email to digitalpebbl...@googlegroups.com.

To post to this group, send an email to digita...@googlegroups.com.

DigitalPebble

unread,
Apr 17, 2013, 3:52:52 PM4/17/13
to digita...@googlegroups.com
Hi Vignesh

Currently the script relies exclusively on the content of the job.jar file but ideally it should detect whether something like '-conf XXX' has been passed as part of the arguments and/or rely on an environment variable such as $BEHEMOTH_CONF.

Could you please open an issue on Github (and attach a patch if you can)?

Thanks

Julien 

On 17 April 2013 16:49, <vigne...@gmail.com> wrote:
Thanks Julien,

I'm getting the annotations properly now. The problem was, behemoth wasn't picking up the config file correctly. I edited the behemoth script file to point to the correct conf directory and its working fine now.

-- Vignesh

vka...@gmail.com

unread,
Jul 19, 2013, 7:39:08 PM7/19/13
to digita...@googlegroups.com, jul...@digitalpebble.com
Vignesh,
Could you elaborate a little bit on what exactly you changed? When I use corpus reader, I essentially see the url or the hdfs path to the input file.

Would appreciate your comments!
Vinod

On Wednesday, April 17, 2013 10:49:28 AM UTC-5, Vignesh wrote:
Thanks Julien,

I'm getting the annotations properly now. The problem was, behemoth wasn't picking up the config file correctly. I edited the behemoth script file to point to the correct conf directory and its working fine now.

-- Vignesh

On Tuesday, 16 April 2013 14:31:27 UTC+5:30, DigitalPebble wrote:
Reply all
Reply to author
Forward
0 new messages