uima-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eddie Epstein <eaepst...@gmail.com>
Subject Re: Data read problem in UIMA-AS
Date Fri, 20 Aug 2010 22:20:37 GMT
Lokesh,

The scenario in figure 3 uses a collection reader at the driver, and
no collection reader in the remote service. Please try that.

regards,
Eddie


On Fri, Aug 20, 2010 at 9:21 AM, lokesh chanana
<lokesh.chanana@orkash.com> wrote:
> Hello Eddie,
>
> My situation is quite similar to fig 3 but i m using only one queue, one
> service and one runRemotAE instance.
>
> My deployment descriptor is as follows:
>
> /<?xml version="1.0" encoding="UTF-8"?>
> <analysisEngineDeploymentDescription
> xmlns="http://uima.apache.org/resourceSpecifier">
> <name>deploymentDescriptor</name>
> <description/>
> <version>1.0</version>
> <vendor/>
> <deployment protocol="jms" provider="activemq">
> <casPool numberOfCASes="1" initialFsHeapSize="2000000"/>
> <service>
> <inputQueue endpoint="sentenceQueue" brokerURL="${defaultBrokerURL}"
> prefetch="0"/>
> <topDescriptor>
> <import location="flowControllerDescriptor.xml"/>
> </topDescriptor>
> <analysisEngine async="false">
> <scaleout numberOfInstances="1"/>
> <asyncPrimitiveErrorConfiguration>
> <processCasErrors thresholdCount="0" thresholdWindow="0"
> thresholdAction="terminate"/>
> <collectionProcessCompleteErrors timeout="0"
> additionalErrorAction="terminate"/>
> </asyncPrimitiveErrorConfiguration>
> </analysisEngine>
> </service>
> </deployment>
> </analysisEngineDeploymentDescription>/
>
> And my flocontroller descriptor is..
>
> /<?xml version="1.0" encoding="UTF-8"?>
> <analysisEngineDescription xmlns="http://uima.apache.org/resourceSpecifier">
> <frameworkImplementation>org.apache.uima.java</frameworkImplementation>
> <primitive>false</primitive>
> <delegateAnalysisEngineSpecifiers>
> <delegateAnalysisEngine key="CollectionReader">
> <import
> location="../../examples/descriptors/collection_reader/FileSystemCollectionReader.xml"/>
> </delegateAnalysisEngine>
> <delegateAnalysisEngine key="Sentence">
> <import location="OpenNLPSentenceDetector.xml"/>
> </delegateAnalysisEngine>
> <delegateAnalysisEngine key="CasConsumer">
> <import
> location="../../examples/descriptors/cas_consumer/XmiWriterCasConsumer.xml"/>
> </delegateAnalysisEngine>
> </delegateAnalysisEngineSpecifiers>
> <analysisEngineMetaData>
> <name>sentence</name>
> <description>Implements a collection processing engine including collection
> reader,
>                 analysis engines and cas consumer.</description>
> <configurationParameters/>
> <configurationParameterSettings/>
> <flowConstraints>
> <fixedFlow>
> <node>CollectionReader</node>
> <node>Sentence</node>
> <node>CasConsumer</node>
> </fixedFlow>
> </flowConstraints>
> <capabilities>
> <capability>
> <inputs/>
> <outputs>
> <type
> allAnnotatorFeatures="true">org.apache.uima.examples.tokenizer.Sentence</type>
> </outputs>
> <languagesSupported>
> <language>en</language>
> </languagesSupported>
> </capability>
> </capabilities>
> <operationalProperties>
> <modifiesCas>true</modifiesCas>
> <multipleDeploymentAllowed>false</multipleDeploymentAllowed>
> <outputsNewCASes>false</outputsNewCASes>
> </operationalProperties>
> </analysisEngineMetaData>
> </analysisEngineDescription>/
>
> As i am quite new so kindly guide me where i m wrong...
>
> Regards
> Lokesh
>
>
>
> On Friday 20 August 2010 06:33 PM, Eddie Epstein wrote:
>>
>> Sorry, it is a bit hard to understand your scenario. Please clarify
>> using the examples
>> in http://uima.apache.org/doc-uimaas-what.html and identify which best
>> fits your situation,
>> Figure 3, 4 or 5.
>>
>> Thanks,
>> Eddie
>>
>> On Thu, Aug 19, 2010 at 6:12 AM, lokesh chanana
>> <lokesh.chanana@orkash.com>  wrote:
>>
>>>
>>> Hello,
>>>
>>> I am deploying UIMA-AS for testing perpose by now.
>>>
>>> My configurations includes
>>>
>>> !. Broker on one system<say Broker>
>>> 2. Service on one system<say Service>
>>> 3. Client on one system<say client>
>>>
>>> My configuration
>>> =>  my deployment descriptor calls the flowControlAggregate.
>>> =>  Flow control defines 3 delegates
>>>    1. collection Reader that reads data from
>>> /opt/apache-uima/examples/data
>>>    2. my core analysis Engine that process the data
>>>    3. XmiWriterCasConsumer.xml class to generete output.
>>>
>>> i deployed them as
>>> <On Broker>  startBroker.sh
>>>
>>> <On Service>  deployAsyncService.sh /Path/to/deployment/descriptor.xml
>>> -brokerURL tcp://<BROKER>:61616
>>>
>>> <on client>  runRemoteAsyncAE.sh tcp://<BROKER>:61616 MeetingFinderQueue
>>> -c
>>> /Path/to/collectin/reader.xml -o result.xmi
>>>
>>>
>>> NOW the problems i m concerned about are...
>>>
>>> If I don't define collection_Reader then i get no error but it says only
>>> one
>>> file is processed. (I knw that it is due to fact that client send an
>>> empty
>>> CAS). this means that client is sending the data in form of CAS. if so
>>> then
>>> y a cas collector if compulsaryin my flowControlAggregator???i tried
>>> emoving
>>> it but the service didn't started.
>>>
>>>    I found one answer to this that CAS at client is just sending the
>>> reference for CAS at SERVICE.
>>>    To test this i deleted the whole data at
>>> SERVICE(/opt/apache-uima/examples/data).
>>>    Unexpectedly my processing time got reduced to some 20 sec. (which was
>>> earlier 3200 sec. for 500 documents).
>>>
>>>    now i have two cases
>>> case 1: when the data at client and service is same.
>>>
>>>>
>>>> processing is too slow.
>>>> xmi file ganerated are of shoter length
>>>> the XMI files are different then those created by the basic UIMA
>>>> application
>>>> processing at SERVICE is too high
>>>> neglegible processing at client.
>>>>
>>>
>>> case 2: when the data there is no data at SERVICE end
>>>
>>>>
>>>> Processing is fast
>>>> XMI are same as those creaed by local uima.
>>>> no much proocessiong either on client or SERVICE end.
>>>>
>>>
>>> Now as things are not as i expected I am sure i m somewhere wrong
>>> conceptually. As i am quite new to UIMA-As any help is apprecable.
>>>
>>> Regards
>>> Lokesh
>>>
>>>
>>>
>
>

Mime
View raw message