blob: 12de6b963f370093104b5b0dee49ce921e5ac583 [file] [log] [blame]
<?xml version="1.0" encoding="UTF-8"?>
<pgeConfig>
<!-- use file staging -->
<fileStaging dir="[JobInputDir]" force="[ForceStaging]">
<stageFiles metadataKey="InputFiles" />
</fileStaging>
<!-- How to run the PGE -->
<exe dir="[JobDir]" shell="/bin/bash">
<cmd>export PATH=$HOME/bin/:${PATH}</cmd>
<cmd>shopt -s expand_aliases</cmd>
<cmd>alias rat="java -jar [DRAT_HOME]/rat/lib/apache-rat-0.13.jar"</cmd>
<cmd>echo "Creating working dirs"</cmd>
<cmd>mkdir -p [JobInputDir] ; mkdir -p [JobOutputDir]; mkdir -p [JobLogDir]</cmd>
<!-- <cmd>echo "Staging input to [JobInputDir]"</cmd>
<cmd>FILELIST='[InputFiles]'</cmd>
<cmd>FSEP="_|_"</cmd>
<cmd>bash [DRAT_HOME]/pge/bin/rat_audit/copy_files.sh ${FILELIST} "${FSEP}" "[JobInputDir]"</cmd>-->
<cmd>echo "Running Apache RAT on [JobInputDir]"</cmd>
<cmd>rat [JobInputDir] > [JobOutputDir]/rat_[MimeType]_[DateMilis].log</cmd>
</exe>
<!-- Files to ingest -->
<output>
<!-- one or more of these -->
<dir path="[JobOutputDir]" createBeforeExe="false">
<files regExp=".*\.log" metFileWriterClass="org.apache.oodt.cas.pge.writers.metlist.MetadataListPcsMetFileWriter" args="[DRAT_HOME]/pge/policy/metout/rat_log_metout.xml"/>
<files regExp=".*\.log" metFileWriterClass="org.apache.oodt.cas.pge.writers.FilenameExtractorWriter" args="[DRAT_HOME]/pge/extractors/filename/rat_log_filename_extractor.xml"/>
</dir>
</output>
<!-- Custom metadata to add to output files -->
<customMetadata>
<!-- automatically perform file staging -->
<metadata key="ForceStaging" val="true"/>
<metadata key="PGETask_FileStager" val="org.apache.oodt.cas.pge.staging.HashingOrigFileStager"/>
<!-- helpful keys -->
<metadata key="LessThan" val="&#x3C;"/>
<metadata key="LessThanOrEqualTo" val="[LessThan]="/>
<metadata key="GreaterThan" val="&#x3E;"/>
<metadata key="GreaterThanOrEqualTo" val="[GreaterThan]="/>
<metadata key="Exclamation" val="&#33;"/>
<metadata key="Ampersand" val="&#38;"/>
<metadata key="NotEqualTo" val="[Ampersand]="/>
<metadata key="LogicalAnd" val="[Ampersand][Ampersand]"/>
<metadata key="CshPipeToStdOutAndError" val="[GreaterThan][Ampersand][Exclamation]"/>
<metadata key="ProductionDateTime" val="[DATE.UTC]"/>
<metadata key="DateMilis" val="[DATE_TO_MILLIS([ProductionDateTime],UTC_FORMAT,1970-01-01)]"/>
<metadata key="JobDir" val="[DRAT_HOME]/data/jobs/rat/[DateMilis]"/>
<metadata key="JobInputDir" val="[JobDir]/input"/>
<metadata key="JobOutputDir" val="[JobDir]/output"/>
<metadata key="JobLogDir" val="[JobDir]/logs"/>
<metadata key="ProductType" val="RatLog"/>
</customMetadata>
</pgeConfig>