- Timestamp:
- 2016-10-25T14:49:36+13:00 (8 years ago)
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
other-projects/hathitrust/solr-extracted-features/trunk/RUN.bash
r30912 r30918 1 1 #!/bin/bash 2 2 3 spark-submit --class org.hathitrust.PrepareForIngest --master local[4] target/htrc-ef-ingest-0.9-jar-with-dependencies.jar --json-filelist=pd-file-listing-step10000.txt pd-ef-json-files pd-solr-files $* 3 input_dir=pd-ef-json-files 4 output_dir=pd-solr-json-files 5 6 master_opt="--master local[4]" 7 self_contained_jar=target/htrc-ef-ingest-0.9-jar-with-dependencies.jar 8 base_cmd="spark-submit --class org.hathitrust.PrepareForIngest $master_opt $self_contained_jar" 9 10 if [ $# -ge 1 ] ; then 11 file_listing=shift $* 12 $base_cmd --json-filelist="$file_listing" $input_dir $output_dir $* 13 else 14 echo "****" 15 echo "* Processing all files in: $input_dir" 16 echo "****" 17 $base_cmd $input_dir $output_dir $* 18 fi 19 20 # spark-submit --class org.hathitrust.PrepareForIngest --master local[4] target/htrc-ef-ingest-0.9-jar-with-dependencies.jar --json-filelist=pd-file-listing-step10000.txt pd-ef-json-files pd-solr-json-files $* 4 21 5 22 # spark-submit --class org.hathitrust.PrepareForIngest --master local[4] target\htrc-ef-ingest-0.9-jar-with-dependencies.jar --json-filelist=pd-file-listing-step1000.txt json-files solr-files $*
Note:
See TracChangeset
for help on using the changeset viewer.