Changeset 30975

Timestamp:

2016-10-28T14:28:51+13:00 (7 years ago)

Author:

davidb

Message:

Introduction of new solr-url command line argument, leading to some other adjustments in RUN scripts

Location:

other-projects/hathitrust/solr-extracted-features/trunk

Files:

• RUN-PD-CLUSTER.bash (modified) (1 diff)
• _RUN.bash (modified) (3 diffs)
• src/main/java/org/hathitrust/PagedJSON.java (modified) (3 diffs)
• src/main/java/org/hathitrust/PrepareForIngest.java (modified) (7 diffs)

other-projects/hathitrust/solr-extracted-features/trunk/RUN-PD-CLUSTER.bash

@@ -7 +7 @@
 #input_dir="hdfs://10.10.0.52:9000/user/htrc/pd-ef-json-files"
 
-output_dir=hdfs://master:9000/user/htrc/pd-solr-json-files
-
+#output_dir=hdfs://master:9000/user/htrc/pd-solr-json-files
+solr_url="http://10.11.0.53:8983/solr/htrc-pd-ef/update"
 
 master_opt="--master spark://10.10.0.52:7077"

other-projects/hathitrust/solr-extracted-features/trunk/_RUN.bash

@@ -19 +19 @@
 fi
 
-if [ "x$output_dir" = "x" ] ; then
-    echo "_RUN.bash: Failed to set 'output_dir'" 1>&2
-    exit
-fi
+#if [ "x$output_dir" = "x" ] ; then
+#    echo "_RUN.bash: Failed to set 'output_dir'" 1>&2
+#    exit
+#fi
 
 run_jps=0
@@ -71 +71 @@
     
 self_contained_jar=target/htrc-ef-ingest-0.9-jar-with-dependencies.jar
-base_cmd="spark-submit --class org.hathitrust.PrepareForIngest $master_opt $self_contained_jar"
+cmd="spark-submit --class org.hathitrust.PrepareForIngest $master_opt $self_contained_jar"
 
-cmd="$base_cmd --verbosity 1 $json_filelist $input_dir $output_dir $*"
+if [ "x$solr_url" != "x" ] ; then
+    cmd="$cmd --solr-url $solr_url"
+fi
+
+if [ "x$output_dir" != "x" ] ; then
+    cmd="$cmd --output-dir $output_dir"
+fi
+
+
+cmd="$cmd --verbosity 1 $input_dir $json_filelist $*"
 
 echo "****"
@@ -79 +88 @@
 echo "*   $cmd"
 echo "****"
+
 if [ "$run_jps" = "1" ] ; then
   echo "* Monitor progress on Spark cluster through:"

other-projects/hathitrust/solr-extracted-features/trunk/src/main/java/org/hathitrust/PagedJSON.java

@@ -31 +31 @@
 
     protected String _input_dir;
+    protected String _solr_url;
     protected String _output_dir;
     protected int    _verbosity;
     
-    public PagedJSON(String input_dir, String output_dir, int verbosity)
+    public PagedJSON(String input_dir, String solr_url, String output_dir, int verbosity)
     {
         _input_dir  = input_dir;
+        _solr_url   = solr_url;
         _output_dir = output_dir;
         _verbosity  = verbosity;
@@ -170 +172 @@
         */
         
-        //return solr_doc_json;
         return solr_update_json;
     }
@@ -288 +289 @@
                 */
                             
-                saveSolrDoc(solr_add_doc_json,output_json_bz2);
-                //postSolrDoc(solr_add_doc_json);
-            
+                if (_solr_url != null) {
+                    postSolrDoc(solr_add_doc_json);
+                }
+
+                if (_output_dir != null) {
+                    saveSolrDoc(solr_add_doc_json,output_json_bz2);
+                }
             }
             else {

other-projects/hathitrust/solr-extracted-features/trunk/src/main/java/org/hathitrust/PrepareForIngest.java

@@ -15 +15 @@
     protected String _input_dir;
     protected String _json_list_filename;
+    protected String _solr_url;
     protected String _output_dir;
     
     protected int    _verbosity;
 
-    public PrepareForIngest(String input_dir, String json_list_filename, String output_dir, int verbosity)
+    public PrepareForIngest(String input_dir, String json_list_filename, 
+                            String solr_url, String output_dir, int verbosity)
     {
         _input_dir = input_dir;
@@ -42 +44 @@
         JavaRDD<String> json_list_data = jsc.textFile(_json_list_filename,NUM_PARTITIONS).cache();
 
-        PagedJSON paged_json = new PagedJSON(_input_dir,_output_dir,_verbosity);
+        PagedJSON paged_json = new PagedJSON(_input_dir,_solr_url, _output_dir,_verbosity);
         JavaRDD<String> json_ids = json_list_data.flatMap(paged_json).cache();
 
@@ -52 +54 @@
         System.out.println("");
 
-        System.out.println("############");
         String rdd_save_file = "rdd-solr-json-page-files";
         json_ids.saveAsTextFile(rdd_save_file);
+        System.out.println("############");
         System.out.println("# Saved RDD of Solr JSON page files, top-level, as:");
         System.out.println("#  " + rdd_save_file);
@@ -63 +65 @@
     }
 
+    public static void print_usage(HelpFormatter formatter, Options options)
+    {
+        formatter.printHelp("RUN.bash [options] input-dir json-filelist.txt", options);
+    }
     public static void main(String[] args) {
 
@@ -86 +92 @@
         options.addOption(verbosity_opt);
         
-        //CommandLineParser parser = new DefaultParser(); // 1.3 and above
+        Option output_dir_opt = new Option("o", "output-dir", true, 
+                "If specified, save BZipped Solr JSON files to this directory");
+        output_dir_opt.setRequired(false);
+        options.addOption(output_dir_opt);
+        
+        Option solr_url_opt = new Option("u", "solr-url", true, 
+                "If specified, the URL to post the Solr JSON data to");
+        solr_url_opt.setRequired(false);
+        options.addOption(solr_url_opt);
+        
+        Option dry_run_opt = new Option("r", "dry-run", false, 
+                "Used to initiate a 'dry-run' where the files are all read in, but nothing is ingested/saved");
+        dry_run_opt.setRequired(false);
+        options.addOption(dry_run_opt);
         
         // need to work with CLI v1.2 as this is the JAR that is bundled with Hadoop/Spark 
         CommandLineParser parser = new GnuParser(); 
+        //CommandLineParser parser = new DefaultParser(); // if working with CLI v1.3 and above
+        
         HelpFormatter formatter = new HelpFormatter();
-        CommandLine cmd;
+        CommandLine cmd = null;
 
         try {
@@ -98 +119 @@
         catch (ParseException e) {
             System.err.println(e.getMessage());
-            formatter.printHelp("RUN.bash [options] json-file-list.txt input-dir output-dir", options);
+            print_usage(formatter,options);
             System.exit(1);
-            return;
+            //return;  // prevents 'cmd may not be assigned' compiler error in Eclipse
         }
 
@@ -111 +132 @@
         int verbosity = Integer.parseInt(verbosity_str);
 
+        String output_dir = cmd.getOptionValue("output-dir",null);
+        String solr_url   = cmd.getOptionValue("solr-url",null);
+        boolean dry_run   = cmd.hasOption("dry-run");
+        
         String[] filtered_args = cmd.getArgs();
 
-        if (filtered_args.length != 3) {
-            formatter.printHelp("RUN.bash [options] json-filelist.txt input-dir output-dir", options);
+        if (filtered_args.length != 2) {
+            print_usage(formatter,options);
             System.exit(1);
+        }
+        
+        if (!dry_run && ((output_dir == null) && (solr_url==null))) {
+            System.err.println("Need to specify either --solr-url or --output-dir otherwise generated files are not ingested/saved");
+            print_usage(formatter,options);
         }
         String json_list_filename = filtered_args[0];
         String input_dir  = filtered_args[1];
-        String output_dir = filtered_args[2];
+        //String output_dir = filtered_args[2];
 
-        PrepareForIngest prep_for_ingest = new PrepareForIngest(input_dir,json_list_filename,output_dir,verbosity);
+        PrepareForIngest prep_for_ingest 
+            = new PrepareForIngest(input_dir,json_list_filename,solr_url,output_dir,verbosity);
         prep_for_ingest.exec();