Changeset 31374
- Timestamp:
- 2017-01-31T00:22:39+13:00 (7 years ago)
- File:
-
- 1 edited
Legend:
- Unmodified
- Added
- Removed
-
other-projects/hathitrust/wcsa/extracted-features-solr/trunk/solr-ingest/src/main/java/org/hathitrust/extractedfeatures/ProcessForSolrIngest.java
r31372 r31374 34 34 35 35 protected String _input_dir; 36 protected String _json_list_filename;36 //protected String _json_list_filename; 37 37 protected String _whitelist_filename; 38 38 protected String _solr_url; … … 41 41 protected int _verbosity; 42 42 43 public ProcessForSolrIngest(String input_dir, String json_list_filename,43 public ProcessForSolrIngest(String input_dir, /*String json_list_filename,*/ 44 44 String solr_url, String output_dir, int verbosity) 45 45 { 46 46 _input_dir = input_dir; 47 _json_list_filename = (json_list_filename != null) ? json_list_filename : input_dir;47 //_json_list_filename = (json_list_filename != null) ? json_list_filename : input_dir; 48 48 49 49 boolean use_whitelist = Boolean.getBoolean("wcsa-ef-ingest.use-whitelist"); … … 58 58 { 59 59 String spark_app_name = "[" + exec_mode + "] Extract Features: Process for Solr Ingest"; 60 spark_app_name += " [" + _json_list_filename + "]";60 //spark_app_name += " [" + _json_list_filename + "]"; 61 61 62 62 if (_solr_url != null) { … … 101 101 102 102 //String packed_sequence_path = "hdfs:///user/capitanu/data/packed-ef"; 103 String packed_sequence_path = _ json_list_filename;103 String packed_sequence_path = _input_dir; 104 104 105 105 JavaPairRDD<Text, Text> input_pair_rdd = jsc.sequenceFile(packed_sequence_path, Text.class, Text.class); … … 309 309 String[] filtered_args = cmd.getArgs(); 310 310 311 if (filtered_args.length != 2) {311 if (filtered_args.length != 1) { 312 312 print_usage(formatter,options); 313 313 System.exit(1); … … 343 343 344 344 String input_dir = filtered_args[0]; 345 String json_list_filename = filtered_args[1];345 //String json_list_filename = filtered_args[1]; 346 346 347 347 ProcessForSolrIngest prep_for_ingest 348 = new ProcessForSolrIngest(input_dir, json_list_filename,solr_url,output_dir,verbosity);348 = new ProcessForSolrIngest(input_dir,/*json_list_filename,*/solr_url,output_dir,verbosity); 349 349 350 350 prep_for_ingest.execPerVolumeSequenceFile();
Note:
See TracChangeset
for help on using the changeset viewer.