Changeset 31220 for other-projects/hathitrust/wcsa/extracted-features-solr/trunk/solr-ingest/src/main/java/org/hathitrust/extractedfeatures/PerPageJSONFlatmap.java
- Timestamp:
- 2016-12-12T20:18:04+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/PerPageJSONFlatmap.java
r31096 r31220 25 25 26 26 protected String _input_dir; 27 28 protected WhitelistBloomFilter _whitelist_bloomfilter; 29 27 30 protected String _solr_url; 28 31 protected String _output_dir; … … 34 37 boolean _strict_file_io; 35 38 36 public PerPageJSONFlatmap(String input_dir, String solr_url, String output_dir, int verbosity, 39 public PerPageJSONFlatmap(String input_dir, String whitelist_filename, 40 String solr_url, String output_dir, int verbosity, 37 41 DoubleAccumulator progress_accum, double progress_step, 38 42 boolean strict_file_io) 39 43 { 40 44 _input_dir = input_dir; 45 46 if (whitelist_filename != null) { 47 _whitelist_bloomfilter = new WhitelistBloomFilter(whitelist_filename,true); 48 } 49 41 50 _solr_url = solr_url; 42 51 _output_dir = output_dir; … … 51 60 52 61 public Iterator<JSONObject> call(String json_file_in) throws IOException 53 //public void call(String json_file_in)54 62 { 55 //ClusterFileIO.memory_usage("Before BZIP2 JSON file read"); 56 String full_json_file_in = _input_dir + "/" + json_file_in; 63 String full_json_file_in = _input_dir + "/" + json_file_in; 57 64 JSONObject extracted_feature_record = JSONClusterFileIO.readJSONFile(full_json_file_in); 58 //ClusterFileIO.memory_usage("After BZIP2 JSON file read"); 59 65 60 66 ArrayList<JSONObject> json_pages = new ArrayList<JSONObject>(); 61 67
Note:
See TracChangeset
for help on using the changeset viewer.