Changeset 31375 for other-projects/hathitrust/wcsa/extracted-features-solr/trunk/solr-ingest/src/main/java/org/hathitrust/extractedfeatures/PerPageJSONFlatmap.java

Timestamp:

2017-01-31T21:35:50+13:00 (7 years ago)

Author:

davidb

Message:

Initial cut at including POS information to solr index

File:

• other-projects/hathitrust/wcsa/extracted-features-solr/trunk/solr-ingest/src/main/java/org/hathitrust/extractedfeatures/PerPageJSONFlatmap.java (modified) (2 diffs)

other-projects/hathitrust/wcsa/extracted-features-solr/trunk/solr-ingest/src/main/java/org/hathitrust/extractedfeatures/PerPageJSONFlatmap.java

@@ -33 +33 @@
     
     protected WhitelistBloomFilter _whitelist_bloomfilter;
-
+    protected UniversalPOSLangMap _universal_langmap = null;
     
     protected DoubleAccumulator _progress_accum;
@@ -136 +136 @@
                     // Convert to Solr add form
                     JSONObject solr_add_doc_json 
-                        = SolrDocJSON.generateSolrDocJSON(volume_id, page_id, ef_page, _whitelist_bloomfilter,_icu_tokenize);
+                        = SolrDocJSON.generateSolrDocJSON(volume_id, page_id, ef_page, 
+                                                    _whitelist_bloomfilter, _universal_langmap,
+                                                    _icu_tokenize);
                     solr_add_doc_json.put("filename_json_bz2", output_json_bz2);