source: other-projects/hathitrust/solr-extracted-features/trunk/src/main/java/org/hathitrust/extractedfeatures/PagedJSONForeach.java@ 31002

Last change on this file since 31002 was 31002, checked in by davidb, 7 years ago

Need to separate flatMap and foreach calls in PagedJSON

  • Property svn:executable set to *
File size: 2.1 KB
Line 
1package org.hathitrust.extractedfeatures;
2
3import java.util.ArrayList;
4import java.util.Iterator;
5import org.apache.spark.api.java.function.FlatMapFunction;
6import org.apache.spark.api.java.function.VoidFunction;
7import org.apache.spark.util.DoubleAccumulator;
8import org.json.JSONArray;
9import org.json.JSONObject;
10
11/*
12class PagedJSON implements Function<String, Boolean> {
13
14 private static final long serialVersionUID = 1L;
15
16 public Boolean call(String s) { return s.contains("a"); }
17}
18 */
19
20
21class PagedJSONForeach extends BasePerJSON implements VoidFunction<JSONObject>
22//public class PagedJSON implements VoidFunction<String>
23{
24 private static final long serialVersionUID = 1L;
25
26 public PagedJSONForeach(String input_dir, String solr_url, String output_dir, int verbosity,
27 DoubleAccumulator progress_accum, double progress_step)
28 {
29 super(input_dir,solr_url,output_dir,verbosity,progress_accum,progress_step);
30 }
31
32
33 public void call(JSONObject solr_add_doc_json)
34 {
35 String output_json_bz2 = solr_add_doc_json.getString("filename_json_bz2");
36 solr_add_doc_json.remove("filename_json_bz2");
37
38 boolean random_test = (Math.random()>0.999); // every 1000
39
40 if ((_verbosity >=2) && (random_test)) {
41 System.out.println("==================");
42 System.out.println("Sample output Solr add JSON [random test 1/1000]: " + solr_add_doc_json.toString());
43 System.out.println("==================");
44 }
45
46
47 if (_solr_url != null) {
48 if ((_verbosity >=2) && (random_test)) {
49 System.out.println("==================");
50 System.out.println("Posting to: " + _solr_url);
51 System.out.println("==================");
52 }
53 JSONSolrTransform.postSolrDoc(_solr_url, solr_add_doc_json);
54 }
55
56 if (_output_dir != null) {
57 if ((_verbosity >=2) && (random_test)) {
58 System.out.println("==================");
59 System.out.println("Saving to: " + _output_dir);
60 System.out.println("==================");
61 }
62 JSONSolrTransform.saveSolrDoc(solr_add_doc_json, _output_dir + "/" + output_json_bz2);
63 }
64 }
65
66}
67
Note: See TracBrowser for help on using the repository browser.