1 | package org.hathitrust.extractedfeatures;
|
---|
2 |
|
---|
3 | import java.io.IOException;
|
---|
4 |
|
---|
5 | import org.apache.hadoop.io.Text;
|
---|
6 | import org.apache.spark.api.java.function.Function;
|
---|
7 | import org.json.JSONObject;
|
---|
8 |
|
---|
9 | class PerVolumeCatalogLangSequenceFileMap implements Function<Text, String>
|
---|
10 | {
|
---|
11 | private static final long serialVersionUID = 1L;
|
---|
12 |
|
---|
13 | protected String _input_dir;
|
---|
14 | protected int _verbosity;
|
---|
15 |
|
---|
16 | boolean _strict_file_io;
|
---|
17 |
|
---|
18 | public PerVolumeCatalogLangSequenceFileMap(String input_dir, int verbosity, boolean strict_file_io)
|
---|
19 | {
|
---|
20 | _input_dir = input_dir;
|
---|
21 | _verbosity = verbosity;
|
---|
22 |
|
---|
23 | _strict_file_io = strict_file_io;
|
---|
24 | }
|
---|
25 |
|
---|
26 | public String call(Text json_text) throws IOException
|
---|
27 | {
|
---|
28 | String catalog_lang = null;
|
---|
29 |
|
---|
30 | try {
|
---|
31 | JSONObject extracted_feature_record = new JSONObject(json_text.toString());
|
---|
32 |
|
---|
33 | String volume_id = extracted_feature_record.getString("id");
|
---|
34 |
|
---|
35 | JSONObject ef_metadata = extracted_feature_record.getJSONObject("metadata");
|
---|
36 |
|
---|
37 | if (_verbosity >= 1) {
|
---|
38 | System.out.println("Processing: " + volume_id);
|
---|
39 | }
|
---|
40 |
|
---|
41 | if (ef_metadata != null) {
|
---|
42 | String ef_catalog_language = ef_metadata.getString("language");
|
---|
43 | if (!ef_catalog_language.equals("")) {
|
---|
44 |
|
---|
45 | catalog_lang = ef_catalog_language;
|
---|
46 | }
|
---|
47 | else {
|
---|
48 | System.err.println("No catalog 'language' metadata => Skipping id: " + volume_id);
|
---|
49 | }
|
---|
50 | }
|
---|
51 | else {
|
---|
52 | System.err.println("No 'metadata' section in JSON file => Skipping id: " + volume_id);
|
---|
53 | }
|
---|
54 |
|
---|
55 | }
|
---|
56 | catch (Exception e) {
|
---|
57 | if (_strict_file_io) {
|
---|
58 | throw e;
|
---|
59 | }
|
---|
60 | else {
|
---|
61 | e.printStackTrace();
|
---|
62 | }
|
---|
63 | }
|
---|
64 |
|
---|
65 | return catalog_lang;
|
---|
66 | }
|
---|
67 | }
|
---|
68 |
|
---|