annotate software/mpdl-services/mpiwg-mpdl-cms/src/de/mpg/mpiwg/berlin/mpdl/cms/translator/MicrosoftTranslator.java @ 23:e845310098ba

diverse Korrekturen
author Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
date Tue, 27 Nov 2012 12:35:19 +0100
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
23
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
1 package de.mpg.mpiwg.berlin.mpdl.cms.translator;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
2
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
3 import java.util.ArrayList;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
4
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
5 import com.memetix.mst.language.Language;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
6 import com.memetix.mst.translate.Translate;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
7 import com.memetix.mst.detect.Detect;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
8
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
9 import de.mpg.mpiwg.berlin.mpdl.exception.ApplicationException;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
10
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
11 public class MicrosoftTranslator {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
12 private static final String KEY = "474A4E72DB217E37031EC190ACB4159378A6917C"; // free key to use with Microsoft server
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
13 private static String[] DEST_LANGUAGES = {"eng", "ger", "fra"};
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
14
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
15 private MicrosoftTranslator() {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
16 // nothing: prevent instantiation
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
17 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
18
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
19 public static String[] getDestLanguages() {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
20 return DEST_LANGUAGES;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
21 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
22
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
23 public static String translate(String query, String toLanguageStr) throws ApplicationException {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
24 if (toLanguageStr == null)
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
25 throw new ApplicationException("MicrosoftTranslator: toLanguageStr is null");
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
26 String translation = null;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
27 try {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
28 Translate.setKey(KEY); // Set the API key once per JVM. It is set statically and applies to all services
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
29 String langId = de.mpg.mpiwg.berlin.mpdl.lt.general.Language.getInstance().getLanguageId(toLanguageStr); // e.g. "de" is delivered from "deu"
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
30 Language toLanguage = Language.fromString(langId);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
31 translation = Translate.execute(query, toLanguage);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
32 translation = translation.toLowerCase();
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
33 } catch (Exception e) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
34 throw new ApplicationException(e);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
35 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
36 return translation;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
37 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
38
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
39 public static String translate(String query, String fromLanguageStr, String toLanguageStr) throws ApplicationException {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
40 if (fromLanguageStr == null)
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
41 throw new ApplicationException("MicrosoftTranslator: toLanguageStr is null");
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
42 if (toLanguageStr == null)
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
43 throw new ApplicationException("MicrosoftTranslator: fromLanguageStr is null");
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
44 String translation = null;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
45 try {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
46 Translate.setKey(KEY);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
47 String fromLangId = de.mpg.mpiwg.berlin.mpdl.lt.general.Language.getInstance().getLanguageId(fromLanguageStr); // e.g. "de" is delivered from "deu"
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
48 String toLangId = de.mpg.mpiwg.berlin.mpdl.lt.general.Language.getInstance().getLanguageId(toLanguageStr); // e.g. "de" is delivered from "deu"
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
49 Language fromLanguage = Language.fromString(fromLangId);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
50 Language toLanguage = Language.fromString(toLangId);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
51 translation = Translate.execute(query, fromLanguage, toLanguage);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
52 translation = translation.toLowerCase();
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
53 } catch (Exception e) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
54 throw new ApplicationException(e);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
55 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
56 return translation;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
57 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
58
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
59 public static ArrayList<String> translate(String[] query, String fromLanguageStr, String[] toLanguagesStr) throws ApplicationException {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
60 ArrayList<String> translations = new ArrayList<String>();
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
61 for (int i=0; i<toLanguagesStr.length; i++) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
62 String toLanguageStr = toLanguagesStr[i];
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
63 String[] translationsArray = translate(query, fromLanguageStr, toLanguageStr);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
64 for (int j=0; j<translationsArray.length; j++) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
65 String translation = translationsArray[j];
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
66 translations.add(translation);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
67 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
68 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
69 return translations;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
70 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
71
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
72 public static String[] translate(String[] query, String fromLanguageStr, String toLanguageStr) throws ApplicationException {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
73 if (fromLanguageStr == null)
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
74 throw new ApplicationException("MicrosoftTranslator: toLanguageStr is null");
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
75 if (toLanguageStr == null)
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
76 throw new ApplicationException("MicrosoftTranslator: fromLanguageStr is null");
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
77 String[] translation = null;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
78 try {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
79 Translate.setKey(KEY);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
80 String fromLangId = de.mpg.mpiwg.berlin.mpdl.lt.general.Language.getInstance().getLanguageId(fromLanguageStr); // e.g. "de" is delivered from "deu"
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
81 String toLangId = de.mpg.mpiwg.berlin.mpdl.lt.general.Language.getInstance().getLanguageId(toLanguageStr); // e.g. "de" is delivered from "deu"
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
82 Language fromLanguage = Language.fromString(fromLangId);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
83 Language toLanguage = Language.fromString(toLangId);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
84 translation = Translate.execute(query, fromLanguage, toLanguage);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
85 for (int i=0; i<translation.length; i++) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
86 translation[i] = translation[i].toLowerCase();
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
87 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
88 } catch (Exception e) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
89 throw new ApplicationException(e);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
90 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
91 return translation;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
92 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
93
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
94 public static String detectLanguageCode(String query) throws ApplicationException {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
95 String langCode = null;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
96 try {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
97 Detect.setKey(KEY);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
98 Language detectedLanguage = Detect.execute(query);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
99 if (detectedLanguage != null)
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
100 langCode = detectedLanguage.toString();
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
101 } catch (Exception e) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
102 throw new ApplicationException(e);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
103 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
104 return langCode;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
105 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
106
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
107 public static String detectLanguageName(String query) throws ApplicationException {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
108 String langName = null;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
109 try {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
110 Detect.setKey(KEY);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
111 Language detectedLanguage = Detect.execute(query);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
112 if (detectedLanguage != null)
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
113 langName = detectedLanguage.getName(Language.GERMAN);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
114 } catch (Exception e) {
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
115 throw new ApplicationException(e);
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
116 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
117 return langName;
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
118 }
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
119
e845310098ba diverse Korrekturen
Josef Willenborg <jwillenborg@mpiwg-berlin.mpg.de>
parents:
diff changeset
120 }