Examples of LangProfile

Examples of com.cybozu.labs.langdetect.util.LangProfile

98
99
100
101
102
103
104
105
106
107
108
    assertThat(langProfile.getFreq().size(), is(equalTo(freqSize)));
  }

  private static LangProfile readProfileFile(File profileFile) throws FileNotFoundException, IOException {
    FileInputStream input = null;
    final LangProfile langProfile;
    try {
      input = new FileInputStream(profileFile);
      langProfile = LangProfileFactory.readProfile(input);
    } finally {
      IOUtils.closeQuietly(input);
View Full Code Here

Examples of com.cybozu.labs.langdetect.util.LangProfile

32
33
34
35
36
37
38
39
40
41
42
43
        writer.println("Oui ça va. Et toi ?");
      } finally {
        IOUtils.closeQuietly(writer);
      }
     
      LangProfile trucProfile = GenProfile.generate("truc", inputFile);
      HashMap<String, Integer> freqs = trucProfile.getFreq();
      assertThat(freqs, is(notNullValue()));
      assertThat(freqs.get("t"), is(equalTo(8)));
      assertThat(freqs.get("to"), is(equalTo(4)));
      assertThat(freqs.get("out"), is(equalTo(2)));
      assertThat(freqs.get("o"), is(equalTo(7)));
View Full Code Here

Examples of com.cybozu.labs.langdetect.util.LangProfile

72
73
74
75
76
77
78
79
80
81
82
        for (File file: listFiles) {
            if (file.getName().startsWith(".") || !file.isFile()) continue;
            FileInputStream is = null;
            try {
                is = new FileInputStream(file);
                LangProfile profile = JSON.decode(is, LangProfile.class);
                addProfile(profile, index, langsize);
                ++index;
            } catch (JSONException e) {
                throw new LangDetectException(ErrorCode.FormatError, "profile format error in '" + file.getName() + "'");
            } catch (IOException e) {
View Full Code Here

Examples of com.cybozu.labs.langdetect.util.LangProfile

102
103
104
105
106
107
108
109
110
111
112
        if (langsize < 2)
            throw new LangDetectException(ErrorCode.NeedLoadProfileError, "Need more than 2 profiles");
           
        for (String json: json_profiles) {
            try {
                LangProfile profile = JSON.decode(json, LangProfile.class);
                addProfile(profile, index, langsize);
                ++index;
            } catch (JSONException e) {
                throw new LangDetectException(ErrorCode.FormatError, "profile format error");
            }
View Full Code Here

Examples of com.cybozu.labs.langdetect.util.LangProfile

31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
     * @return Language profile instance
     * @throws LangDetectException
     */
    public static LangProfile loadFromWikipediaAbstract(String lang, File file) throws LangDetectException {

        LangProfile profile = new LangProfile(lang);

        BufferedReader br = null;
        try {
            InputStream is = new FileInputStream(file);
            if (file.getName().endsWith(".gz")) is = new GZIPInputStream(is);
            br = new BufferedReader(new InputStreamReader(is, "utf-8"));

            TagExtractor tagextractor = new TagExtractor("abstract", 100);

            XMLStreamReader reader = null;
            try {
                XMLInputFactory factory = XMLInputFactory.newInstance();
                reader = factory.createXMLStreamReader(br);
                while (reader.hasNext()) {
                    switch (reader.next()) {
                    case XMLStreamReader.START_ELEMENT:
                        tagextractor.setTag(reader.getName().toString());
                        break;
                    case XMLStreamReader.CHARACTERS:
                        tagextractor.add(reader.getText());
                        break;
                    case XMLStreamReader.END_ELEMENT:
                        String text = tagextractor.closeTag();
                        if (text != null) profile.update(text);
                        break;
                    }
                }
            } catch (XMLStreamException e) {
                throw new LangDetectException(ErrorCode.TrainDataFormatError, "Training database file '" + file.getName() + "' is an invalid XML.");
View Full Code Here

Examples of com.cybozu.labs.langdetect.util.LangProfile

88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
     * @return Language profile instance
     * @throws LangDetectException
     */
    public static LangProfile loadFromText(String lang, File file) throws LangDetectException {

        LangProfile profile = new LangProfile(lang);

        BufferedReader is = null;
        try {
            is = new BufferedReader(new InputStreamReader(new FileInputStream(file), "utf-8"));

            int count = 0;
            while (is.ready()) {
                String line = is.readLine();
                profile.update(line);
                ++count;
            }

            System.out.println(lang + ":" + count);

View Full Code Here

Examples of com.cybozu.labs.langdetect.util.LangProfile

131
132
133
134
135
136
137
138
139
140
141
142
                continue;
            }

            FileOutputStream os = null;
            try {
                LangProfile profile = GenProfile.loadFromWikipediaAbstract(lang, file);
                profile.omitLessFreq();

                File profile_path = new File(get("directory") + "/profiles/" + lang);
                os = new FileOutputStream(profile_path);
                JSON.encode(profile, os);
            } catch (JSONException e) {
View Full Code Here

Examples of com.cybozu.labs.langdetect.util.LangProfile

178
179
180
181
182
183
184
185
186
187
188
189
            return;
        }

        FileOutputStream os = null;
        try {
            LangProfile profile = GenProfile.loadFromText(lang, file);
            profile.omitLessFreq();

            File profile_path = new File(lang);
            os = new FileOutputStream(profile_path);
            JSON.encode(profile, os);
        } catch (JSONException e) {
View Full Code Here
TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.