Package org.languagetool

Examples of org.languagetool.JLanguageTool

30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
public class UppercaseSentenceStartRuleTest extends TestCase {

  public void testUkrainian() throws IOException {
    final Ukrainian ukrainian = new Ukrainian();
    final UppercaseSentenceStartRule rule = new UppercaseSentenceStartRule(TestTools.getEnglishMessages(), ukrainian);
    final JLanguageTool lt = new JLanguageTool(ukrainian);

    assertEquals(0, rule.match(lt.getAnalyzedSentence("Автор написав це речення з великої літери.")).length);

    final RuleMatch[] matches = rule.match(lt.getAnalyzedSentence("автор написав це речення з маленької літери."));
    assertEquals(1, matches.length);
    assertEquals(1, matches[0].getSuggestedReplacements().size());
    assertEquals("Автор", matches[0].getSuggestedReplacements().get(0));
   
    assertEquals(0, lt.check("Це список з декількох рядків:\n\nрядок 1,\n\nрядок 2,\n\nрядок 3.").size());
    assertEquals(0, lt.check("Це список з декількох рядків:\n\nрядок 1;\n\nрядок 2;\n\nрядок 3.").size());
    assertEquals(0, lt.check("Це список з декількох рядків:\n\n 1) рядок 1;\n\n2) рядок 2;\n\n3)рядок 3.").size());
  }
View Full Code Here
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
  public void testRule() throws IOException {
    SimpleReplaceRule rule = new SimpleReplaceRule(TestTools.getEnglishMessages());

    RuleMatch[] matches;
    JLanguageTool langTool = new JLanguageTool(new Ukrainian());
   
    // correct sentences:
    matches = rule.match(langTool.getAnalyzedSentence("Ці рядки повинні збігатися."));
    assertEquals(0, matches.length);

    // incorrect sentences:
    matches = rule.match(langTool.getAnalyzedSentence("Ці рядки повинні співпадати."));
    assertEquals(1, matches.length);
    assertEquals(2, matches[0].getSuggestedReplacements().size());
    assertEquals(Arrays.asList("збігатися", "сходитися"), matches[0].getSuggestedReplacements());
   
    matches = rule.match(langTool.getAnalyzedSentence("Нападаючий"));
    assertEquals(1, matches.length);
    assertEquals(Arrays.asList("Нападник", "Нападальний", "Нападний"), matches[0].getSuggestedReplacements());

    matches = rule.match(langTool.getAnalyzedSentence("Нападаючого"));
    assertEquals(1, matches.length);
    assertEquals(Arrays.asList("Нападник", "Нападальний", "Нападний"), matches[0].getSuggestedReplacements());

  }
View Full Code Here
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
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
  @Test
  public void testMorfologikSpeller() throws IOException {
    MorfologikUkrainianSpellerRule rule = new MorfologikUkrainianSpellerRule (TestTools.getMessages("Ukrainian"), new Ukrainian());

    JLanguageTool langTool = new JLanguageTool(new Ukrainian());

    // correct sentences:
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("До вас прийде заввідділу!")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence(",")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("123454")).length);

    assertEquals(0, rule.match(langTool.getAnalyzedSentence("До нас приїде The Beatles!")).length);

    //incorrect sentences:

    RuleMatch[] matches = rule.match(langTool.getAnalyzedSentence("атакуючий"));
    // check match positions:
    assertEquals(1, matches.length);

    matches = rule.match(langTool.getAnalyzedSentence("шкляний"));

    assertEquals(1, matches.length);
    assertEquals("скляний", matches[0].getSuggestedReplacements().get(0));

    assertEquals(0, rule.match(langTool.getAnalyzedSentence("а")).length);

    // mix alphabets
    matches = rule.match(langTool.getAnalyzedSentence("прийдешнiй"));   // latin 'i'

    assertEquals(1, matches.length);
    assertEquals("прийдешній", matches[0].getSuggestedReplacements().get(0));

    // compounding
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Жакет був синьо-жовтого кольору")).length);

    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Він багато сидів на інтернет-форумах")).length);

    assertEquals(1, rule.match(langTool.getAnalyzedSentence("Він багато сидів на інтермет-форумах")).length);

    // abbreviations

    RuleMatch[] match = rule.match(langTool.getAnalyzedSentence("Читання віршів Т.Г.Шевченко і Г.Тютюнника"));
    assertEquals(new ArrayList<RuleMatch>(), Arrays.asList(match));

    match = rule.match(langTool.getAnalyzedSentence("Читання віршів Т. Г. Шевченко і Г. Тютюнника"));
    assertEquals(new ArrayList<RuleMatch>(), Arrays.asList(match));

    match = rule.match(langTool.getAnalyzedSentence("Англі́йська мова (англ. English language, English) належить до германської групи"));
    assertEquals(new ArrayList<RuleMatch>(), Arrays.asList(match));

    match = rule.match(langTool.getAnalyzedSentence("Англі́йська мова (англ English language, English) належить до германської групи"));
    assertEquals(1, match.length);
  }
View Full Code Here
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
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
    public void testMorfologikSpeller() throws IOException {
        MorfologikCatalanSpellerRule rule =
                new MorfologikCatalanSpellerRule (TestTools.getMessages("Catalan"), new Catalan());

        RuleMatch[] matches;
        JLanguageTool langTool = new JLanguageTool(new Catalan());

        // prefixes and suffixes.
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("S'autodefineixin com a populars.")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Redibuixen el futur.")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("L'exdirigent del partit.")).length);

        // correct sentences:
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Abacallanada")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Abatre-les-en")).length);
       
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Allò que més l'interessa.")).length);
        // checks that "WORDCHARS ·-'" is added to Hunspell .aff file
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Porta'n quatre al col·legi.")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Has de portar-me'n moltes.")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence(",")).length);
        // Spellcheck dictionary contains Valencian and general accentuation
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Francès i francés.")).length);
        // checks abbreviations
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Viu al núm. 23 del carrer Nou.")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("N'hi ha de color vermell, blau, verd, etc.")).length);
             
       
        // Test for Multiwords.
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Era vox populi.")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Aquell era l'statu quo.")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("Va ser la XIV edició.")).length);
       
        //test for "LanguageTool":
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("LanguageTool!")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence(",")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("123454")).length);
       
        //tests for mixed case words
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("pH")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("McDonald")).length);

        //incorrect sentences:
        //TODO: quna --> quan ?, uan --> una, Qie-Què, línees->línies
       
        matches = rule.match(langTool.getAnalyzedSentence("quna"));
        assertEquals(1, matches.length);
        //assertEquals("quan", matches[0].getSuggestedReplacements().get(0));
       
        matches = rule.match(langTool.getAnalyzedSentence("Video"));
        assertEquals(1, matches.length);
        assertEquals("vídeo", matches[0].getSuggestedReplacements().get(0));
       
        matches = rule.match(langTool.getAnalyzedSentence("bànner"));
        assertEquals(1, matches.length);
        assertEquals("bàner", matches[0].getSuggestedReplacements().get(0));
       
        matches = rule.match(langTool.getAnalyzedSentence("colaborassió"));
        assertEquals(1, matches.length);
        assertEquals("col·laboració", matches[0].getSuggestedReplacements().get(0));
       
        matches = rule.match(langTool.getAnalyzedSentence("sssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssssss"));
        assertEquals(1, matches.length);
       
        matches = rule.match(langTool.getAnalyzedSentence("plassa"));
        assertEquals(1, matches.length);
        assertEquals("plaça", matches[0].getSuggestedReplacements().get(0));    
       
        matches = rule.match(langTool.getAnalyzedSentence("Deú"));
        assertEquals(1, matches.length);
        assertEquals("Deu", matches[0].getSuggestedReplacements().get(0));
        assertEquals("Déu", matches[0].getSuggestedReplacements().get(1));

        matches = rule.match(langTool.getAnalyzedSentence("joan"));
        assertEquals(1, matches.length);
        assertEquals(0, matches[0].getFromPos());
        assertEquals(4, matches[0].getToPos());
        assertEquals("Joan", matches[0].getSuggestedReplacements().get(0));
       
        matches = rule.match(langTool.getAnalyzedSentence("abatusats"));
        assertEquals(1, matches.length);
        assertEquals(0, matches[0].getFromPos());
        assertEquals(9, matches[0].getToPos());
        assertEquals("abatussats", matches[0].getSuggestedReplacements().get(0));
       
        // incomplete multiword
        matches = rule.match(langTool.getAnalyzedSentence("L'statu"));
        assertEquals(1, matches.length);
        assertEquals(2, matches[0].getFromPos());
        assertEquals(7, matches[0].getToPos());
        assertEquals("tato", matches[0].getSuggestedReplacements().get(0));
       
        matches = rule.match(langTool.getAnalyzedSentence("argüit"));
        assertEquals(1, matches.length);
        assertEquals(0, matches[0].getFromPos());
        assertEquals(6, matches[0].getToPos());
        assertEquals("arguït", matches[0].getSuggestedReplacements().get(0));
        assertEquals("arguïa", matches[0].getSuggestedReplacements().get(1));
        assertEquals("arguïm", matches[0].getSuggestedReplacements().get(2));

       
        matches = rule.match(langTool.getAnalyzedSentence("ángel"));
        assertEquals(1, matches.length);
        assertEquals("Àngel", matches[0].getSuggestedReplacements().get(0));
        assertEquals("àngel", matches[0].getSuggestedReplacements().get(1));
       
        matches = rule.match(langTool.getAnalyzedSentence("caçessim"));
        assertEquals(1, matches.length);
        assertEquals("cacéssim", matches[0].getSuggestedReplacements().get(0));
        assertEquals("caséssim", matches[0].getSuggestedReplacements().get(1));
        assertEquals("casséssim", matches[0].getSuggestedReplacements().get(2));
        assertEquals("cesséssim", matches[0].getSuggestedReplacements().get(3));
       
       
        matches = rule.match(langTool.getAnalyzedSentence("cantaríà"));
        assertEquals(1, matches.length);
        assertEquals("cantaria", matches[0].getSuggestedReplacements().get(0));
        assertEquals("Cantàbria", matches[0].getSuggestedReplacements().get(1));
       
        //best suggestion first
        matches = rule.match(langTool.getAnalyzedSentence("poguem"));
        assertEquals(1, matches.length);
        assertEquals("puguem", matches[0].getSuggestedReplacements().get(0));
       
        //incorrect mixed case words
        assertEquals(1, rule.match(langTool.getAnalyzedSentence("PH")).length);
        assertEquals(1, rule.match(langTool.getAnalyzedSentence("Ph")).length);
        assertEquals(1, rule.match(langTool.getAnalyzedSentence("MCDonald")).length);
       
        matches = rule.match(langTool.getAnalyzedSentence("tAula"));
        assertEquals(1, matches.length);
        assertEquals("taula", matches[0].getSuggestedReplacements().get(0));
       
        matches = rule.match(langTool.getAnalyzedSentence("TAula"));
        assertEquals(1, matches.length);
        assertEquals("taula", matches[0].getSuggestedReplacements().get(0));
       
        matches = rule.match(langTool.getAnalyzedSentence("col·Labora"));
        assertEquals(1, matches.length);
        assertEquals("col·labora", matches[0].getSuggestedReplacements().get(0));
       
        matches = rule.match(langTool.getAnalyzedSentence("col·laborÀ"));
        assertEquals(1, matches.length);
        assertEquals("col·laborà", matches[0].getSuggestedReplacements().get(0));
       
        //capitalized wrong words
        matches = rule.match(langTool.getAnalyzedSentence("En la Pecra"));
        assertEquals(1, matches.length);
        assertEquals("Macra", matches[0].getSuggestedReplacements().get(0));
        assertEquals("Pacià", matches[0].getSuggestedReplacements().get(1));

        assertEquals(1, rule.match(langTool.getAnalyzedSentence("aõh")).length);
        assertEquals(0, rule.match(langTool.getAnalyzedSentence("a")).length);
    }
View Full Code Here
55
56
57
58
59
60
61
    assertMatches("A fost plecat „pentru... «puțin” timp».", 4);
  }

  private void setUpRule(Language language) throws IOException {
    rule = new GenericUnpairedBracketsRule(TestTools.getEnglishMessages(), language);
    langTool = new JLanguageTool(language);
  }
View Full Code Here
44
45
46
47
48
49
50
  @Override
  protected void setUp() throws Exception {
    super.setUp();
    rule = new SimpleReplaceRule(TestTools.getMessages("ro"));
    langTool = new JLanguageTool(new Romanian());
  }
View Full Code Here
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
    // TODO: detect all the cases not preceded with 'das'
  }

  public void testPhraseExceptions() throws IOException {
    CaseRule rule = new CaseRule(null);
    JLanguageTool langTool = new JLanguageTool(new German());

    // correct sentences:
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Das gilt ohne Wenn und Aber.")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("ohne Wenn und Aber")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Das gilt ohne Wenn und Aber bla blubb.")).length);
    // as long as phrase exception isn't complete, there's no error:
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Das gilt ohne wenn")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Das gilt ohne wenn und")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("wenn und aber")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("und aber")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("aber")).length);
    // incorrect sentences:
    // error not found here as it's in the XML rules:
    //assertEquals(1, rule.match(langTool.getAnalyzedSentence("Das gilt ohne wenn und aber.")).length);
  }
View Full Code Here
31
32
33
34
35
36
37
38
public class CompoundRuleTest extends CompoundRuleTestAbs {

  @Override
  protected void setUp() throws Exception {
    super.setUp();
    langTool = new JLanguageTool(new Romanian());
    rule = new CompoundRule(TestTools.getMessages("ro"));
  }
View Full Code Here
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
*/
public class DashRuleTest extends TestCase {

  public void testRule() throws IOException {
    DashRule rule = new DashRule(null);
    JLanguageTool langTool = new JLanguageTool(new German());

    // correct sentences:
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Die große Diäten-Erhöhung kam dann doch.")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Die große Diätenerhöhung kam dann doch.")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Die große Diäten-Erhöhungs-Manie kam dann doch.")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Die große Diäten- und Gehaltserhöhung kam dann doch.")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Die große Diäten- sowie Gehaltserhöhung kam dann doch.")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Die große Diäten- oder Gehaltserhöhung kam dann doch.")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Erst so - Karl-Heinz dann blah.")).length);
    assertEquals(0, rule.match(langTool.getAnalyzedSentence("Erst so -- Karl-Heinz aber...")).length);
   
    // incorrect sentences:
    assertEquals(1, rule.match(langTool.getAnalyzedSentence("Die große Diäten- Erhöhung kam dann doch.")).length);
    assertEquals(1, rule.match(langTool.getAnalyzedSentence("Die große Diäten-  Erhöhung kam dann doch.")).length);
    assertEquals(1, rule.match(langTool.getAnalyzedSentence("Die große Diäten-Erhöhungs- Manie kam dann doch.")).length);
    assertEquals(1, rule.match(langTool.getAnalyzedSentence("Die große Diäten- Erhöhungs-Manie kam dann doch.")).length);
  }
View Full Code Here
33
34
35
36
37
38
39
40
public class RussianCompoundRuleTest extends CompoundRuleTestAbs {

  @Override
  protected void setUp() throws Exception {
    super.setUp();
    langTool = new JLanguageTool(new Russian());
    rule = new RussianCompoundRule(null);
  }
View Full Code Here
TOP

Related Classes of org.languagetool.JLanguageTool

  • org.languagetool.chunking.EnglishChunkerTest
  • org.languagetool.clientexample.Example
  • org.languagetool.commandline.Main
  • org.languagetool.dev.blogs.BlogChecker
  • org.languagetool.dev.ConfusionSetCoverage
  • org.languagetool.dev.conversion.RuleCoverage
  • org.languagetool.dev.dumpcheck.SentenceSourceChecker
  • org.languagetool.dev.eval.AfterTheDeadlineEvaluator
  • org.languagetool.dev.eval.CheckBNC
  • org.languagetool.dev.eval.LanguageToolEvaluator
    • Copyright © 2018 www.massapicom. 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.