Package it.cnr.isti.hpc.wikipedia.article

Examples of it.cnr.isti.hpc.wikipedia.article.Article

72
73
74
75
76
77
78
79
80
81
82
83
84
  }
 
 
  @Test
  public void testNotRedirect() throws IOException {
    Article a = new Article();
    String mediawiki = IOUtils.getFileAsUTF8String("./src/test/resources/en/liberalism.txt");
    parser.parse(a, mediawiki);
    System.out.println(a.getRedirect());
    assertTrue(! a.isRedirect());
   
   
  }
View Full Code Here
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
        return;
      }
      if (page.isMain())
        type = Type.ARTICLE;

      Article article = new Article();
      article.setTitle(title);
      article.setWikiId(Integer.parseInt(id));
      article.setNamespace(namespace);
      article.setIntegerNamespace(integerNamespace);
      article.setTimestamp(timestamp);
      article.setType(type);
      parser.parse(article, page.getText());

      try {
        out.write(article.toJson());
        out.write("\n");
      } catch (IOException e) {
        logger.error("writing the output file {}", e.toString());
        System.exit(-1);
      }
View Full Code Here
42
43
44
45
46
47
48
49
50
51
52
  public void testParsing() throws UnsupportedEncodingException, FileNotFoundException, IOException, SAXException {
    URL u = this.getClass().getResource("/en/mercedes.xml");
    WikipediaArticleReader wap = new WikipediaArticleReader(u.getFile(),"/tmp/mercedes.json.gz", Language.EN);
    wap.start();
    String json = IOUtils.getFileAsUTF8String("/tmp/mercedes.json.gz");
    Article a = Article.fromJson(json);
    assertTrue(a.getCleanText().startsWith("Mercedes-Benz"));
    assertEquals(15, a.getCategories().size());
   
   
  }
View Full Code Here
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
  }

 
  @Test
  public void testInfobox() throws IOException {
    Article articleWithInfobox = new Article();

    String text = readFileAsString("/it/xml-dump/article-with-infobox.txt");
    articleParser.parse(articleWithInfobox, text);
   
    assertTrue(articleWithInfobox.hasInfobox());
    Template infobox = articleWithInfobox.getInfobox();
    assertEquals(12,infobox.getSchema().size());
    assertEquals("Infobox_fiume", infobox.getName());
    assertEquals("Adige", infobox.get("nome"));
    assertEquals("12200", infobox.get("bacino"));
   
View Full Code Here
91
92
93
94
95
96
97
98
99
100
101
102
103
104
  }

  @Test
  public void table() throws IOException {
    Article articleWithTable = new Article();
    String text = readFileAsString("/it/xml-dump/table.txt");
    articleParser.parse(articleWithTable, text);
    assertEquals("Nome italiano", articleWithTable.getTables().get(0)
        .getColumn(1).get(0));
    assertEquals("15 agosto", articleWithTable.getTables().get(0)
        .getColumn(0).get(10));

  }
View Full Code Here
105
106
107
108
109
110
111
112
113
114
115
  @Test
  public void list() throws IOException {

    String text = readFileAsString("/it/xml-dump/list.txt");
    Article articleWithList = new Article();
    articleParser.parse(articleWithList, text);
    List<String> list = articleWithList.getLists().get(2);
    assertEquals("Antropologia culturale e Antropologia dei simboli", list.get(0));
   
  }
View Full Code Here
37
38
39
40
41
42
43
44
45
46
47
48
49
50
  ArticleParser parser = new ArticleParser(Language.EN);
 
 
  @Test
  public void testParsing() throws IOException {
    Article a = new Article();
    String mediawiki = IOUtils.getFileAsUTF8String("./src/test/resources/en/article.txt");
    parser.parse(a, mediawiki);
    assertTrue("Wrong parsed text",a.getCleanText().trim().startsWith("Albedo (), or reflection coefficient, is the diffuse reflectivity or reflecting power of a surface."));
    assertEquals(5, a.getCategories().size());
    assertEquals(7,a.getSections().size());
    assertEquals(74,a.getLinks().size());
   
  }
View Full Code Here
51
52
53
54
55
56
57
58
59
60
61
62
 

 
  @Test
  public void testMercedes() throws IOException {
    Article a = new Article();
    String mediawiki = IOUtils.getFileAsUTF8String("./src/test/resources/en/mercedes.txt");
    parser.parse(a, mediawiki);
    assertTrue(a.getCleanText().startsWith("Mercedes-Benz"));
    assertEquals(15, a.getCategories().size());
   
  }
View Full Code Here
62
63
64
65
66
67
68
69
70
71
72
  }
 
 
  @Test
  public void testDisambiguation() throws IOException {
    Article a = new Article();
    String mediawiki = IOUtils.getFileAsUTF8String("./src/test/resources/en/hdis.txt");
    parser.parse(a, mediawiki);
    assertTrue(a.isDisambiguation());
   
  }
View Full Code Here
TOP

Related Classes of it.cnr.isti.hpc.wikipedia.article.Article

  • it.cnr.isti.hpc.wikipedia.article.en.ArticleTest
  • it.cnr.isti.hpc.wikipedia.article.it.ArticleTest
  • it.cnr.isti.hpc.wikipedia.reader.WikipediaArticleReader$JsonConverter
  • it.cnr.isti.hpc.wikipedia.reader.WikipediaArticleReaderTest
    • 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.