Package pignlproc.markup

Examples of pignlproc.markup.AnnotatingMarkupParser

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
88
89
90
91
92
93
94
95
96
97
98
99
        String markup = reader.getCurrentValue().toString();
        assertEquals(
                "#REDIRECT [[Computer accessibility]] {{R from CamelCase}}",
                markup);

        AnnotatingMarkupParser converter = new AnnotatingMarkupParser();
        String simpleText = converter.parse(markup);
        assertEquals("", simpleText);
        assertTrue(converter.getWikiLinkAnnotations().isEmpty());
        assertEquals("http://en.wikipedia.org/wiki/Computer_accessibility",
                converter.getRedirect());

        // second article
        assertTrue(reader.nextKeyValue());
        assertEquals("Anarchism", reader.getCurrentKey().toString());
        markup = reader.getCurrentValue().toString();
        converter = new AnnotatingMarkupParser();
        simpleText = converter.parse(markup);
        assertTrue(simpleText.startsWith("\nAnarchism is a political philosophy"
                + " encompassing theories and attitudes"));
        assertEquals(465, converter.getWikiLinkAnnotations().size());
        Annotation firstLink = converter.getWikiLinkAnnotations().get(0);
        assertEquals("political philosophy", firstLink.label);
        assertEquals("http://en.wikipedia.org/wiki/Political_philosophy",
                firstLink.value);
        assertEquals(16, firstLink.begin);
        assertEquals(36, firstLink.end);
        assertEquals("political philosophy",
                simpleText.substring(firstLink.begin, firstLink.end));

        // third article
        assertTrue(reader.nextKeyValue());
        assertEquals("AfghanistanHistory", reader.getCurrentKey().toString());
        markup = reader.getCurrentValue().toString();
        converter = new AnnotatingMarkupParser();
        simpleText = converter.parse(markup);
        assertEquals("", simpleText);
        assertEquals(0, converter.getWikiLinkAnnotations().size());
        assertEquals("http://en.wikipedia.org/wiki/History_of_Afghanistan",
                converter.getRedirect());

        // fourth article
        assertTrue(reader.nextKeyValue());
        assertEquals("Autism", reader.getCurrentKey().toString());
        markup = reader.getCurrentValue().toString();
        converter = new AnnotatingMarkupParser();
        simpleText = converter.parse(markup);
        assertTrue(simpleText.contains("Autism is a brain development disorder"
                + " characterized by impaired social interaction and communication"));
        assertEquals(234, converter.getWikiLinkAnnotations().size());
        firstLink = converter.getWikiLinkAnnotations().get(0);
        assertEquals("Neurodevelopmental disorder", firstLink.label);
        assertEquals(
                "http://en.wikipedia.org/wiki/Neurodevelopmental_disorder",
                firstLink.value);
        assertEquals(15, firstLink.begin);
        assertEquals(41, firstLink.end);
        assertEquals("brain development disorder",
                simpleText.substring(firstLink.begin, firstLink.end));

        for (Annotation a: converter.getWikiLinkAnnotations()) {
            // internal anchors are not extracted as links
            assertFalse(a.value.startsWith("#"));
        }

        // there is no fifth article in this test file
View Full Code Here
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
        // first article
        assertTrue(reader.nextKeyValue());
        assertEquals("Antoine Meillet", reader.getCurrentKey().toString());
        String markup = reader.getCurrentValue().toString();
        AnnotatingMarkupParser converter = new AnnotatingMarkupParser();
        String simpleText = converter.parse(markup);
        // TODO: handle date templates
        assertTrue(simpleText.startsWith("Paul Jules Antoine Meillet, né le  à Moulins,"
                + " Allier, mort le  à Châteaumeillant"));
        assertEquals(48, converter.getWikiLinkAnnotations().size());
        List<String> headers = converter.getHeaders();
        assertEquals("Biographie", headers.get(0));
        assertEquals("Antoine Meillet et les études arméniennes",
                headers.get(1));
        assertEquals("Antoine Meillet et les études homériques", headers.get(2));
        assertEquals(8, headers.size());
        List<String> paragraphs = converter.getParagraphs();
        assertEquals(
                "Paul Jules Antoine Meillet, né le  à Moulins, Allier, mort"
                        + " le  à Châteaumeillant, Cher, est le principal linguiste"
                        + " français des premières décennies du .",
                paragraphs.get(0));
        assertEquals(
                "D'origine bourbonnaise, fils d'un notaire de Châteaumeillant (Cher),"
                        + " il fait ses études secondaires au lycée"
                        + " Théodore-de-Banville de Moulins.\n"
                        + "Étudiant à la faculté des lettres de Paris à partir de 1885,"
                        + " il suit notamment les cours de Louis Havet à la Sorbonne, de"
                        + " Michel Bréal au Collège de France et de Ferdinand de Saussure"
                        + " à l'École pratique des hautes études. Il assure à la suite"
                        + " de Saussure le cours de grammaire comparée, qu'il complète à"
                        + " partir de 1894 par une conférence sur l'iranien.",
                paragraphs.get(1));
        assertEquals(15, paragraphs.size());
        assertNull(converter.getRedirect());

        // go to the last article wich is a redirect
        assertTrue(reader.nextKeyValue());
        assertTrue(reader.nextKeyValue());
        assertTrue(reader.nextKeyValue());
        assertTrue(reader.nextKeyValue());
        assertTrue(reader.nextKeyValue());

        assertEquals("Amenophis IV", reader.getCurrentKey().toString());
        markup = reader.getCurrentValue().toString();
        converter = new AnnotatingMarkupParser();
        assertEquals("", converter.parse(markup));
        assertEquals(0, converter.getWikiLinkAnnotations().size());
        assertEquals("http://en.wikipedia.org/wiki/Akh%C3%A9naton",
                converter.getRedirect());

        // this was the last article
        assertFalse(reader.nextKeyValue());
    }
View Full Code Here
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
            }
            String title = reader.getCurrentKey().toString();
            String uri = AnnotatingMarkupParser.titleToUri(title, languageCode);
            String rawMarkup = reader.getCurrentValue().toString();

            AnnotatingMarkupParser converter = new AnnotatingMarkupParser(
                    languageCode);
            String text = converter.parse(rawMarkup);
            String redirect = converter.getRedirect();
            DataBag links = bagFactory.newDefaultBag();
            for (Annotation link : converter.getWikiLinkAnnotations()) {
                links.add(tupleFactory.newTupleNoCopy(Arrays.asList(link.value,
                        link.begin, link.end)));
            }
            DataBag headers = bagFactory.newDefaultBag();
            for (Annotation h : converter.getHeaderAnnotations()) {
                headers.add(tupleFactory.newTupleNoCopy(Arrays.asList(h.value,
                        h.begin, h.end)));
            }
            DataBag paragraphs = bagFactory.newDefaultBag();
            for (Annotation p : converter.getParagraphAnnotations()) {
                paragraphs.add(tupleFactory.newTupleNoCopy(Arrays.asList(
                        p.value, p.begin, p.end)));
            }
            return tupleFactory.newTupleNoCopy(Arrays.asList(title, uri, text,
                    redirect, links, headers, paragraphs));
View Full Code Here
TOP

Related Classes of pignlproc.markup.AnnotatingMarkupParser

  • info.bliki.htmlcleaner.ContentToken
  • info.bliki.htmlcleaner.TagNode
  • info.bliki.wiki.model.ImageFormat
  • info.bliki.wiki.model.WikiModel
  • info.bliki.wiki.tags.WPATag
  • pignlproc.format.TestWikipediaParsing
  • pignlproc.storage.ParsingWikipediaLoader
  • java.util.regex.Matcher
    • 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.