Package org.vietspider.common.text

Examples of org.vietspider.common.text.TextCounter

70
71
72
73
74
75
76
77
78
79
80
   
    if(links.size() < 1) {
      return null;
    }

    TextCounter counter = new TextCounter();
    if(counter.countSentence(builder) > 3) return null;
    if(counter.countWords(builder) > 30) return null;
   
    return nodes;
  }
View Full Code Here
95
96
97
98
99
100
101
102
103
104
    String text = new String(links.get(0).getValue());
    text = text.trim();
    char ch = text.charAt(0);
    if(Character.isLetter(ch) && Character.isLowerCase(ch)) return false;
   
    TextCounter counter = new TextCounter();
    if(counter.countSentence(text) > 1) return false;
   
    return true;
  }
View Full Code Here
48
49
50
51
52
53
54
55
56
57
58
    return searchUpper(parent, names);
  }
 
  public boolean isValidText(HTMLNode node, int size) {
    TextRenderer textRenderer = new TextRenderer(node, null);
    TextCounter counter = new TextCounter();
//    System.out.println(textRenderer.getTextValue());
//    System.out.println(counter.count(textRenderer.getTextValue()));
    if(counter.countSentence(textRenderer.getTextValue()) > size) return false;
    return true;
  }
View Full Code Here
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
  public int getScore() {  return score; }
 
  private void analytics() {
    if(maxPattern > -1) return;
   
    TextCounter textCounter = new TextCounter();
    int s = start;
    int index = start;
    String text = renderer.getTextValue();
    int pattern = 0;
    maxPattern = 0;
    minPattern = -1;
   
    while(index < end) {
      char c = text.charAt(index);
      if(c == '\\') {
        //tinh toan diem cho phan truoc
        int sentence = textCounter.countSentence(text, s, index);
        int word = textCounter.countWord(text, s, index);
//        if(sentence > 5) {
//          System.out.println("========================================");
//          System.out.println(text.substring(s, index));
//          System.out.println(" ===== > "+ sentence+  " : "
//              + word + " : "+ calculate(pattern, sentence, word));
//          System.out.println("========================================");
//        }
        score += new ScoreCalculator().calculate(pattern, sentence, word);
        //end ket thuc
        pattern = 0;
        while(c == '\\') {
          pattern++;
          index++;
          if(index >= end) break;
          c = text.charAt(index);
        }
        s = index;
        if(pattern > maxPattern) maxPattern = pattern;
        if(pattern < minPattern || minPattern < 0) minPattern =  pattern;
        timePattern++;
        totalSeparator += pattern;
      }
      index++;
    }
   
    if(s < index) {
      int sentence = textCounter.countSentence(text, s, index);
      int word = textCounter.countWord(text, s, index);
      score += new ScoreCalculator().calculate(pattern, sentence, word);
    }
   
//    System.out.println("================================================================");
    for(int i = 0;  i < nodes.size(); i++) {
View Full Code Here
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
 
  public int getScore() { return score; }
 
  private void processTextValue() {
    String [] elements = text.split("\n");
    TextCounter textCounter = new TextCounter();
    for(int i = 0; i < elements.length; i++) {
      int value  = textCounter.countSentence(elements[i]);
      if(value > 1) {
//        System.out.println(" : == =======================================");
//        System.out.println("papa a "+ elements[i]);
        paragraph++;
      }
      sentence += value;
     
      word += textCounter.countWords(elements[i]);
    }
  }
View Full Code Here
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
    return true;
  }

  private void processTextValue() {
    String [] elements = text.split("\n");
    TextCounter textCounter = new TextCounter();
    for(int i = 0; i < elements.length; i++) {
      int value  = textCounter.countSentence(elements[i]);
      if(value > 1) {
//        System.out.println(" : == =======================================");
//        System.out.println("papa a "+ elements[i]);
        paragraph++;
      }
      sentence += value;
     
      word += textCounter.countWords(elements[i]);
    }
  }
View Full Code Here
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
    }
   
    List<AnalyticsModel> models = new ArrayList<AnalyticsModel>(nodes.length);
    int maxScore = 0;
   
    TextCounter textCounter = new TextCounter();
    ScoreCalculator scoreCalculator = new ScoreCalculator();
    for(int i = 0; i < nodes.length; i++) {
      if(isTextElement(nodes[i])) return;
      else if(isBlockElement(nodes[i])) {
        AnalyticsRenderer renderer = new AnalyticsRenderer(nodes[i], true);
        String text = renderer.getTextValue().toString();
        int sentence = textCounter.countSentence(text, 0, text.length());
        int word = textCounter.countWord(text, 0, text.length());

        int score = scoreCalculator.calculate(sentence, word);
        models.add(new AnalyticsModel(nodes[i], score));
        if(score > maxScore) maxScore = score;
      }
View Full Code Here
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
  public boolean hasParagraph(HTMLNode node, int max) {
    AnalyticsRenderer renderer = new AnalyticsRenderer(node, true);
    if(renderer.getParagraph() < 1) return false;

    String [] elements = renderer.getTextValue().toString().split("\n");
    TextCounter textCounter = contentChecker.getTextCounter();
    for(String element : elements) {
      int counter = textCounter.countSentence(element);
      if(counter > max) return true;

      counter = textCounter.countWord(element, 0, element.length());
      if(counter >= 15) return true;
    }
    return false;
  }
View Full Code Here
TOP

Related Classes of org.vietspider.common.text.TextCounter

  • org.vietspider.document.util.NodeRemover
  • org.vietspider.document.util.OtherLinkRemover2
  • org.vietspider.html.renderer.checker.LinkBlockChecker
  • org.vietspider.html.renderer.content.AnalyticsRenderer
  • org.vietspider.html.renderer.content.BlockAnalytics
  • org.vietspider.html.renderer.content.TextAnalyticsRenderer
  • org.vietspider.html.renderer.NodeRenderer
  • 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.