Package org.htmlcleaner

Examples of org.htmlcleaner.HtmlCleaner.clean()

21
22
23
24
25
26
27
28
29
30
31
//    String page = HttpServer.getHttpServer().httpGetPage( "http://s3.travian.net/dorf1.php");
     
    HtmlCleaner cleaner = new HtmlCleaner();
    CleanerProperties props = cleaner.getProperties();
    props.setRecognizeUnicodeChars( true);
        TagNode node = cleaner.clean( new File("tests/overview8.html"));
//        TagNode node = cleaner.clean( page);

        updateVillages2( villages, node);

  }
View Full Code Here
19
20
21
22
23
24
25
26
27
28
29
  {
    // 1. goto village
    // 2. goto inner village
    // 3. goto rally point
    HtmlCleaner cleaner = new HtmlCleaner();
    TagNode node = cleaner.clean( new File("tests/materias.html"));

    int  numNode = 0;
        Object[] domTroops = node.evaluateXPath( "//div[@id='content']//table");
        for( Object t : domTroops) {
          if( t instanceof TagNode) {
View Full Code Here
19
20
21
22
23
24
25
26
27
28
29
//        String page = HttpServer.getHttpServer().getMapPage( coordX, coordY);
       
        HtmlCleaner cleaner = new HtmlCleaner();
        CleanerProperties props = cleaner.getProperties();
        props.setRecognizeUnicodeChars( true);
        TagNode node = cleaner.clean( new File("tests/karte.htm"));
//        TagNode node = cleaner.clean( page);

        return getSquareUrl2( node);
    }
    private static URL getSquareUrl2( TagNode node) throws Exception
View Full Code Here
43
44
45
46
47
48
49
50
51
52
53
        Thread.sleep( (long) (1000 + 2000*Math.random()));
       
        HtmlCleaner cleaner = new HtmlCleaner();
        CleanerProperties props = cleaner.getProperties();
        props.setRecognizeUnicodeChars( true);
        TagNode node = cleaner.clean( new File("tests/karte_village.htm"));
//        TagNode node = cleaner.clean( page);

        return getSquareType2( node);

    }
View Full Code Here
18
19
20
21
22
23
24
25
26
27
28
  }
 
  public void recherche(String lien) throws MalformedURLException, IOException, XPatherException
  {
    HtmlCleaner cleaner = new HtmlCleaner();
    TagNode node = cleaner.clean(new URL(lien));
    info = new String();
    int count = 0;
   
    for (Object o : node.evaluateXPath("//body//p"))
    {
View Full Code Here
98
99
100
101
102
103
104
105
106
107
108
    props.setTranslateSpecialEntities(true);
    props.setTransSpecialEntitiesToNCR(true);
   
    HtmlCleaner cleaner = new HtmlCleaner(props);
     
    TagNode node = cleaner.clean(new URL(lien));
    //System.out.println("Title: " + ((TagNode)(node.evaluateXPath("//title")[0])).getText());
    for (Object o : node.evaluateXPath("//ul[@id='acces_1']//li/a"))
    {
      String dUrl = ((TagNode)(o)).getAttributeByName("href");
      //System.out.println("LI: " + org.apache.commons.lang3.StringEscapeUtils.unescapeHtml4(((TagNode)(o)).getText().toString()));
View Full Code Here
161
162
163
164
165
166
167
168
169
170
171
    props.setTransSpecialEntitiesToNCR(true);
    props.setOmitComments(true);
     
    HtmlCleaner cleaner = new HtmlCleaner(props);

    TagNode node = cleaner.clean(new URL(lien));
   
    for (Object o : node.evaluateXPath("//div[@id='retour_accueil']/a/img"))
    {
      //System.out.println(((TagNode)(o)).getAllChildren());
      lien_logo = ((TagNode)(o)).getAttributeByName("src");
View Full Code Here
206
207
208
209
210
211
212
213
214
215
216
    props.setOmitComments(true);
     
    HtmlCleaner cleaner = new HtmlCleaner(props);
     
    int i=0;
    TagNode node = cleaner.clean(new URL(lien));
   
    for (Object o : node.evaluateXPath("//div[@class='encadre_fiche firstencadre']/div/div/a"))
    {
      lien_site_spe = ((TagNode)(o)).getAttributeByName("href");
      //System.out.println("lien spe "+lien_site_spe);
View Full Code Here
110
111
112
113
114
115
116
117
118
119
120
 
 
  public boolean recherche(String mot, String lien) throws MalformedURLException, IOException, XPatherException
  {
    HtmlCleaner cleaner = new HtmlCleaner();
    TagNode node = cleaner.clean(new URL(lien));
    boolean bool = false;

    String[] decoupage = mot.split(" ");
    if(decoupage.length == 1)
    {
View Full Code Here
92
93
94
95
96
97
98
99
100
101
102
  }
 
  private static String getError(String response) throws IOException{
    String error = null;
    HtmlCleaner cleaner = new HtmlCleaner();
    TagNode html = cleaner.clean(response);
    TagNode errortag = html.findElementByAttValue("id", "error", true, true);
    if (errortag != null){
      error = errortag.getAttributeByName("title");
    }
    return error;
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.