Package org.vietspider.html

Examples of org.vietspider.html.NodeIterator

278
279
280
281
282
283
284
285
286
287
288
289
290
//    return images;
  }
 
  public String getCharset(HTMLDocument document) {
    HTMLNode root = document.getRoot();
    NodeIterator iterator = root.iterator();
    while(iterator.hasNext()) {
      HTMLNode n = iterator.next();
      if(!n.isNode(Name.META)) continue;
      if(n.isNode(Name.BODY)) break;
      Attributes attributes = n.getAttributes();
      Attribute attribute = attributes.get("http-equiv");
      if(attribute == null || attribute.getValue() == null) continue;
View Full Code Here
74
75
76
77
78
79
80
81
82
83
84
85
86
    }
  }
 
  public String getCharset(HTMLDocument document) throws Exception {
    HTMLNode root = document.getRoot();
    NodeIterator iterator = root.iterator();
    while(iterator.hasNext()) {
      HTMLNode n = iterator.next();
      if(!n.isNode(Name.META)) continue;
      if(n.isNode(Name.BODY)) break;
      Attributes attributes = n.getAttributes();
      Attribute attribute = attributes.get("http-equiv");
      if(attribute == null || attribute.getValue() == null) continue;
View Full Code Here
35
36
37
38
39
40
41
42
43
44
45
46
47
    HTMLParser2 parser2 = new HTMLParser2();
    HTMLDocument document  = parser2.createDocument(file, null);
   
    List<String> attrValues = new ArrayList<String>();
   
    NodeIterator iterator = document.getRoot().iterator();
    while(iterator.hasNext()) {
      HTMLNode node = iterator.next();
      if(node.isNode(nodeName)) {
        Attributes attributes = node.getAttributes();
        Attribute attribute = attributes.get(attrName);
        if(attribute == null) continue;
        attrValues.add(attribute.getValue());
View Full Code Here
TOP

Related Classes of org.vietspider.html.NodeIterator

  • AttributeExplorer
  • JavascriptExtractor
  • org.vietspider.html.renderer.checker.ContentChecker
  • org.vietspider.html.renderer.checker.FormChecker
  • org.vietspider.html.renderer.checker.FormNodeRemover
  • org.vietspider.html.renderer.checker.LinkBlockChecker
  • org.vietspider.html.renderer.checker.ULNodeChecker
  • org.vietspider.html.renderer.ContentRegionSearcher
  • org.vietspider.html.renderer.ContentRegionSearcher2
  • org.vietspider.html.renderer.ContentRendererBak
    • 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.