Package org.vietspider.html

Examples of org.vietspider.html.HTMLNode

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
    return newDocument;
  }

  public HTMLDocument[] extractRow(HTMLDocument document, NodePath[] nodePaths) {
    List<List<HTMLNode>> listHtmlValues = new ArrayList<List<HTMLNode>>();
    HTMLNode root = document.getRoot();

    for(int i = 0; i < nodePaths.length; i++) {
      List<HTMLNode> matchValues = matchNodes(root, nodePaths[i]);
      if(matchValues != null) listHtmlValues.add(matchValues);
    }

    if(listHtmlValues.size() == 0 || listHtmlValues.get(0) == null) return new HTMLDocument[0];

    List<HTMLNode> htmlValues = listHtmlValues.get(0);
    HTMLDocument [] newDocuments = new HTMLDocument[htmlValues.size()];

    for(int i = 0; i < htmlValues.size(); i++) {
      HTMLNode html = HTMLParser2.clone(root);
      if(htmlValues.get(i) == null) continue;
      html.addChild(htmlValues.get(i));
//      htmlValues.get(i).setParent(html);
      for(int j = 1; j < listHtmlValues.size(); j++) {
        List<HTMLNode> newHtmlValues = listHtmlValues.get(j);
        if(i > newHtmlValues.size()) break;
        try {
          if(newHtmlValues.get(i) == null) continue;
          html.addChild(newHtmlValues.get(i));
//          newHtmlValues.get(i).setParent(html);
        } catch (Exception e) {
          continue;
        }
      }
View Full Code Here
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
    for(; start > -1; start--) {
      NodeImpl node = tokens.get(start);
      if(node.isNode(Name.FORM)) break;
    }

    HTMLNode form = null;
    boolean md5 = false;
    List<HTMLNode> inputs = new ArrayList<HTMLNode>();
    String formValue = null;

    for(int i = start; i < tokens.size(); i++) {
      NodeImpl node = tokens.get(i);
      if(node.isNode(Name.FORM)) {
        if(node.isOpen()) {
          if(!md5) {
            String value = new String(node.getValue());
            md5 = value.toLowerCase().indexOf("md5") > -1;
          }
          form = node;
          formValue = new String(form.getValue()).toLowerCase();
        } else {
          break;
        }
      } else if(node.isNode(Name.INPUT)) {
        if(!md5 && formValue != null) {
View Full Code Here
125
126
127
128
129
130
131
132
133
134
135
      }
     
      Node node = (Node)inodes[i];
      for(HTMLNode htmlNode : htmlNodes) {
        if(htmlNode == null) continue;
        HTMLNode test = lookNode(htmlNode, node);
        if(test == null) {
          continue;
        }
        htmlValues.add(lookNode(htmlNode, node));
      }
View Full Code Here
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
//    System.out.println("node expresstion "+nodeExp.toString());
//    System.out.println("attributes length "+nodeExp.getAttributes().length);
    int counter  = 0;
    NodeMatcher matcher = new NodeMatcher();
    for(int i = 0; i < htmlChildren.size(); i++) {
      HTMLNode childNode =  htmlChildren.get(i);
      if(nodeExp.getName() != childNode.getName()) continue;
      if(matcher.match(nodeExp.getPattern(), counter)) {
        Attribute [] attrs = nodeExp.getAttributes();
        if(attrs == null || attrs.length < 1) {
//          System.out.println(" da xay ra roi ");
          htmlValues.add(childNode);
        } else {
//          System.out.println(" xay ra ");
          Attributes nodeAttributes = childNode.getAttributes();
          if(matcher.contains(nodeAttributes, attrs)) htmlValues.add(childNode);
        }
      }
      counter++;
    } 
View Full Code Here
191
192
193
194
195
196
197
198
199
200
201
202
      if(matchValues != null) nodes.addAll(matchValues);
    }

    for(HTMLNode node : nodes) {
      if(node == null) continue;
      HTMLNode parent  = node.getParent();
      if(parent == null) continue;
//      System.out.println(" truoc " + parent.getChildren().size());
      parent.removeChild(node);
//      System.out.println(" sau " + parent.getChildren().size());
    }
  }
View Full Code Here
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
//      System.out.println(" sau " + parent.getChildren().size());
    }
  }

  public void removeFrom(HTMLNode root, NodePath path){
    HTMLNode element = lookNode(root, path);
    if (element == null) return;
    java.util.Iterator<HTMLNode> iter =  element.getParent().getChildren().iterator();
    boolean remove = false;
    while(iter.hasNext()){
      HTMLNode ele = iter.next();
      if(!remove) remove = ele == element;
      if(remove) iter.remove();
    }
  }
View Full Code Here
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
    this.linkChecker = linkChecker;
    StringBuilder builder = new StringBuilder();
   
    NodeIterator iterator = root.iterator(ignores);
    while(iterator.hasNext()) {
      HTMLNode node = iterator.next();
      switch (node.getName()) {
      case CONTENT:
        char [] chars = node.getValue();
        if(!isEmpty(chars)) {
//          if(isValid(contents, node, constain)) {
            int start = builder.length();
            for(int k = 0; k < chars.length; k++) {
              builder.append(chars[k] == '\n' ? ' ' : chars[k]);
            }
            HTMLNode parent = node.getParent();
            if(parent != null && parent.isNode(Name.SPAN)) builder.append(' ');
           
            int end = builder.length();
            positions.add(new NodePosition(node, start, end));
          }
//        }
View Full Code Here
199
200
201
202
203
204
205
206
207
208
209
 
  private boolean isWrapperContent(HTMLNode node){
    List<HTMLNode> children = node.getChildren();
    if(children == null) return false;
    for(int i = 0; i < children.size(); i++) {
      HTMLNode child = children.get(i);
      if(child.isNode(Name.CONTENT)
         || isWrapperContent(child)) return true;
    }
    return false;
  }
View Full Code Here
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
  public static HTMLNode searchBody(HTMLDocument document) throws Exception {
    RefsDecoder decoder = new RefsDecoder();
    NodeIterator iterator = document.getRoot().iterator();
    while(iterator.hasNext()) {
      HTMLNode node = iterator.next();
      if(!node.isNode(Name.CONTENT)) continue;
      char [] chars = node.getValue();
      chars = decoder.decode(chars);

      chars = CharsUtil.cutAndTrim(chars, 0, chars.length);
      chars =  java.text.Normalizer.normalize(new String(chars), Normalizer.Form.NFC).toCharArray();
      node.setValue(chars);             
    } 

    HTMLExtractor extractor  = new HTMLExtractor();
    NodePathParser pathParser = new NodePathParser();
View Full Code Here
40
41
42
43
44
45
46
47
48
49
50
51
//      if(isValid(contents, node, constain)) {
        int start = builder.length();
        for(int k = 0; k < chars.length; k++) {
          builder.append(chars[k] == '\n' ? ' ' : chars[k]);
        }
        HTMLNode parent = node.getParent();
        if(parent != null && parent.isNode(Name.SPAN)) builder.append(' ');

        int end = builder.length();
        positions.add(new NodePosition(node, start, end));
      }
//    }
View Full Code Here
TOP

Related Classes of org.vietspider.html.HTMLNode

  • AttributeExplorer
  • ExtractForum
  • HTMLNodeExplorer
  • JavascriptExtractor
  • org.vietspider.browser.LoginUtils
  • org.vietspider.browser.LoginWebsiteHandler
  • org.vietspider.browser.VirtualFormPost
  • org.vietspider.browser.VirtualFormPostBak
  • org.vietspider.document.util.ImageDescRemover
  • org.vietspider.document.util.NodeRemover
    • 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.