Package org.vietspider.html.parser

Examples of org.vietspider.html.parser.NodeImpl

85
86
87
88
89
90
91
92
93
94
95
96
97
98
      nextPath = nextPath.substring(commonPath.length());
      nextPath = nextPath.substring(1, nextPath.indexOf('.', 1));
      int nextIdx = Integer.parseInt(nextPath);

      HTMLNode newCommonNode =
        new NodeImpl(commonNode.getValue(), commonNode.getName(), TypeToken.TAG);
      List<HTMLNode> children = commonNode.getChildren();
      for (i = idx; i < nextIdx; i++) {
        newCommonNode.addChild(children.get(i));
      }
      return newCommonNode;
    } catch (Exception e) {
      // LogService.getInstance().setMessage(e.toString());
      return root;
View Full Code Here
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
      // if(idx > 1) path = path.substring(2, idx);
      // if(path.charAt(0) == '.') path = path.substring(1);
      // if(path.trim().isEmpty()) return root;
      idx = Integer.parseInt(path);

      HTMLNode newCommonNode = new NodeImpl(commonNode.getValue(), commonNode.getName(),
          TypeToken.TAG);
      List<HTMLNode> children = commonNode.getChildren();
      for (i = idx; i < children.size(); i++) {
        newCommonNode.addChild(children.get(i));
        // children.get(i).setParent(newCommonNode);
      }
      return newCommonNode;
    } catch (Exception e) {
      StringBuilder builder = new StringBuilder();
View Full Code Here
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
  @Deprecated()
  public List<String> scanScriptLink(List<NodeImpl> tokens) {
    IdentifierAttributeHandler handler = new IdentifierAttributeHandler(null, null,  "*", "onclick");
    for(int i = 0; i < tokens.size(); i++) {
      NodeImpl token = tokens.get(i);
      handler.handle(token);
    }
   
    List<String> list = handler.getValues();
    handler = new IdentifierAttributeHandler(list, jsVerifier, "a", "href");
   
    for(int i = 0; i < tokens.size(); i++) {
      NodeImpl token = tokens.get(i);
      handler.handle(token);
    }
   
    return list;
  }
View Full Code Here
118
119
120
121
122
123
124
125
126
127
 
  public synchronized List<String> scanSiteLink(List<NodeImpl> tokens) {
    LinkAttributeHandler handler = new LinkAttributeHandler(null, siteLinkVerifier, linkAttributeMap);
//    MapAttributeHandler handler = new MapAttributeHandler(null, siteLinkVerifier, linkAttributeMap);
    for(int i = 0; i < tokens.size(); i++) {
      NodeImpl token = tokens.get(i);
      handler.handle(token);
    }
    return handler.getValues();
  }
View Full Code Here
164
165
166
167
168
169
170
171
172
173
    }
  }
 
  public  synchronized void createFullNormalLink(List<NodeImpl> tokens, URL home) {
    for(int i = 0; i < tokens.size(); i++) {
      NodeImpl nodeImpl = tokens.get(i);
      if(!nodeImpl.isTag()) continue;
      createFullSingleLink(nodeImpl, pageAttributeFullMap, home, normalLinkVerifier);
    }
  }
View Full Code Here
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
  public TextRendererBak (HTMLNode root, List<HTMLNode> contents, int type, boolean constain) {
    List<HTMLNode> tokens =  new ArrayList<HTMLNode>();
//    root.buildTokens(tokens);

    for(int i = 0; i < tokens.size(); i++) {
      NodeImpl node = (NodeImpl)tokens.get(i);
      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]);
            }
            int end = builder.length();
            positions.add(new NodePosition(node, start, end));
          }
        }
        break;
      case IMG:
        positions.add(new NodePosition(node, -1, -1));
        break;
      case H1:
      case H2:
      case H3:
      case H4:
      case H5:
      case H6:
      case TR:
      case TABLE:
      case TD:
      case P:
      case DIV:
      case BR:
        if(!isEndWithNewLine(builder)) {
          builder.append('\n');
          if(type == RENDERER) builder.append('\n');
        }
        break;
      case SCRIPT:
      case STYLE:
        if(node.getType() == TypeToken.TAG) i++;
        break;
      default:
        if(builder.length() > 0) {
          char c = builder.charAt(builder.length()-1);
          if(!(Character.isWhitespace(c)
View Full Code Here
64
65
66
67
68
69
70
71
72
73
74
  }
 
  public List<String> scanScriptLink(List<NodeImpl> tokens) {
    List<String> values  = new ArrayList<String>();
    for(int i = 0; i < tokens.size(); i++) {
      NodeImpl token = tokens.get(i);
      getAttributes(token, values, "*", "onclick", null);
      values.addAll(getAttributes(token, null, "a", "href", jsVerifier));
    }
    return values;
  }
View Full Code Here
82
83
84
85
86
87
88
89
90
91
  }
 
  public synchronized List<String> getSiteLink(List<NodeImpl> tokens) {
    List<String> values  = new ArrayList<String>();
    for(int i = 0; i < tokens.size(); i++) {
      NodeImpl token = tokens.get(i);
      getAttributes(token, values, linkAttributeMap, siteLinkVerifier);      
    }
    return values;
  }
View Full Code Here
92
93
94
95
96
97
98
99
100
101
102
103
104
//    resources.add(new Resource("link", "href"));
//    resources.add(new Resource("script", "src"));
    downloadResources(address, tokens, resources);

    for(int i = 0; i < tokens.size(); i++) {
      NodeImpl token = tokens.get(i);
      if(token.getType() != TypeToken.TAG || !token.isNode(Name.A)) continue;
      Attributes attributes = token.getAttributes();
      Attribute attribute = attributes.get("href");
      if(attribute == null) continue;
      String link  = attribute.getValue();
      if(link == null || link.trim().length() < 1) continue;
      link  = urlUtils.createURL(parent, link);
View Full Code Here
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
    builder.append('>');
  }

  public void downloadResources(String referer, List<NodeImpl> tokens, List<Resource> resources) {
    for(int i = 0; i < tokens.size(); i++) {
      NodeImpl token = tokens.get(i);
      if(token.getType() != TypeToken.TAG) continue;
      for(Resource resource : resources) {
        if(!token.isNode(resource.tag))  continue;
        try {
          Attributes attributes = token.getAttributes();
          for(Attribute attribute : attributes) {
            if(!attribute.getName().equalsIgnoreCase(resource.attr))  continue;
            String rscName = toName(attribute.getValue());
            File file = new File(folder, rscName);
            String link = urlUtils.createURL(url, attribute.getValue());
View Full Code Here
TOP

Related Classes of org.vietspider.html.parser.NodeImpl

  • org.vietspider.browser.form.FormUtils
  • org.vietspider.browser.LoginWebsiteHandler
  • org.vietspider.browser.VirtualFormPost
  • org.vietspider.browser.VirtualFormPostBak
  • org.vietspider.chars.CharsEncoder
  • org.vietspider.html.NodeConfig
  • org.vietspider.html.renderer.ContentRendererBak
  • org.vietspider.html.renderer.TextRenderer
  • org.vietspider.html.util.HTMLAnchorUtil
  • org.vietspider.html.util.HyperLinkUtil
    • 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.