Package org.vietspider.token.attribute

Examples of org.vietspider.token.attribute.Attributes

94
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
    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);
      String subName = "";
      try {
        URL subUrl = new URL(link);
        String temp = subUrl.getPath();
        if(temp != null) subName += temp;
        temp = subUrl.getQuery();
        if(temp != null) subName += "_"+temp;
        subName = toName(subName);
      }catch (Exception e) {
      }
      if(subName == null || subName.trim().length() < 1) subName = toName(link);
      try{
        crawl(new URL(address), link, subName, level+1, depth);
      }catch (Exception e) {
        System.err.println(e);
      }
      attribute.setValue(subName+".html");
      attributes.set(attribute);
    }


    StringBuilder contentBuilder = new StringBuilder();
    for(int i = 0; i < tokens.size(); i++) {
View Full Code Here
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
      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());
            System.out.println("downloading "+link+" ...");
            if(!file.exists()) loadResource(referer, link, file);
            attribute.setValue(rscName);
            attributes.set(attribute);
          }
        } catch (Exception e) {
          System.err.println(e);
        }
      }
View Full Code Here
87
88
89
90
91
92
93
94
95
96
97
    if (map == null) return ;
    toXMLValue(clazz, object, node);
  }
 
  private void toXMLValue(Class<?> clazz, Object bean, XMLNode node)  throws Exception  {
    Attributes attrs = AttributeParser.parse(node);
    for(Attribute attr : attrs) {
      Field field = getField(attr.getName(), clazz);
      if(field == null) continue;
      Class<?> type = field.getType();
      if(type.isPrimitive() || reflectUtil.isPrimitiveType(type) || type.isEnum()) {
View Full Code Here
141
142
143
144
145
146
147
148
149
150
151
152
      Collection collection = null;
     
      if(current != null && current instanceof Collection) {
        collection = (Collection) current;
      } else if(type.isInterface())  {
        Attributes attributes = AttributeParser.parse(node);
        Attribute attr = attributes.get("type");
        if(attr != null) {
          String typeName = attr.getValue();
          Class<?> clazz = Thread.currentThread().getContextClassLoader().loadClass(typeName);
          collection = (Collection) clazz.newInstance();
        } else {
View Full Code Here
203
204
205
206
207
208
209
210
211
212
213
214
    Map<Object, Object> map = null;
   
    if(current != null && current instanceof Map<?,?>) {
      map = (Map<Object, Object>) current;
    } else if(type.isInterface())  {
      Attributes attributes = AttributeParser.parse(node);
      Attribute attr = attributes.get("type");
      if(attr != null) {
        String typeName = attr.getValue();
        Class<?> clazz = Thread.currentThread().getContextClassLoader().loadClass(typeName);
        map =  (Map<Object, Object>) clazz.newInstance();
      } else {
View Full Code Here
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
    node.addChild(new XMLNode(value.toCharArray(), null, TypeToken.CONTENT));
    children.add(node);
  }
 
  private void setAttribute(XMLNode parent, String name, String value) {
    Attributes attrs = AttributeParser.parse(parent);
    for(Attribute ele : attrs) {
      if(ele.getName().equals(name)){
        ele.setValue(value);
        return ;
      }
    }
    Attribute attr = new Attribute(name, value);
    attrs.add(attr);
  }
View Full Code Here
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
    if(children == null) return ;
    for(XMLNode child : children) {
      if(child.isNode(LINK) && child.getChildren().size() > 0) {
        item.addLink(new MetaLink(child.getChild(0).getTextValue()));
      } else if(child.isNode(ENCLOSURE)) {
        Attributes attributes = AttributeParser.parse(child);
        MetaLink metaLink  = new MetaLink();
        Attribute attribute = attributes.get(URL);
        if(attribute != null) metaLink.setHref(attribute.getValue());
        attribute = attributes.get(TYPE);
        if(attribute != null) metaLink.setType(attribute.getValue());
        item.addLink(metaLink);
      }
    }
  }
View Full Code Here
121
122
123
124
125
126
127
128
129
130
131
132
   
    for(XMLNode child : children) {
      if(!child.isNode(LINK)) continue;
      MetaLink metaLink = toBean.toBean(MetaLink.class, child);
      if(metaLink.getHref() == null || metaLink.getHref().trim().isEmpty()) {
        Attributes attributes = AttributeParser.parse(child);
        Attribute attribute = attributes.get(REF);
        if(attribute != null) metaLink.setHref(attribute.getValue());
      }
      item.addLink(metaLink);
    }
  }
View Full Code Here
609
610
611
612
613
614
615
616
617
618
619
620
    }
  }

  protected void removeIFrameSource(HTMLNode node) {
    if(node.isNode(Name.IFRAME)) {
      Attributes attributes = node.getAttributes();
      attributes.remove("src");
    }
    List<HTMLNode> children = node.getChildren();
    if(children == null || children.size() < 1) return;
    for(int i = 0; i < children.size(); i++) {
      removeIFrameSource(children.get(i));
View Full Code Here
25
26
27
28
29
30
31
32
33
34
35
36
37
38
          System.out.println("===================================================");
          System.out.println(node.getChild(0).getTextValue());
        }
        continue;
      }
      Attributes attributes = node.getAttributes();
      for(int i = 0; i < attributes.size(); i++) {
        String value = attributes.get(i).getValue();
        if(attributes.get(i).getName().startsWith("on")
            || value.toLowerCase().startsWith("javascript")) {
          System.out.println("===================================================");
          System.out.println(value);
        }
      }
View Full Code Here
TOP

Related Classes of org.vietspider.token.attribute.Attributes

  • AttributeExample
  • AttributeExplorer
  • JavascriptExtractor
  • org.vietspider.browser.form.FormUtils
  • org.vietspider.browser.LoginWebsiteHandler
  • org.vietspider.browser.VirtualFormPost
  • org.vietspider.browser.VirtualFormPostBak
  • org.vietspider.html.path2.HTMLExtractor
  • org.vietspider.html.path2.LookupNode
  • org.vietspider.html.renderer.checker.ContentChecker
    • 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.