/**
* 递归钻取正文信息
*
* @param nodeP
* @return
*/
@SuppressWarnings("unchecked")
protected List<Node> extractHtml(Node nodeP, String type) throws Exception {
NodeList nodeList = nodeP.getChildren();
if ((nodeList == null) || (nodeList.size() == 0)) {
return null;
}
ArrayList tableList = new ArrayList();
try {
for (NodeIterator e = nodeList.elements(); e.hasMoreNodes(); ) {
Node node = (Node) e.nextNode();
if (node instanceof LinkTag) {
tableList.add(node);
} else if (node instanceof ScriptTag
|| node instanceof StyleTag
|| node instanceof SelectTag) {
} else if (node instanceof TextNode) {
if (node.getText().length() > 0) {
tableList.add(node);
}
} else {
List tempList = extractHtml(node, type);
if ((tempList != null) && (tempList.size() > 0)) {
Iterator ti = tempList.iterator();
while (ti.hasNext()) {
tableList.add(ti.next());
}
}
}
}
} catch (Exception e) {
return null;
}
if ((tableList != null) && (tableList.size() > 0)) {
TableContext tc = new TableContext();
tc.setLinkList(new ArrayList());
tc.setTextBuffer(new StringBuffer());
tableNumber++;
tc.setTableRow(tableNumber);
Iterator ti = tableList.iterator();
// 得到设置的搜索URL
String baseUrl = Config.getSingleConfig(ConfigItem.SEARCH_BASE_URL);
while (ti.hasNext()) {
Node node = (Node) ti.next();
if (node instanceof LinkTag) {
LinkTag linkTag = (LinkTag) node;
if (!"1".equalsIgnoreCase(type)) {
linkTag.setAttribute(
"href", baseUrl + SearchHelper.encrypt(linkTag.getAttribute("href")));
}
tc.getLinkList().add(linkTag);
} else {
tc.getTextBuffer().append(node.getText());
}
}
return tableList;
}
return null;
}
public ContentModel view(String url) {
ContentModel model = new ContentModel();
try {
NodeFilter filter = new TagNameFilter("html");
Parser parser = new Parser();
parser.setURL(SearchHelper.decrypt(url));
parser.setEncoding(parser.getEncoding());
// parser.setEncoding("gb2312");
NodeList list = parser.extractAllNodesThatMatch(filter);
for (int i = 0; i < list.size(); i++) {
String s = list.elementAt(i).toHtml();
model.setContent(s);
}
} catch (Exception e) {
e.printStackTrace();
}
return model;
}