jsoup

import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.select.Elements;

public class jsOup {

	private String getCssFun(String selector) {
		char[] ch = selector.toCharArray();
		boolean bool = false; // 默认没有找到CSS_FUN元素
		String findFun = null; // 找到CSS_FUN元素下标
		for (int i = 0; i < ch.length && ch.length > 5 && !bool; i++) {
			if (ch[i] == ':') {
				if (ch[i + 1] == 'e' && ch[i + 2] == 'q' && ch[i + 3] == '(') {
					// 找到:eq(
					findFun = ":eq(";
					bool = true;
				} else if (ch[i + 1] == 'g' && ch[i + 2] == 't'
						&& ch[i + 3] == '(') {
					// 找到:gt(
					findFun = ":gt(";
					bool = true;
				} else if (ch[i + 1] == 'l' && ch[i + 2] == 't'
						&& ch[i + 3] == '(') {
					// 找到:lt(
					findFun = ":lt(";
					bool = true;
				}
			}
		}
		return findFun;
	}

	/**
	 * 从父级元素中获取指定选择表达式的元素集合。 考虑情况： 多个EQ，多个LT，多个GT的情况，还有日后增加一个办法
	 * 
	 * @param selector
	 *            选择表达式
	 * @param elements
	 *            parentElement
	 * @return 返回符合选择表达式的元素集合。
	 */
	public Elements getElements(String selector, Elements parentElement)
			throws Exception {
		if (selector == null || selector.length() == 0) {
			return parentElement;
		}
		if (parentElement == null) {
			return parentElement;
		}
		selector = selector.trim();
		String findFun = getCssFun(selector);
		if (findFun == null) { // 已经找不到CSS_FUN元素
			return parentElement.select(selector);
		} else {
			int cssFunIndex = selector.indexOf(findFun);
			String first = selector.substring(0, cssFunIndex); // 获取(EQ|GT|LT)前面的字符串
			String center = selector.substring(cssFunIndex + findFun.length(),
					selector.length());
			String end = center.substring(center.indexOf(")") + 1,
					center.length()); // 获取EQ后面的字符串
			String cssFunNum = center.substring(0, center.indexOf(")"));
			if (first.length() != 0) {
				parentElement = parentElement.select(first);
			}
			if (findFun.equals(":eq(")) {
				parentElement = parentElement.eq(Integer.parseInt(cssFunNum));
			} else if (findFun.equals(":lt(")) {
				parentElement = new Elements(parentElement.subList(0,
						Integer.parseInt(cssFunNum)));
			} else if (findFun.equals(":gt(")) {
				parentElement = new Elements(parentElement.subList(
						Integer.parseInt(cssFunNum), parentElement.size()));
			}
			if (end.length() == 0) { // 处于最后，selector设置为null
				selector = null;
			} else { // CSS_FUN元素不处于最后
				selector = end;
			}
			return getElements(selector, parentElement);
		}
	}

	public void test() {
		Document doc;
		try {
			String url = "http://club.history.sina.com.cn/viewthread.php?tid=5483877&sudaref=bbs.sina.com.cn&retcode=0";
			url = "http://www.agrilink.cn/";
			doc = Jsoup.connect(url).get();
			String query = "div.myInfo_up:lt(1)";
			query = "div.myInfo_up:lt(3):lt(1)";
			query = "div.myInfo_up:lt(3):gt(1)";
			query = "div.myInfo_up:lt(3):gt(2) a";
			query = "div.myInfo_up:lt(3):gt(2) a:eq(1)";
			Elements newsHeadlines = getElements(query, doc.select("html"));// getString(query,
			// doc);//
			//
			doc.select(query);// doc.select("div.myInfo_up:eq(0)").select("font");
			int i = 0;
			for (Element element : newsHeadlines) {
				System.out.println("div=" + element.html());
				System.out.println("------------" + (i++) + "----------"
						+ element.id());
			}
		} catch (Exception e) {
			// TODO Auto-generated catch block
			e.printStackTrace();
		}
	}

	public static void main(String args[]) {
		jsOup jsOup = new jsOup();
		jsOup.test();
	}

}