Nutch之GoogleSuggest前缀匹配_google chche 前缀匹配-CSDN博客

本文链接：https://blog.csdn.net/ArcticFoxHan/article/details/37302661

本代码是在Nutch1.2的环境下运行成功，仅仅是加入前缀匹配，搜索时会自动提示数据库中有的词条，如果数据库中没有的词条则会自动补充进去。

关键字处理，如果搜索关键字在数据库中，则不添加；如果搜索关键字不在数据库中，则自动添加关键字到数据库中。代码1是处理的servlet，代码2是修改后的index.jsp

//提交数据接收处理
package com.wgh.servlet;
import com.wgh.tools.ConnDB;
import java.io.IOException;
import java.io.PrintWriter;
import java.sql.ResultSet;
import java.sql.SQLException;
import javax.jms.Session;
import javax.servlet.RequestDispatcher;
import javax.servlet.ServletException;
import javax.servlet.http.HttpServlet;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
import org.dom4j.Document;
import org.dom4j.DocumentHelper;
import org.dom4j.Element;
import org.dom4j.io.OutputFormat;
import org.dom4j.io.XMLWriter;
public class KeyServlet extends HttpServlet
{
	private static final long serialVersionUID = 1L;
	protected void doGet(HttpServletRequest request,
			HttpServletResponse response) throws ServletException, IOException {
		this.doPost(request, response);
	}
	private void getKey(HttpServletRequest request, HttpServletResponse response)
			throws ServletException, IOException {
		request.setCharacterEncoding("UTF-8");
		response.setCharacterEncoding("UTF-8");
		String inputKey = request.getParameter("inputKey");
		if (!inputKey.equals("") && inputKey != null) {
			inputKey = java.net.URLDecoder.decode(inputKey, "UTF-8");
			response.setContentType("application/xml");
			Document document = DocumentHelper.createDocument();
			Element keys = DocumentHelper.createElement("keys");
			document.setRootElement(keys);
			ConnDB conn = new ConnDB();
			String sql = "SELECT * FROM tb_key WHERE keyWord LIKE '" + inputKey
					+ "%' AND keyWord !='" + inputKey
					+ "' ORDER BY keyWord ASC";
			ResultSet rs = conn.executeQuery(sql);
			try {
				while (rs.next()) {
					Element key = keys.addElement("keyWord");
					String keyWord = rs.getString(2);
					key.setText(keyWord);
				}
			} catch (SQLException e) {
				e.printStackTrace();
			}
			conn.close();
			OutputFormat format = new OutputFormat();
			format.setEncoding("UTF-8");
			PrintWriter out = response.getWriter();
			XMLWriter writer = new XMLWriter(out, format);
			writer.write(document);
			writer.close();
		}
	}
	protected void doPost(HttpServletRequest request,
			HttpServletResponse response) throws ServletException, IOException {
		String action = request.getParameter("action");
		if ("getKey".equals(action))
			getKey(request, response);
		if ("saveKey".equals(action))
			saveKey(request, response);
	}
	private void saveKey(HttpServletRequest request,HttpServletResponse response)throws ServletException, IOException{
			String inputKey = new String(request.getParameter("suggest").getBytes("ISO-8859-1"), "UTF-8");
			ConnDB conn = new ConnDB();
			if ((inputKey != null) && (!"".equals(inputKey))) {
			String sql = "SELECT * FROM tb_key WHERE keyWord ='"+ inputKey + "'";
			ResultSet rs = conn.executeQuery(sql);
			try
			{
			if (!rs.next()) {
			String sqlIns = "INSERT INTO tb_key (keyWord) VALUES('"+ inputKey + "')";
			int i = conn.executeUpdate(sqlIns);
				}
			}
			catch (SQLException e) {
			e.printStackTrace();}
			}
			conn.close();
			String inputKey1 = new String(inputKey.getBytes("UTF-8"), "ISO-8859-1");
			response.sendRedirect("search.jsp?query=" + inputKey1);
	}
}

<%@ page language="java" contentType="text/html; charset=UTF-8"
	pageEncoding="UTF-8"%>
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<title>北极狐搜</title>
<link href="CSS/style.css" rel="stylesheet">

<script src="js/request.js"></script>
<script language="javascript">
	var request = false;
	var size = 0;//总条目
	var curSize = -1;//当前条目
	//Ajax请求方法
	function getSuggest() {
		if (event.keyCode != 13 && event.keyCode != 38 && event.keyCode != 40) {

			curSize = -1;//当前条目
			var suggest = document.getElementById("suggest").value;
			var url = "KeyServlet"; //服务器地址
			var param = "action=getKey&inputKey=" + encodeURI(suggest)
					+ "&nocache=" + new Date().getTime();
			request = httpRequest("post", url, true, callbackFunc, param); //调用请求方法 	

		}
	}
	//Ajax回调函数
	function callbackFunc() {
		if (request.readyState == 4) { //判断响应是否完成 

			if (request.status == 200) { //判断响应是否成功

				var doc = request.responseXML; //获取XML格式的关键字
				var contentObj = document.getElementById("content");//获取content对象
				size = contentObj.childNodes.length; //设置条目数
				//清空提示内容对象
				while (true) {
					if (contentObj.hasChildNodes()) {
						contentObj.removeChild(contentObj.firstChild);//移除指定的节点
					} else {
						break;//跳出while循环
					}
				}
				var content = doc.getElementsByTagName("keyWord");//获取XML文档中的keyWord节点
				if (content.length > 0) {
					for ( var i = 0; i < content.length; i++) {
						addContent(content(i).firstChild.nodeValue);
					}
					document.getElementById("toolBox").style.visibility = "visible";//显示提示框
				} else {
					document.getElementById("toolBox").style.visibility = "hidden";//隐藏提示框
				}
			}
		}
	}
	window.onload = function() {
		document.getElementsByTagName('body')[0].onpropertychange = ctrlMove;//监听onkeydown事件
	}
	//通过↑键和↓键选择所需列表项
	function ctrlMove() {
		if (size > 0) {
			var contentObj = document.getElementById("content");//获取content对象
			size = contentObj.childNodes.length; //设置条目数
			switch (event.keyCode) {
			case 13:
				document.getElementById("toolBox").style.visibility = "hidden";//隐藏提示框
				size = 0;
				return false;
				break;
			case 38://向上方向键
				if (curSize == -1) {
					curSize = size - 1;
				} else if (curSize > 0) {
					curSize--;
				}
				break;
			case 40://向下方向键
				if (curSize == -1) {
					curSize = 0;
				} else if (curSize < size - 1) {
					curSize++;
				}
				break;
			}
			if (event.keyCode == 38 || event.keyCode == 40) {
				contentObj.childNodes[curSize].style.background = "#009900";//当鼠标在关键词上时改变背景色 
				contentObj.childNodes[curSize].style.color = "#FFFFFF";//当鼠标在关键词上时改变文字颜色
				document.getElementById("suggest").value = contentObj.childNodes[curSize].innerHTML;
				for ( var i = 0; i < size; i++) {
					if (i != curSize) {
						contentObj.childNodes[i].style.background = "#FFFFFF";//当鼠标在关键词上时改变背景色 
						contentObj.childNodes[i].style.color = "#000000";//当鼠标在关键词上时改变文字颜色
					}
				}
			}
		}
	}
	//在content中添加一条信息 
	function addContent(content) {
		var contentObj = document.getElementById("content");//获取content对象
		var divContent = document.createElement("div");//创建一个DIV对象
		divContent.innerHTML = content;//设置div标记的内容
		divContent.style.paddingLeft = "2px";//设置文字与左边框的距离
		divContent.onmouseover = function() {//添加鼠标移入事件
			this.style.background = "#009900";//当鼠标在关键词上时改变背景色 
			this.style.color = "#FFFFFF";//当鼠标在关键词上时改变文字颜色
		};
		divContent.onmouseout = function() {//添加鼠标移出事件
			this.style.background = "#009900";//鼠标移开时变为原来的背景色 
			this.style.color = "#000000";//当鼠标在关键词上时改变文字颜色
		};
		divContent.onclick = function() {//添加单击事件
			document.getElementById("suggest").value = this.innerHTML;
			document.getElementById("toolBox").style.visibility = "hidden";
		};
		contentObj.appendChild(divContent);//将对象添加在contentObj对象中 
	}
</script>
</head>
<body>
	<div id="box">
		<div id="top"></div>
		<div id="center"
			style="float: left; width:753px; height: 300px; background-repeat:no-repeat; background-position:top center;">
			<div style="position: absolute;padding-left: 100px;">
				<ul>
					<form action="KeyServlet?action=saveKey" name="form1" method="post"
						style="margin:0px;">
						<li style="padding-right:5px;"><input type="text"
							id="suggest" name="suggest" class="inputText"
							onpropertychange="getSuggest();" autocomplete="off">
						</li>
						<li><input type="submit" name="button" class="inputButton"
							value=""></li>
						<li style="color: skyblue; clear:left;"><b><font
								color="skyblue">北極狐热门搜索：</font> </b>QQ
							show    QQ game     QQ
							飞车    <br>
							<div id="toolBox" align="left">
								<span id="content"></span>
							</div>
						</li>
					</form>
				</ul>
			</div>
		</div>
	</div>
</body>
</html>

/** request.js
 * 构建XMLHttpRequest对象并请求服务器
 * 
 * @param reqType：请求类型（GET或POST）
 * @param url：服务器地址
 * @param async：是否异步请求
 * @param resFun：响应的回调函数
 * @param parameter
 *            :请求参数
 * @return :XMLHttpRequest对象
 */
function httpRequest(reqType, url, async, resFun, parameter) {
	var request = null;
	if (window.XMLHttpRequest) { // 非IE浏览器，创建XMLHttpRequest对象
		request = new XMLHttpRequest();
	} else if (window.ActiveXObject) { // IE浏览器，创建XMLHttpRequest对象
		var arrSignatures = [ "Msxml2.XMLHTTP", "Microsoft.XMLHTTP",
				"Microsoft.XMLHTTP", "MSXML2.XMLHTTP.5.0",
				"MSXML2.XMLHTTP.4.0", "MSXML2.XMLHTTP.3.0", "MSXML2.XMLHTTP" ];
		for ( var i = 0; i < arrSignatures.length; i++) {
			request = new ActiveXObject(arrSignatures[i]);
			if (request || typeof (request) == "object")
				break;
		}
	}
	if (request || typeof (request) == "object") {
		if (reqType.toLowerCase() == "post") { // 以POST方式提交
			request.open(reqType, url, true); // 打开服务器连接
			// 设置MIME类型
			request.setRequestHeader("Content-Type",
					"application/x-www-form-urlencoded");
			request.onreadystatechange = resFun;// 设置处理响应的回调函数
			parameter = encodeURI(parameter); // 将参数字符串进行编码
			request.send(parameter); // 发送请求
		} else { // 以GET方式提交
			url = url + "?" + parameter; // 重新设置URL，GET方式需要在URL中附带请求参数
			request.open(reqType, url, true); // 打开服务器连接
			request.onreadystatechange = resFun;// 响应回调函数
			request.send(null); // 发送请求
		}
	} else {
		alert("该浏览器不支持Ajax！");
	}
	return request;
}