您的位置:首页 > 其它

.NET开发过程中的全文索引使用技巧之Solr

2013-11-28 10:45 267 查看
前言:相信许多人都听说过.net开发过程中基于Lucene.net实现的全文索引,而Solr是一个高性能,基于Lucene的全文搜索服务器。同时对其进行了扩展,提供了比Lucene更为丰富的查询语言,同时实现了可配置、可扩展并对查询性能进行了优化,并且提供了一个完善的功能管理界面,是一款非常优秀的全文搜索引引擎,这里我就绕过Lucene,直接说Solr的应用了,总之,Solr比Lucene更加方便简洁好用,而且上手快,开发效率高。

Solr应用场景:涉及到大数据的全文搜索。尤其是电子商务平台还有现在流行的云计算,物联网等都是需要强大的数据量作为支撑的,使用Solr来进行数据检索最合适不过了,而且Solr是免费开源的,门槛低、投资少见效快。关于Solr的一些优点我这里就不在累赘陈述了,园子里也有很多大神也写了很多关于Solr的技术博文,我这里也只是抛砖引玉,见笑了。

好了,这里就开始Solr的奇幻之旅吧

基于.NET平台下的Solr开发步骤

一、搭建Solr服务器,具体步骤如下:

1.安装JDK,因为是.NET平台,不需要安装JRE、JAVA虚拟机,只安装JDK即可,而且安装JDK不需要手动去配置环境变量,它会自动帮我们配置好环境变量,很方便,这里我安装的是jdk1.7,官网地址:http://www.oracle.com/technetwork/java/javase/downloads/index.html

2.安装Tomcat8.0,官网地址:http://tomcat.apache.org/download-80.cgi,安装完成后启动Monitor Tomcat,浏览器地址栏输入http://localhost:8080/,能进入说明安装成功

3.下载Solr,这里我用的是Solr4.4版本,下载后进行下列配置

  (1)解压Solr4.4,创建Solr目录,比如D:/SorlServer/one,将解压后的Solr4.4中的example目录下的Solr文件夹中的所有文件拷贝到创建的目录中

  (2)创建Solr Web应用,具体步骤,将解压后的Solr4.4中的dist目录下的Solr-4.4.0.war文件拷贝到Tomcat下,比如C:\Program Files\Apache Software Foundation\Tomcat 7.0\webapps下,重命名为one.war,启动Tomcat后该文件会自动解压,进入到D:\SorlServer\one\collection1\conf下,打开solrconfig.xml文件,找到 <dataDir>节点改为<dataDir>${solr.data.dir:c:/SorlServer/one/data}</dataDir>

注意:这一步很重要:打开C:\Program Files\Apache Software Foundation\Tomcat 7.0\webapps\One\WEB-INF下的web.xml文件,找到<env-entry>节点开启,

将env-entry-value值改为D:/SorlServer/one,如下:

<env-entry>

<env-entry-name>solr/home</env-entry-name>

<env-entry-value>D:/SorlServer/one</env-entry-value>

<env-entry-type>java.lang.String</env-entry-type>

</env-entry>

   (3)将解压后的Solr4.4下的/dist/solrj-lib目录中的所有jar包拷贝到C:\Program Files\Apache Software Foundation\Tomcat 7.0\lib中

  (4)停止Tomcat,然后再启动,访问http://localhost:8080/one,即可打开

注意:如果是开发英文网站,我们就不需要使用第三方的分词配置,Solr本身就内置支持英文分词,如果是其他语种比如小语种(日语、意大利、法语等等),大家可以去网上找相关的分词包,这里我们以中文分词为例,毕竟国内大部分网站都是中文为主的。

[b] 4.配置中文分词,国内常用的分词器(庖丁解牛mmseg4jIKAnalyzer),这里我用的是IKAnalyzer,这个分词器比较活跃而且更新也快,挺好用的,[/b]具体步骤如下:

(1)将IKAnalyzer的jar包以及IKAnalyzer.cfg.xml都复制到C:\Program Files\Apache Software Foundation\Tomcat 7.0\webapps\one\WEB-INF\lib下

(2)配置D:\SorlServer\one\collection1\conf下的schema.xml,添加如下配置:

<!-- 分词配置 -->

<fieldType name="text_IKFENCHI" class="solr.TextField">

<analyzer class="org.wltea.analyzer.lucene.IKAnalyzer"/>

</fieldType>

(3)停止Tomcat,然后再启动,访问http://localhost:8080/one/#/collection1/analysis,即可进行测试

以上是Solr服务器端的相关配置工作

二、开始基于.NET平台的Solr开发:

1.下载Solr客户端组件,我用的是园子里的Terry大哥的[b]EasyNet.Solr,地址在微软开源站:http://easynet.codeplex.com/,[/b]

Terry大哥已经把solr客户端封装的很完善了,里面封装了很多现成的方法和参数配置,我们直接可以拿过来用,利用Easynet.solr创建索引,然后再查询索引,具体使用方法如下:

  (1)下载EasyNet.Solr源码直接放到项目中,也可以将源码生成Dll组件后添加到项目引用进行使用,把源码放到项目中最好不过了,我们也可以对其进行调整来满足自己的需要

  (2)创建索引实体类,就是我们要保存的索引数据,比如创建一个产品实体类   

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using Seek.SearchIndex;
using System.Data;
using System.Threading;
using System.Configuration;
using System.Reflection;
using EasyNet.Solr;
using EasyNet.Solr.Impl;
using EasyNet.Solr.Commons;
using System.Xml.Linq;
using EasyNet.Solr.Commons.Params;
using System.Threading.Tasks;

namespace Seek.SearchIndex
{
/// <summary>
/// 索引器
/// </summary>
public class Indexer
{
private readonly static OptimizeOptions optimizeOptions = new OptimizeOptions();
private readonly static CommitOptions commitOptions = new CommitOptions() { SoftCommit = true };
private readonly static ISolrResponseParser<NamedList, EasyNet.Solr.ResponseHeader> binaryResponseHeaderParser = new BinaryResponseHeaderParser();
private readonly static IUpdateParametersConvert<NamedList> updateParametersConvert = new BinaryUpdateParametersConvert();
private readonly static ISolrQueryConnection<NamedList> connection = new SolrQueryConnection<NamedList>() { ServerUrl = ConfigurationManager.AppSettings["SolrServer"] };
private readonly static ISolrUpdateConnection<NamedList, NamedList> solrUpdateConnection = new SolrUpdateConnection<NamedList, NamedList>() { ServerUrl = ConfigurationManager.AppSettings["SolrServer"], ContentType = "application/javabin" };
private readonly static ISolrUpdateOperations<NamedList> solr = new SolrUpdateOperations<NamedList, NamedList>(solrUpdateConnection, updateParametersConvert) { ResponseWriter = "javabin" };
private readonly static ISolrQueryOperations<NamedList> solrQuery = new SolrQueryOperations<NamedList>(connection) { ResponseWriter = "javabin" };

public enum State
{
/// <summary>
/// 运行中
/// </summary>
Runing,
/// <summary>
/// 停止
/// </summary>
Stop,
/// <summary>
/// 中断
/// </summary>
Break
}
/// <summary>
/// 窗口
/// </summary>
private Main form;
/// <summary>
/// 线程
/// </summary>
public Thread t;
/// <summary>
/// 消息状态
/// </summary>
public State state = State.Stop;
/// <summary>
/// 当前索引
/// </summary>
private long currentIndex = 0;

public long CurrentIndex
{
get { return currentIndex; }
set { currentIndex = value; }
}

private int _startId = AppCongfig.StartId;

public int StartId
{
get { return _startId; }
set { _startId = value; }
}

/// <summary>
/// 产品总数
/// </summary>
private int productsCount = 0;
/// <summary>
/// 起始时间
/// </summary>
private DateTime startTime = DateTime.Now;
/// <summary>
/// 结束时间
/// </summary>
private DateTime endTime = DateTime.MinValue;
private static object syncLock = new object();
#region 单利模式
private static Indexer instance = null;

private Indexer(Main _form)
{
form = _form;
productsCount = DataAccess.GetCount(0);       //产品数统计
form.fullerTsslMaxNum.Text = productsCount.ToString();
form.fullerProgressBar.Minimum = 0;
form.fullerProgressBar.Maximum = productsCount;
}
public static Indexer GetInstance(Main form)
{
if (instance == null)
{
lock (syncLock)
{
if (instance == null)
{
instance = new Indexer(form);
}
}
}
return instance;
}
#endregion

/// <summary>
/// 启动
/// </summary>
public void Start()
{
ThreadStart ts = new ThreadStart(FullerRun);
t = new Thread(ts);
t.Start();
}
/// <summary>
/// 停止
/// </summary>
public void Stop()
{
state = State.Stop;
}
/// <summary>
/// 中断
/// </summary>
public void Break()
{
state = State.Break;
}

/// <summary>
/// 创建索引
/// </summary>
public void InitIndex(object data)
{
var docs = new List<SolrInputDocument>();
DataTable list = data as DataTable;
foreach (DataRow pro in list.Rows)
{
var model = new SolrInputDocument();

PropertyInfo[] properites = typeof(IndexProductModel).GetProperties();//得到实体类属性的集合
string[] dateFields = { "CreateDate", "LastEditDate" };
string field = string.Empty;//存储fieldname
foreach (PropertyInfo propertyInfo in properites)//遍历数组
{
object val = pro[propertyInfo.Name];
if (val != DBNull.Value)
{
model.Add(propertyInfo.Name, new SolrInputField(propertyInfo.Name, val));
}
}
docs.Add(model);

StartId = Convert.ToInt32(pro["ID"]);
}
GetStartId();
lock (syncLock)
{
if (currentIndex <= productsCount)
{
form.fullerProgressBar.Value = (int)currentIndex;
}
form.fullerTsslCurrentNum.Text = currentIndex.ToString();
}
var result = solr.Update("/update", new UpdateOptions() {  Docs = docs });
}

/// <summary>
/// 创建索引
/// </summary>
public void CreateIndexer(DataTable dt)
{
GetStartId();
Parallel.ForEach<DataRow>(dt.AsEnumerable(), (row) =>
{
//从数据库查询商品详细属性
if (row != null)
{
var docs = new List<SolrInputDocument>();
var model = new SolrInputDocument();

PropertyInfo[] properites = typeof(IndexProductModel).GetProperties();//得到实体类属性的集合
string[] dateFields = { "CreateDate", "LastEditDate" };
string field = string.Empty;//存储fieldname
foreach (PropertyInfo propertyInfo in properites)//遍历数组
{
object val = row[propertyInfo.Name];
if (val != DBNull.Value)
{
model.Add(propertyInfo.Name, new SolrInputField(propertyInfo.Name, val));
}
}
docs.Add(model);

StartId = Convert.ToInt32(row["ID"]);
var result = solr.Update("/update", new UpdateOptions() { Docs = docs });
}
});

//GetStartId();
lock (syncLock)
{
if (currentIndex <= productsCount)
{
form.fullerProgressBar.Value = (int)currentIndex;
}
form.fullerTsslCurrentNum.Text = currentIndex.ToString();
}
}

/// <summary>
/// 全部索引运行
/// </summary>
public void FullerRun()
{
//GetStartId();
//form.fullerTsslCurrentNum.Text = currentIndex.ToString();
DataTable dt = DataAccess.GetNextProductsInfo(StartId);
StartId = AppCongfig.StartId;
if (state == State.Break)
{
this.SendMesasge("完全索引已继续,起始ID[" + StartId + "]...");
}
else
{
startTime = DateTime.Now;
this.SendMesasge("完全索引已启动,起始ID[" + StartId + "]...");
}
state = State.Runing;
form.btnInitIndex.Enabled = false;
form.btnSuspend.Enabled = true;
form.btnStop.Enabled = true;

while (dt != null && dt.Rows.Count > 0 && state == State.Runing)
{
try
{
InitIndex(dt);//单线程
// CreateIndexer(dt);//多线程
}
catch (Exception ex)
{
state = State.Stop;
form.btnInitIndex.Enabled = true;
form.btnSuspend.Enabled = false;
form.btnStop.Enabled = false;
GetStartId();
this.SendMesasge(ex.Message.ToString());
}
form.fullerTsslTimeSpan.Text = "已运行 :" + GetTimeSpanShow(DateTime.Now - startTime) + ",预计还需:" + GetTimeSpanForecast();

try
{
dt = DataAccess.GetNextProductsInfo(StartId);//获取下一组产品
}
catch (Exception err)
{
this.SendMesasge("获取下一组产品出错,起始ID[" + StartId + "]:" + err.Message);
}
}
if (state == State.Runing)
{
state = State.Stop;
form.btnInitIndex.Enabled = true;
form.btnSuspend.Enabled = false;
form.btnStop.Enabled = false;
AppCongfig.SetValue("StartId", StartId.ToString());
this.SendMesasge("完全索引已完成,总计索引数[" + currentIndex + "]结束的产品Id" + StartId);
}
else if (state == State.Break)
{
GetStartId();
state = State.Break;
form.btnInitIndex.Enabled = true;
form.btnSuspend.Enabled = false;
form.btnStop.Enabled = false;
AppCongfig.SetValue("StartId", StartId.ToString());
this.SendMesasge("完全索引已暂停,当前索引位置[" + currentIndex + "]结束的产品Id" + StartId);
}
else if (state == State.Stop)
{
GetStartId();
state = State.Stop;
this.SendMesasge("完全索引已停止,已索引数[" + currentIndex + "]结束的产品Id" + StartId);
form.btnInitIndex.Enabled = true;
form.btnSuspend.Enabled = false;
form.btnStop.Enabled = false;
AppCongfig.SetValue("StartId", StartId.ToString());
productsCount = DataAccess.GetCount(StartId);       //产品数统计
form.fullerTsslMaxNum.Text = productsCount.ToString();
form.fullerProgressBar.Minimum = 0;
form.fullerProgressBar.Maximum = productsCount;
}
endTime = DateTime.Now;
}

/// <summary>
/// 多线程构建索引数据方法
/// </summary>
/// <param name="threadDataParam"></param>
public void MultiThreadCreateIndex(object threadDataParam)
{
InitIndex(threadDataParam);
}

/// <summary>
/// 获取最大的索引id
/// </summary>
private void GetStartId()
{
IDictionary<string, ICollection<string>> options = new Dictionary<string, ICollection<string>>();
options[CommonParams.SORT] = new string[] { "ProductID DESC" };
options[CommonParams.START] = new string[] { "0" };
options[CommonParams.ROWS] = new string[] { "1" };
options[HighlightParams.FIELDS] = new string[] { "ProductID" };
options[CommonParams.Q] = new string[] { "*:*" };
var result = solrQuery.Query("/select", null, options);
var solrDocumentList = (SolrDocumentList)result.Get("response");
currentIndex = solrDocumentList.NumFound;
if (solrDocumentList != null && solrDocumentList.Count() > 0)
{
StartId = (int)solrDocumentList[0]["ProductID"];
//AppCongfig.SetValue("StartId", solrDocumentList[0]["ProductID"].ToString());
}
else
{
StartId = 0;
// AppCongfig.SetValue("StartId", "0");
}
}

/// <summary>
/// 优化索引
/// </summary>
public void Optimize()
{
this.SendMesasge("开始优化索引,请耐心等待...");
var result = solr.Update("/update", new UpdateOptions() { OptimizeOptions = optimizeOptions });
var header = binaryResponseHeaderParser.Parse(result);
this.SendMesasge("优化索引耗时:" + header.QTime + "毫秒");
}

/// <summary>
/// 发送消息到界面
/// </summary>
/// <param name="message">发送消息到界面</param>
protected void SendMesasge(string message)
{
form.fullerDgvMessage.Rows.Add(form.fullerDgvMessage.Rows.Count + 1, message, DateTime.Now.ToString());
}
/// <summary>
/// 获取时间间隔显示
/// </summary>
/// <param name="ts">时间间隔</param>
/// <returns></returns>
protected string GetTimeSpanShow(TimeSpan ts)
{
string text = "";
if (ts.Days > 0)
{
text += ts.Days + "天";
}
if (ts.Hours > 0)
{
text += ts.Hours + "时";
}
if (ts.Minutes > 0)
{
text += ts.Minutes + "分";
}
if (ts.Seconds > 0)
{
text += ts.Seconds + "秒";
}
return text;
}
/// <summary>
/// 获取预测时间
/// </summary>
/// <returns></returns>
protected string GetTimeSpanForecast()
{
if (currentIndex != 0)
{
TimeSpan tsed = DateTime.Now - startTime;
double d = ((tsed.TotalMilliseconds / currentIndex) * productsCount) - tsed.TotalMilliseconds;
return GetTimeSpanShow(TimeSpan.FromMilliseconds(d));
}
return "";
}
}
}


View Code
  (5)运行索引器,创建索引,这里是我的索引器界面,如图



  可以随时跟踪索引生成的情况

  (6)索引创建完毕后,可以进入Solr服务器界面http://localhost:8080/one/#/collection1/query进行测试

以上就是Solr的前期工作,主要是Solr服务器搭建和客户端调用生成索引,后期再对客户端的查询进行详细的说明,下期预告

1.全文搜索,分词配置,以及类似于谷歌和百度那种输入关键字自动完成功能

2.Facet查询
内容来自用户分享和网络整理,不保证内容的准确性,如有侵权内容,可联系管理员处理 点击这里给我发消息
标签: