lucenes
成功将99收藏夹实现基于Lucene的站内全文搜索99收藏夹原来的站内收藏搜索是基于数据库的like%%语句,发现速度有点慢,于是想到用Lucene来实现站内收藏的按名字和共享收藏的连接标题来搜索的功能,昨天下午做好了,但是结果有点牵强人意。
首先我用Lucene把数据从数据库导出建立索引,代码如下,注:里面的ChinaTokene方法是调用www.sj110.com的中文分词组件
1public class IntranetIndexer
2{
3 private IndexWriter writer;
4 public IntranetIndexer(string dictory)
5 {
6 if (Directory.GetFiles(dictory).Length == 0)
7 {
8 writer = new IndexWriter(dictory, new WhitespaceAnalyzer(), true);
9 }
10 else
11 {
12 writer = new IndexWriter(dictory,new WhitespaceAnalyzer(),false);
13 }
14 writer.SetUseCompoundFile(true);
15 }
16 public void AddDataReaderToIndex(SqlDataReader dr)
17 {
18 if (dr.HasRows)
19 {
20 string EndId="0";
21 while (dr.Read())
22 {
23 /**//*int fieldCount = dr.FieldCount;
24 for (int i = 0; i < fieldCount; i++)
25 {
26 Document doc = new Document();
27 doc.Add(Field.Text(dr.GetName(i).ToString(),ChinaTokene(dr[i].ToString())));
28 writer.AddDocument(doc);
29 EndId = dr[0].ToString();
30 }*/
31 Document doc = new Document();
32 doc.Add(Field.Keyword("this_url",dr["this_url"].ToString()));
33 doc.Add(Field.Keyword("this_name",dr["this_name"].ToString()));
34 doc.Add(Field.Text("this_title", ChinaTokene(dr["this_title"].ToString())));
35 doc.Add(Field.UnIndexed("all_title",dr["this_title"].ToString()));
36 writer.AddDocument(doc);
37 EndId = dr["this_id"].ToString();
38 }
39 using (StreamWriter sw = new StreamWriter(@"C:\Inetpub\wwwroot\pwqdream\endid.txt"))
40 {
41 sw.WriteLine(EndId);
42 }
43 }
44 }
45
46 private string ChinaTokene(string s)
47 {
48 System.Text.StringBuilder sb = new System.Text.StringBuilder();
49 List<string> results = Sj110.Com.Chinese.Tokenizer.Tokenize(s);
50 foreach (string str in results)
51 {
52 sb.AppendFormat("{0} ",str);
53 }
54 sb = sb.Remove(sb.Length-2,1);
55 return sb.ToString();
56 }
57 public void Close()
58 {
59 writer.Optimize();
60 writer.Close();
61 }
当然,最后面要把最后的数据库的那条记录的ID写入文本文件
因为在前面的得到DataReader的时候我们首先是根据这个文本文件来获得最后条记录的ID再查询的
代码如下
1protected void buttonAddIndex_Click(object sender, EventArgs e)
2 {
3 int endId;
4 using (StreamReader sr = new StreamReader(@"C:\Inetpub\wwwroot\pwqdream\endid.txt"))
5 {
6 try
7 {
8 endId = Convert.ToInt32(sr.ReadLine());
9 if (endId < 0)
10 endId = 0;
11 }
12 catch
13 {
14 endId = 0;
15 }
16 }
17 string connStr = ConfigurationManager.AppSettings[0].ToString();
18 SqlConnection conn = new SqlConnection(connStr);
19 string selStr = "select this_id,this_title,this_url,this_name from userFav where this_id>@this_id";
20 SqlCommand comm = new SqlCommand(selStr,conn);
21 comm.Parameters.AddWithValue("@this_id",endId);
22 conn.Open();
23 SqlDataReader dr = comm.ExecuteReader();
24 IntranetIndexer writer = new IntranetIndexer(@"C:\Inetpub\wwwroot\pwqdream\index");
25 writer.AddDataReaderToIndex(dr);
26 writer.Close();
27 dr.Close();
28 dr.Dispose();
29 conn.Close();
30 }
查询的时候用的是ajax调用后台方法,多字段搜索,且都先分词了,结果很牵强,如果输入的是名字,那么取不到url和url,如果输入的象标题,则取不到url和名字,还有界面很不友好,下面是代码:
1function showUserShellFav(obj,pageNumber)
2{
3 var selValue = document.getElementById("searchSel").value;
4 if(obj!="")
5 {
6 var t = document.getElementById("titleTd");
7 //var c = document.getElementById("contentTd");
8 t.innerHTML="搜索关于--<font color='red'>"+obj+"</font>--的共享收藏";
9 //c.innerHTML="";
10 showLoad();
11 PwqzcDream.test.SelectUserShellFav(obj,pageNumber,selValue,onSelectUserShellFavCom);
12 }
13}
14function onSelectUserShellFavCom(rel)
15{
16 hideLoad();
17 var pageIndexTd = document.getElementById("pageIndexTd");
18 var c = document.getElementById("contentTd");
19 var dtContent = rel.value.Tables[0];
20 var dtCount = rel.value.Tables[1];
21 pageIndexTd.innerHTML = "";
22 for(var r=0;r<dtCount.Rows.length;r++)
23 {
24 var cu = dtCount.Rows[r];
25 pageIndexTd.innerHTML += cu.html;
26 }
27 if(dtContent.Rows.length>0)
28 {
29 c.innerHTML = "";
30 if(PwqzcDream.MyDefault.IsLogin().value)
31 {
32 for(var row=0;row<dtContent.Rows.length;row++)
33 {
34 var cur = dtContent.Rows[row];
35 c.innerHTML+="<li><a type='"+cur.this_url+"' style='color:#0088e4;position:relative;cursor:pointer;' onmousedown='MouseDownToMove(this,event);' onmousemove='MouseMoveToMove(event);' onmouseup='MouseUpToMove(event);' onclick='openLink(this.type);'>"+cur.this_title+"</a>  <font color='red'>"+cur.this_name+"</font>收藏</li>";
36 }
37 }
38 else
39 {
40 for(var row=0;row<dtContent.Rows.length;row++)
41 {
42 var cur = dtContent.Rows[row];
43 c.innerHTML+="<li><a href='"+cur.this_url+"' style='color:#0088e4;' target='_blank'>"+cur.this_title+"</a>  <font color='red'>"+cur.this_name+"</font>收藏</li>";
44 }
45 }
46 }
47 else
48 {
49 c.innerHTML = "对不起,没有找到相关的收藏!";
50 }
51}
下面是按照用户名字搜索结果的抓图,当然,速度提升是很明显的
下面是按照标题搜索的结果的抓图:
同时也希望大家能够喜欢99收藏夹,谢谢大家!
首先我用Lucene把数据从数据库导出建立索引,代码如下,注:里面的ChinaTokene方法是调用www.sj110.com的中文分词组件
1public class IntranetIndexer
2{
3 private IndexWriter writer;
4 public IntranetIndexer(string dictory)
5 {
6 if (Directory.GetFiles(dictory).Length == 0)
7 {
8 writer = new IndexWriter(dictory, new WhitespaceAnalyzer(), true);
9 }
10 else
11 {
12 writer = new IndexWriter(dictory,new WhitespaceAnalyzer(),false);
13 }
14 writer.SetUseCompoundFile(true);
15 }
16 public void AddDataReaderToIndex(SqlDataReader dr)
17 {
18 if (dr.HasRows)
19 {
20 string EndId="0";
21 while (dr.Read())
22 {
23 /**//*int fieldCount = dr.FieldCount;
24 for (int i = 0; i < fieldCount; i++)
25 {
26 Document doc = new Document();
27 doc.Add(Field.Text(dr.GetName(i).ToString(),ChinaTokene(dr[i].ToString())));
28 writer.AddDocument(doc);
29 EndId = dr[0].ToString();
30 }*/
31 Document doc = new Document();
32 doc.Add(Field.Keyword("this_url",dr["this_url"].ToString()));
33 doc.Add(Field.Keyword("this_name",dr["this_name"].ToString()));
34 doc.Add(Field.Text("this_title", ChinaTokene(dr["this_title"].ToString())));
35 doc.Add(Field.UnIndexed("all_title",dr["this_title"].ToString()));
36 writer.AddDocument(doc);
37 EndId = dr["this_id"].ToString();
38 }
39 using (StreamWriter sw = new StreamWriter(@"C:\Inetpub\wwwroot\pwqdream\endid.txt"))
40 {
41 sw.WriteLine(EndId);
42 }
43 }
44 }
45
46 private string ChinaTokene(string s)
47 {
48 System.Text.StringBuilder sb = new System.Text.StringBuilder();
49 List<string> results = Sj110.Com.Chinese.Tokenizer.Tokenize(s);
50 foreach (string str in results)
51 {
52 sb.AppendFormat("{0} ",str);
53 }
54 sb = sb.Remove(sb.Length-2,1);
55 return sb.ToString();
56 }
57 public void Close()
58 {
59 writer.Optimize();
60 writer.Close();
61 }
当然,最后面要把最后的数据库的那条记录的ID写入文本文件
因为在前面的得到DataReader的时候我们首先是根据这个文本文件来获得最后条记录的ID再查询的
代码如下
1protected void buttonAddIndex_Click(object sender, EventArgs e)
2 {
3 int endId;
4 using (StreamReader sr = new StreamReader(@"C:\Inetpub\wwwroot\pwqdream\endid.txt"))
5 {
6 try
7 {
8 endId = Convert.ToInt32(sr.ReadLine());
9 if (endId < 0)
10 endId = 0;
11 }
12 catch
13 {
14 endId = 0;
15 }
16 }
17 string connStr = ConfigurationManager.AppSettings[0].ToString();
18 SqlConnection conn = new SqlConnection(connStr);
19 string selStr = "select this_id,this_title,this_url,this_name from userFav where this_id>@this_id";
20 SqlCommand comm = new SqlCommand(selStr,conn);
21 comm.Parameters.AddWithValue("@this_id",endId);
22 conn.Open();
23 SqlDataReader dr = comm.ExecuteReader();
24 IntranetIndexer writer = new IntranetIndexer(@"C:\Inetpub\wwwroot\pwqdream\index");
25 writer.AddDataReaderToIndex(dr);
26 writer.Close();
27 dr.Close();
28 dr.Dispose();
29 conn.Close();
30 }
查询的时候用的是ajax调用后台方法,多字段搜索,且都先分词了,结果很牵强,如果输入的是名字,那么取不到url和url,如果输入的象标题,则取不到url和名字,还有界面很不友好,下面是代码:
1function showUserShellFav(obj,pageNumber)
2{
3 var selValue = document.getElementById("searchSel").value;
4 if(obj!="")
5 {
6 var t = document.getElementById("titleTd");
7 //var c = document.getElementById("contentTd");
8 t.innerHTML="搜索关于--<font color='red'>"+obj+"</font>--的共享收藏";
9 //c.innerHTML="";
10 showLoad();
11 PwqzcDream.test.SelectUserShellFav(obj,pageNumber,selValue,onSelectUserShellFavCom);
12 }
13}
14function onSelectUserShellFavCom(rel)
15{
16 hideLoad();
17 var pageIndexTd = document.getElementById("pageIndexTd");
18 var c = document.getElementById("contentTd");
19 var dtContent = rel.value.Tables[0];
20 var dtCount = rel.value.Tables[1];
21 pageIndexTd.innerHTML = "";
22 for(var r=0;r<dtCount.Rows.length;r++)
23 {
24 var cu = dtCount.Rows[r];
25 pageIndexTd.innerHTML += cu.html;
26 }
27 if(dtContent.Rows.length>0)
28 {
29 c.innerHTML = "";
30 if(PwqzcDream.MyDefault.IsLogin().value)
31 {
32 for(var row=0;row<dtContent.Rows.length;row++)
33 {
34 var cur = dtContent.Rows[row];
35 c.innerHTML+="<li><a type='"+cur.this_url+"' style='color:#0088e4;position:relative;cursor:pointer;' onmousedown='MouseDownToMove(this,event);' onmousemove='MouseMoveToMove(event);' onmouseup='MouseUpToMove(event);' onclick='openLink(this.type);'>"+cur.this_title+"</a>  <font color='red'>"+cur.this_name+"</font>收藏</li>";
36 }
37 }
38 else
39 {
40 for(var row=0;row<dtContent.Rows.length;row++)
41 {
42 var cur = dtContent.Rows[row];
43 c.innerHTML+="<li><a href='"+cur.this_url+"' style='color:#0088e4;' target='_blank'>"+cur.this_title+"</a>  <font color='red'>"+cur.this_name+"</font>收藏</li>";
44 }
45 }
46 }
47 else
48 {
49 c.innerHTML = "对不起,没有找到相关的收藏!";
50 }
51}
下面是按照用户名字搜索结果的抓图,当然,速度提升是很明显的
下面是按照标题搜索的结果的抓图:
同时也希望大家能够喜欢99收藏夹,谢谢大家!