基于Lucene站内搜索(转载)

 

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

 

public   class  IntranetIndexer
  ... {
    private IndexWriter writer;
    public IntranetIndexer(string dictory)
     ...{
         if (Directory.GetFiles(dictory).Length == 0)
         ...{
            writer = new IndexWriter(dictory, new WhitespaceAnalyzer(), true);
         }
        else
        ...{
           writer = new IndexWriter(dictory,new WhitespaceAnalyzer(),false);
       }
       writer.SetUseCompoundFile(true);
    }
    public void AddDataReaderToIndex(SqlDataReader dr)
    ...{
        if (dr.HasRows)
       ...{
            string EndId="0";
            while (dr.Read())
           ...{
               /** *//**//*int fieldCount = dr.FieldCount;
               for (int i = 0; i < fieldCount; i++)
               {
                   Document doc = new Document();
                    doc.Add(Field.Text(dr.GetName(i).ToString(),ChinaTokene(dr[i].ToString())));
                   writer.AddDocument(doc);
                   EndId = dr[0].ToString();
                }*/
                Document doc = new Document();
                doc.Add(Field.Keyword("this_url",dr["this_url"].ToString()));
                doc.Add(Field.Keyword("this_name",dr["this_name"].ToString()));
               doc.Add(Field.Text("this_title", ChinaTokene(dr["this_title"].ToString())));
               doc.Add(Field.UnIndexed("all_title",dr["this_title"].ToString()));
                writer.AddDocument(doc);
                EndId = dr["this_id"].ToString();
            }
            using (StreamWriter sw = new StreamWriter(@"C:Inetpubwwwrootpwqdreamendid.txt"))
            ...{
               sw.WriteLine(EndId);
            }
       }
    }

   private string ChinaTokene(string s)
    ...{
        System.Text.StringBuilder sb = new System.Text.StringBuilder();
       List<string> results = Sj110.Com.Chinese.Tokenizer.Tokenize(s);
       foreach (string str in results)
       ...{
            sb.AppendFormat("{0} ",str);
        }
        sb = sb.Remove(sb.Length-2,1);
      return sb.ToString();
    }
   public void Close()
   ...{
        writer.Optimize();
       writer.Close();
    }

 

 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

protected   void  buttonAddIndex_Click(object sender, EventArgs e)
    ... {
        int endId;
     using (StreamReader sr = new StreamReader(@"C:Inetpubwwwrootpwqdreamendid.txt"))
        ...{
          try
             ...{
                endId = Convert.ToInt32(sr.ReadLine());
                 if (endId < 0)
                    endId = 0;
            }
            catch
            ...{
                endId = 0;
            }
        }
        string connStr = ConfigurationManager.AppSettings[0].ToString();
       SqlConnection conn = new SqlConnection(connStr);
       string selStr = "select this_id,this_title,this_url,this_name from userFav where this_id>@this_id";
        SqlCommand comm = new SqlCommand(selStr,conn);
        comm.Parameters.AddWithValue("@this_id",endId);
       conn.Open();
       SqlDataReader dr = comm.ExecuteReader();
        IntranetIndexer writer = new IntranetIndexer(@"C:Inetpubwwwrootpwqdreamindex");
        writer.AddDataReaderToIndex(dr);
        writer.Close();
        dr.Close();
        dr.Dispose();
       conn.Close();
   }

 

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：

function  showUserShellFav(obj,pageNumber)
  ... {
   var selValue = document.getElementById("searchSel").value;
    if(obj!="")
    ...{
       var t = document.getElementById("titleTd");
       //var c = document.getElementById("contentTd");
       t.innerHTML="搜索关于--<font color='red'>"+obj+"</font>--的共享收藏";
       //c.innerHTML="";
      showLoad();
      PwqzcDream.test.SelectUserShellFav(obj,pageNumber,selValue,onSelectUserShellFavCom);
   }
}
function  onSelectUserShellFavCom(rel)
... {
   hideLoad();
   var pageIndexTd = document.getElementById("pageIndexTd"); 
  var c = document.getElementById("contentTd");
  var dtContent = rel.value.Tables[0];
   var dtCount = rel.value.Tables[1];
   pageIndexTd.innerHTML = "";
   for(var r=0;r<dtCount.Rows.length;r++)
  ...{
     var cu = dtCount.Rows[r];
      pageIndexTd.innerHTML += cu.html;
   }
   if(dtContent.Rows.length>0)
   ...{
      c.innerHTML = "";
     if(PwqzcDream.MyDefault.IsLogin().value)
      ...{
        for(var row=0;row<dtContent.Rows.length;row++)
        ...{
             var cur = dtContent.Rows[row];
            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>&nbsp;&nbsp<font color='red'>"+cur.this_name+"</font>收藏</li>";
         }
      }
      else
      ...{
         for(var row=0;row<dtContent.Rows.length;row++)
         ...{
             var cur = dtContent.Rows[row];
            c.innerHTML+="<li><a href='"+cur.this_url+"' style='color:#0088e4;' target='_blank'>"+cur.this_title+"</a>&nbsp;&nbsp<font color='red'>"+cur.this_name+"</font>收藏</li>";
         }
      }
   }
   else
   ...{
      c.innerHTML = "对不起，没有找到相关的收藏!";
   }
}