private void btnUpdateIndex_Click(object sender, EventArgs e)
{
string id = txtFileId.Text.Trim();
if (string.IsNullOrEmpty(id) == true)
{
SetOutput("请输入文件id(整数)");
return;
}
string filename = "think in lucene......";
if (string.IsNullOrEmpty(id))
{
return;
}
bool enableCreate = IsEnableCreated();//是否已经创建索引文件
Term term = new Term("id", id);
Document doc = new Document();
doc = new Document();//创建文档，给文档添加字段，并把文档添加到索引书写器里
doc.Add(new Field("id", id, Field.Store.YES, Field.Index.TOKENIZED));//存储且索引
doc.Add(new Field("filename", filename, Field.Store.YES, Field.Index.TOKENIZED));
doc.Add(new Field("contents", filename, Field.Store.YES, Field.Index.TOKENIZED));
LuceneIO.Directory directory = LuceneIO.FSDirectory.GetDirectory(INDEX_STORE_PATH, enableCreate);
IndexWriter writer = new IndexWriter(directory, new StandardAnalyzer(),IndexWriter.MaxFieldLength.LIMITED);
writer.UpdateDocument(term, doc);
writer.Optimize();
//writer.Commit();
writer.Close();
directory.Close();
SetOutput(string.Format("更新索引.Id:{0}，已经优化成功", id));
}
public static void PrepareIndex(bool isPangu)
{
Analyzer analyzer = null;
if (isPangu)
{
analyzer = new PanGuAnalyzer();//盘古Analyzer
}
else
{
analyzer = new StandardAnalyzer(Version.LUCENE_29);
}
//测试文章测试
DirectoryInfo dirInfo = Directory.CreateDirectory(Config.INDEX_STORE_PATH);
LuceneIO.Directory directory = LuceneIO.FSDirectory.Open(dirInfo);
IndexWriter writer = new IndexWriter(directory, analyzer, true, IndexWriter.MaxFieldLength.LIMITED);
CreateIndex(writer, "jeffreyzhao", "博客园有一个老赵，人格魅力巨大，洋名就叫jeffreyzhao。据我所知，他还是一个胖子，一个钢琴业余爱好者。");
CreateIndex(writer, "lucene测试", "这是一个测试，关于lucene.net的 关注老赵");
CreateIndex(writer, "博客园里有牛人", "Hello World. 我认识的一个高手，他拥有广博的知识，有极客的态度，还经常到园子里来看看");
CreateIndex(writer, "奥巴马", "美国现任总统是奥巴马？确定不是奥巴牛和奥巴羊 不知道问老赵");
CreateIndex(writer, "奥林匹克", "奥林匹克运动会将来到南美美丽热情的国度巴西，也就是亚马逊河流域的一个地方");
CreateIndex(writer, "写给自己", "博客园的jeffwong，新的开始，继续努力了");
writer.Optimize();
writer.Close();
}
private static IndexSearcher searcher; //索引搜索器
#endregion
#region methods
public static IndexSearcher GenerateSearcher()
{
DirectoryInfo dirInfo = Directory.CreateDirectory(Config.INDEX_STORE_PATH);
LuceneIO.Directory directory = LuceneIO.FSDirectory.Open(dirInfo);
searcher = new IndexSearcher(directory, true);
return(searcher);
}
///// <summary>
///// 删除索引所保存的文件夹
///// </summary>
///// <param name="sender"></param>
///// <param name="e"></param>
//private void btnDeleteAllIndex_Click(object sender, EventArgs e)
//{
// DeleteFiles(INDEX_STORE_PATH);
// SetOutput(string.Format("{0}文件夹保存的索引已经全部删除", INDEX_STORE_PATH));
//}
private void btnDeleteIndex_Click(object sender, EventArgs e)
{
string id = txtFileId.Text.Trim();
if (string.IsNullOrEmpty(id) ==true)
{
SetOutput("请输入文件id(整数)");
return;
}
LuceneIO.Directory directory = LuceneIO.FSDirectory.GetDirectory(INDEX_STORE_PATH, false);
IndexModifier modifier = new IndexModifier(directory, new StandardAnalyzer(), false);
Term term = new Term("id", id);
modifier.DeleteDocuments(term);//删除
modifier.Close();
directory.Close();
SetOutput(string.Format("删除文件索引成功,ID为{0}！", id));
}
//创建索引
private void button3_Click(object sender, EventArgs e)
{
bool isPangu = true;
wznr_Servise wznr = new wznr_Servise();
Analyzer analyzer = new PanGuAnalyzer();//盘古Analyzer
DirectoryInfo dirInfo = Directory.CreateDirectory(Config.INDEX_STORE_PATH);
LuceneIO.Directory directory = LuceneIO.FSDirectory.Open(dirInfo);
IndexWriter writer = new IndexWriter(directory, analyzer, true, IndexWriter.MaxFieldLength.LIMITED);
DataTable dt = wznr.GetDataTable("SELECT Title, p FROM test4Table");
for (int i = 0; i < dt.Rows.Count; i++)
{
string title = Convert.ToString(dt.Rows[i][0]);
string content = Convert.ToString(dt.Rows[i][1]);
CreateIndex(writer, title, content);
}
writer.Optimize();
writer.Close();
this.richTextBox1.Text = string.Format("{0}索引创建成功", isPangu ? "盘古分词" : string.Empty);
}
//创建索引导入数据2012-5-29添加--修改类（弃用）
public void CreateIndexImport(GetTotalDelegate getTotalRecordsDelegate, System.Diagnostics.Stopwatch sw, System.Windows.Forms.RichTextBox rich)
{
int totalRecords = 0;
FenCiHelper fch = new FenCiHelper();
wznr_Servise wznr = new wznr_Servise();
DataTable dt = wznr.GetOldbDataTable("select 标题,内容 from Content"); //wznr.GetDataTable1("SELECT Title, Content FROM News");
dt.Columns.AddRange(new DataColumn[] { new DataColumn("keys", typeof(string)), new DataColumn("bztype", typeof(string)) }); //添加新列
DataRow dr = dt.NewRow(); //实例化新行
DataTable newdt = wznr.GetTableSchema();
DataRow newdr;
string line;
string p; //每个p标签
for (int i = 0; i < dt.Rows.Count; i++) //添加新列分词录入
{
totalRecords++;
getTotalRecordsDelegate(totalRecords);
string title = Convert.ToString(dt.Rows[i][0]); //获取到标题
string content = Convert.ToString(dt.Rows[i][1]); //获取到内容
//string bztype = Convert.ToString(dt.Rows[i][2]);//类别
//调用PanguFenCi（aa）进行分词添加到datatable中批量录入到关键词表中。
title = fch.PanguFenCi(title);
//p标签获取获取每个p---2012-5-18日修改
//p = GetPhtml(content);
//获取路径,循环每次都需要读取文本文件里设置的关键词
string path = Directory.GetCurrentDirectory();
string txtpath = path + @"\App_Data\sDict.txt";
//读取文本内容逐行
StreamReader file = new StreamReader(txtpath);
//分类操作，读取文本与标题分词判断
string[] arr = title.Split('/');
while ((line = file.ReadLine()) != null)
{
string[] arrtxt = line.Split(',');
for (int j = 0; j < arr.Length; j++)
{
if (arrtxt[0].Equals(arr[j]))
{
dt.Rows[i][3] += arrtxt[1] + ",";//类别
break;
//strtxt = arrtxt[1];//所属类别
}
}
}
//调用PanGuContentFenCi（content）进行分词同时输出同义词
dt.Rows[i][1] = fch.PanGuContentFenCi(content);
dt.Rows[i][2] = title;
//dt.Rows[i][3] = bztype;
Analyzer analyzer = new PanGuAnalyzer();//盘古Analyzer
DirectoryInfo dirInfo = Directory.CreateDirectory(Config.INDEX_STORE_PATH);
LuceneIO.Directory directory = LuceneIO.FSDirectory.Open(dirInfo);
IndexWriter writer = new IndexWriter(directory, analyzer, true, IndexWriter.MaxFieldLength.LIMITED);
foreach (Match m in Regex.Matches(content, @"<(\w+)>[^P]*[^<]*</(\w+)>"))
{
newdr = newdt.NewRow();
newdr["Title"] = dt.Rows[i][0];
newdr["p"] = m.Value;
newdr["p1"] = content;
newdr["keys"] = title;
newdr["bztype"] = dt.Rows[i][3];
newdt.Rows.Add(newdr);
// CreateIndex(writer, dt.Rows[i][0].ToString(), m.Value);创建索引
}
writer.Optimize();
writer.Close();
rich.Text += dt.Rows[i][2].ToString() + "-----索引创建成功\n";
rich.ForeColor = System.Drawing.Color.Green;//ConsoleColor.Green;
}
string sql = "insert into test3Table (Title,Content,keys,bztype)" +
" SELECT nc.Title,nc.Content,nc.keys,nc.bztype" +
" FROM @NewBulkTestTvp AS nc";
string sqlp = "insert into test4Table (Title,p,p1,keys,bztype)" +
" SELECT nc.Title,nc.p,nc.p1,nc.keys,nc.bztype" +
" FROM @NewBulkTestTvp AS nc";
sw.Start();
wznr.TableValuedToDB(dt, sql, "dbo.test3Udt");
wznr.TableValuedToDB(newdt, sqlp, "dbo.test4Udt");
sw.Stop();
getTotalRecordsDelegate(totalRecords);
}
