/// <summary>
/// 中文分词方法
/// </summary>
/// <param name="text">待分词的文本</param>
/// <returns></returns>
public static List <string> WordSegment(string text)
{
if (text == null || text.Trim().Length <= 0)
{
return(new List <string>(0));
}
if (s_Segment == null)
{
lock (s_Seg_SyncObj)
{
if (s_Segment == null)
{
string path = ConfigurationManager.AppSettings["WordSegmentConfigPath"];
if (path == null || path.Trim().Length <= 0)
{
// 使用默认路径
path = Path.Combine(AppDomain.CurrentDomain.SetupInformation.ApplicationBase, "Configuration\\KTDictSeg.config");
}
else
{
string p = Path.GetPathRoot(path);
if (p == null || p.Trim().Length <= 0) // 说明是相对路径
{
path = Path.Combine(AppDomain.CurrentDomain.SetupInformation.ApplicationBase, path);
}
}
CSimpleDictSeg tmp = new CSimpleDictSeg();
tmp.LoadConfig(path);
tmp.LoadDict();
s_Segment = tmp;
}
}
}
List <T_WordInfo> list = s_Segment.SegmentToWordInfos(text);
List <string> rstList = new List <string>(list.Count);
foreach (T_WordInfo info in list)
{
if (info != null && info.Word.Length > 1)
{
rstList.Add(info.Word);
}
}
return(rstList);
}
private void BigBangDemo()
{
if (m_SimpleDictSeg == null)
{
m_SimpleDictSeg = new CSimpleDictSeg();
m_SimpleDictSeg.LoadConfig("KTDictSeg.xml");
m_SimpleDictSeg.MatchName = true;
m_SimpleDictSeg.FreqFirst = true;
m_SimpleDictSeg.AutoStudy = true;
m_SimpleDictSeg.FilterStopWords = true;
m_SimpleDictSeg.MultiSelect = false;
m_SimpleDictSeg.Redundancy = 1;
m_SimpleDictSeg.LoadDict();
}
List <T_WordInfo> words = m_SimpleDictSeg.SegmentToWordInfos(word);
foreach (T_WordInfo x in words)
{
clbWords.Items.Add(x.ToString());
}
}
public void initDict()
{
//出事话分词词典
m_SimpleDictSeg = new CSimpleDictSeg();
m_SimpleDictSeg.LoadConfig("KTDictSeg.xml");
m_SimpleDictSeg.LoadDict();
}
public void initDict()
{//初始化分词词典
m_SimpleDictSeg = new CSimpleDictSeg();
m_SimpleDictSeg.LoadConfig("KTDictSeg.xml");
m_SimpleDictSeg.LoadDict();
}