C#调用ICTCLAS30

SharpICTCLAS虽说是开源的，但07年以后就没有人再进行维护，跑搜狗的语料问题不少，就连C#版本的作者也承认有不少问题。

想得到更为准确的分词结果，还是研究SharpICTCLAS3.0，也就是ICTCLAS 2009版。dll 文件不是C#开发的，所以引入要通过DllImport

先自己写了个类

using System;
using System.Collections.Generic;
using System.Text;
using System.Runtime.InteropServices;
namespace test
{
    [StructLayout(LayoutKind.Explicit)]
    public struct result_t
    {
        [FieldOffset(0)]
        public int start;
        [FieldOffset(4)]
        public int length;
        [FieldOffset(8)]
        public int sPos;
        [FieldOffset(12)]
        public int sPosLow;
        [FieldOffset(16)]
        public int POS_id;
        [FieldOffset(20)]
        public int word_ID;
        [FieldOffset(24)]
        public int word_type;
        [FieldOffset(28)]
        public int weight;

    }
    class ICTCLAS30
    {
        const string path = @"ICTCLAS30.dll";

        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_Init")]
        public static extern bool Init(String sInitDirPath);

        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_ParagraphProcess")]
        public static extern String ParagraphProcess(String sParagraph, int bPOStagged);

        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_Exit")]
        public static extern bool Exit();

        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_ImportUserDict")]
        public static extern int ImportUserDict(String sFilename);

        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_FileProcess")]
        public static extern bool FileProcess(String sSrcFilename, String sDestFilename, int bPOStagged);

        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_FileProcessEx")]
        public static extern bool FileProcessEx(String sSrcFilename, String sDestFilename);

        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_GetParagraphProcessAWordCount")]
        public static extern int GetParagraphProcessAWordCount(String sParagraph);
        //ICTCLAS_GetParagraphProcessAWordCount
        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_ParagraphProcessAW")]
        public static extern void ParagraphProcessAW(int nCount, [Out, MarshalAs(UnmanagedType.LPArray)] result_t[] result);

        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_AddUserWord")]
        public static extern int AddUserWord(String sWord);

        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_SaveTheUsrDic")]
        public static extern int SaveTheUsrDic();


        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_DelUsrWord")]
        static extern int DelUsrWord(String sWord);

        public ICTCLAS30()
        {

        }


    }
}

 

 

调用：

 if (!ICTCLAS30.Init(null))
            {
                System.Console.WriteLine("Init ICTCLAS failed!");
                return;
            }
            System.Console.WriteLine("Init ICTCLAS Success!");
            String pResult;
            pResult = ICTCLAS30.ParagraphProcess("点击下载超女纪敏佳深受观众喜爱。禽流感爆发在非典之后。", 1);
            System.Console.WriteLine(pResult);
            ICTCLAS30.Exit();

  

注：

使用的时候把ICTCLAS30.dll，Configure.xml和Data文件夹copy到程序exe运行的位置，否则需要制定他们的位置。

 

 

如何把pResult搞成昨天博文里的wordResult格式还是个问题。还需要好好研究。。。。

 

 附上：官方网站的C#调用示例

 

using System;
using System.IO;
using System.Runtime.InteropServices;
namespace win_csharp
{
    [StructLayout(LayoutKind.Explicit)]
    public struct result_t
    {
        [FieldOffset(0)]
        public int start;
        [FieldOffset(4)]
        public int length;
        [FieldOffset(8)]
        public int POS_id;
        [FieldOffset(12)]
        public int word_ID;
    }
    /// <summary>
    /// Class1 的摘要说明。
    /// </summary>
    class Class1
    {
        const string path = @"ICTCLAS30.dll";
        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_Init")]
        public static extern bool ICTCLAS_Init(String sInitDirPath);
        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_ParagraphProcess")]
        public static extern String ICTCLAS_ParagraphProcess(String sParagraph, int bPOStagged);
        [DllImport(path, CharSet = CharSet.Ansi, EntryPoint = "ICTCLAS_Exit")]
        public static extern bool ICTCLAS_Exit();

        /// <summary>
        /// 应用程序的主入口点。
        /// </summary>
        [STAThread]
        static void Main(string[] args)
        {
            //
            // TODO: 在此处添加代码以启动应用程序
            //
            if (!ICTCLAS_Init(null))
            {
                System.Console.WriteLine("Init ICTCLAS failed!");
                return;
            }
            System.Console.WriteLine("Init ICTCLAS Success!");
            String pResult;
            pResult = ICTCLAS_ParagraphProcess("点击下载超女纪敏佳深受观众喜爱。禽流感爆发在非典之后。", 1);
            System.Console.WriteLine(pResult);
            ICTCLAS_Exit();
        }
    }
}

 

 参考：1. format Blog   http://www.cnblogs.com/format

            2.ICTCLAS 官方文档： http://ictclas.org/zy_2.html

            

 

 关于SharpICTCLAS，请参考，修改了部分bug  http://www.cnblogs.com/birdshover/archive/2009/04/03/1122305.html#1494633 

 

      http://blog.csdn.net/lgnlgn/archive/2009/07/24/4377116.aspx( ICTCLAS30做的lucene.net分词接口 )