static void BuildDawgFile(string file)
{
var rootPath = AppDomain.CurrentDomain.BaseDirectory.Replace(@"\test\bin\Debug\", "");
var wordUtil = new WordUtil();
//加载默认的词频
using (var sr = new StreamReader(rootPath + @"\dict\cwsharp.freq", Encoding.UTF8))
{
string line = null;
while ((line = sr.ReadLine()) != null)
{
if (line == string.Empty) continue;
var array = line.Split(' ');
wordUtil.Add(array[0], int.Parse(array[1]));
}
}
//加载新的词典
using (var sr = new StreamReader(rootPath + @"\dict\cwsharp.dic", Encoding.UTF8))
{
string line = null;
while ((line = sr.ReadLine()) != null)
{
if (line == string.Empty) continue;
wordUtil.Add(line);
}
}
//保存新的dawg文件
wordUtil.SaveTo(file);
}
public void TestFromTxtFile(string file)
{
var wordUtil = new WordUtil();
var expectWordCount = 0;
using (var sr = new StreamReader(file, Encoding.UTF8))
{
string line = null;
while ((line = sr.ReadLine()) != null)
{
if (line == string.Empty) continue;
wordUtil.Add(line);
expectWordCount++;
}
}
var watcher = new System.Diagnostics.Stopwatch();
watcher.Start();
var ms = new MemoryStream();
wordUtil.SaveTo(ms);
watcher.Stop();
Console.WriteLine("build dawg elapsed time:" + watcher.Elapsed.TotalMilliseconds + "'ms");
watcher.Reset();
watcher.Start();
ms.Position = 0;
wordUtil = WordUtil.LoadFrom(ms);
watcher.Stop();
Console.WriteLine("load dawg file elapsed time:" + watcher.Elapsed.TotalMilliseconds + "'ms");
Assert.AreEqual(expectWordCount, wordUtil.Count);
}
