static void BuildDawgFile(string file)
{
var rootPath = AppDomain.CurrentDomain.BaseDirectory.Replace(@"\test\bin\Debug\", "");
var wordUtil = new WordDict();
//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);
}
static void BuildDawgFile(string file)
{
var rootPath = AppDomain.CurrentDomain.BaseDirectory.Replace(@"\test\bin\Debug\", "");
var wordUtil = new WordDict();
//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 WordDict();
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 = WordDict.LoadFrom(ms);
watcher.Stop();
Console.WriteLine("load dawg file elapsed time:" + watcher.Elapsed.TotalMilliseconds + "'ms");
Assert.AreEqual(expectWordCount, wordUtil.Count);
}
public static void AddWords(this WordDict wordUtil, params string[] words)
{
foreach (var word in words)
{
wordUtil.Add(word);
}
}
/// <summary>
/// Parse word data from a string and add it to the dictionary.
/// </summary>
/// <param name="str"></param>
/// <returns></returns>
public bool AddWord(string str)
{
if (string.IsNullOrEmpty(str))
{
return(false);
}
string[] parts = str.Split(' ');
if (parts?.Length >= 3)
{
var key = parts[0]; // 字詞
var freq = Convert.ToInt32(parts[1]);
// 如果字詞已經存在表中，則比較頻率。若出現頻率大於或等於既有的字詞，則覆蓋之。
WordData word = null;
if (WordDict.ContainsKey(key)) // 字詞已經存在表中
{
word = WordDict[key];
if (freq < word.Frequency)
{
return(false);
}
word.ZhuyinList.Clear();
}
else
{
word = new WordData();
WordDict.Add(key, word);
}
word.Frequency = freq;
for (int i = 2; i < parts.Length; i++)
{
word.ZhuyinList.Add(parts[i]);
}
Logger.Verbose("加入字詞: {Key} {@Word}", key, word);
return(true);
}
Logger.Warning($"無效的字詞: {str}");
return(false);
}
public void TestFromTxtFile(string file)
{
var wordUtil = new WordDict();
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 = WordDict.LoadFrom(ms);
watcher.Stop();
Console.WriteLine("load dawg file elapsed time:" + watcher.Elapsed.TotalMilliseconds + "'ms");
Assert.AreEqual(expectWordCount, wordUtil.Count);
}