public void IssuesTest_20()
{
string text = "A10021003吃饭";
var keywords = new string[] { "1", "A", "2", "0", "吃" };
var iws = new IllegalWordsSearch();
var ss = new StringSearch();
var sse = new StringSearchEx2();
iws.SetKeywords(keywords);
iws.UseIgnoreCase = true;
iws.UseDBCcaseConverter = true;
var iwsFirst = iws.FindFirst(text);
Assert.AreEqual("吃", iwsFirst.Keyword);
var iwsAll = iws.FindAll(text);
Assert.AreEqual(1, iwsAll.Count);// 因为1A20左右都是英文或数字，所以识别失败
ss.SetKeywords(keywords);
var ssFirst = ss.FindFirst(text);
Assert.AreEqual("A", ssFirst);
var ssAll = ss.FindAll(text);
Assert.AreEqual(9, ssAll.Count);
sse.SetKeywords(keywords);
var sseFirst = sse.FindFirst(text);
Assert.AreEqual("A", sseFirst);
var sseAll = sse.FindAll(text);
Assert.AreEqual(9, sseAll.Count);
}
public void test3()
{
string s = "中国|国人|zg人";
string test = "我是中国人";
StringSearchEx2 iwords = new StringSearchEx2();
iwords.SetKeywords(s.Split('|').ToList());
var b = iwords.ContainsAny(test);
Assert.AreEqual(true, b);
var f = iwords.FindFirst(test);
Assert.AreEqual("中国", f);
var all = iwords.FindAll(test);
Assert.AreEqual("中国", all[0]);
Assert.AreEqual("国人", all[1]);
Assert.AreEqual(2, all.Count);
var str = iwords.Replace(test, '*');
Assert.AreEqual("我是***", str);
}
static List <string> ReadBadWord()
{
List <string> list = new List <string>();
using (StreamReader sw = new StreamReader(File.OpenRead("BadWord.txt"))) {
string key = sw.ReadLine();
while (key != null)
{
if (key != string.Empty)
{
tf1.AddKey(key);
ff.AddKey(key);
list.Add(key);
}
key = sw.ReadLine();
}
}
stringSearch.SetKeywords(list);
stringSearchEx.SetKeywords(list);
stringSearchEx2.SetKeywords(list);
stringSearchEx3.SetKeywords(list);
wordsSearch.SetKeywords(list);
wordsSearchEx.SetKeywords(list);
wordsSearchEx2.SetKeywords(list);
wordsSearchEx3.SetKeywords(list);
//illegalWordsSearch.SetKeywords(list);
list = list.OrderBy(q => q).ToList();
var str = string.Join("|", list);
str = Regex.Replace(str, @"([\\\.\+\*\-\(\)\[\]\{\}!])", @"\$1");
re = new Regex(str, RegexOptions.IgnoreCase);
var str2 = tf1.ToString();
//str2 = Regex.Replace(str2, @"([\.\+\*\-\[\]\{\}!])", @"\$1");
re2 = new Regex(str2);
return(list);
}
public void test4()
{
string s = "中国人|中国|国人|zg人|我是中国人|我是中国|是中国人";
string test = "我是中国人";
StringSearchEx2 iwords = new StringSearchEx2();
iwords.SetKeywords(s.Split('|').ToList());
var all = iwords.FindAll(test);
Assert.AreEqual(6, all.Count);
var str = iwords.Replace(test, '*');
Assert.AreEqual("*****", str);
}