public void Given_never_before_seen_text()
{
var detector = new SpamDetector();
var corpus = @"欧拉";
Assert.AreEqual(detector.IsSpam(corpus), CategorizationResult.Undetermined);
}
public void Given_a_corpus_that_is_half_spam_and_half_unknown()
{
var detector = new SpamDetector();
detector.SpamFound(@"做人了");
var result = detector.IsSpam(@"我不做人了！");
Assert.AreEqual(result, CategorizationResult.Undetermined);
}
public void Given_text_marked_as_spam_then_marked_as_ham()
{
var detector = new SpamDetector();
var spamCorpus = @"你要说的下一句话是……";
detector.SpamFound(spamCorpus);
detector.HamFound(spamCorpus);
Assert.AreEqual(detector.IsSpam(spamCorpus), CategorizationResult.Undetermined);
}
public void LongTest()
{
var corpus = @"不愧是DIO！我们不敢做的事，他毫不在乎地做了！真是佩服，真是我们的偶像！";
var currentPath = @".\";
var spamPath = Path.Combine(currentPath, @"Resources", @"spam.txt");
var hamPath = Path.Combine(currentPath, @"Resources", @"ham.txt");
var spamDetector = new SpamDetector();
spamDetector.Train(spamPath, hamPath);
var utf8 = new UTF8Encoding(false);
var spams = File.ReadAllLines(spamPath, utf8);
var hams = File.ReadAllLines(hamPath, utf8);
var r0 = 0;
Parallel.For(72001, 80000, i =>
{
var result = spamDetector.IsSpam(spams[i]);
if (result == CategorizationResult.First)
{
Interlocked.Increment(ref r0);
}
});
var r = r0 / 8000d * 100;
Console.WriteLine($@"{r}%");
Assert.IsTrue(r > 0.97);
var r1 = 0;
Parallel.For(648001, 720000, i =>
{
var result = spamDetector.IsSpam(hams[i]);
if (result != CategorizationResult.First)
{
Interlocked.Increment(ref r1);
}
});
r = r1 / 72000d * 100;
Console.WriteLine($@"{r}%");
Assert.IsTrue(r > 0.97);
Assert.AreEqual(spamDetector.IsSpam(corpus), CategorizationResult.Second);
}
public void Given_text_that_has_known_ham_and_unknown_words()
{
var detector = new SpamDetector();
var ham = @"我在短暂的人生里发现...一个人越是玩弄阴谋，就越会感到人类的能力是有极限的....";
var mysteryMeat = @"欧拉欧拉欧拉欧拉欧拉欧拉欧拉";
detector.HamFound(ham);
var result = detector.IsSpam(mysteryMeat);
Assert.AreEqual(result, CategorizationResult.Undetermined);
}
public void Given_text_that_has_an_equivalent_number_of_ham_and_spam()
{
var detector = new SpamDetector();
var ham = @"不愧是DIO！";
var spam = @"不过，我拒绝";
detector.HamFound(ham);
detector.SpamFound(spam);
var result = detector.IsSpam(@"不愧是DIO！不过，我拒绝");
Assert.AreEqual(result, CategorizationResult.Undetermined);
}
public void Given_text_previously_marked_as_spam()
{
var detector = new SpamDetector();
var corpus = @"无駄";
var ham = @"我在短暂的人生里发现...一个人越是玩弄阴谋，就越会感到人类的能力是有极限的....";
detector.SpamFound(corpus);
detector.HamFound(ham);
var isSpam = detector.IsSpam(corpus);
Assert.AreEqual(isSpam, CategorizationResult.First);
}
public void Test()
{
var corpus = @"不愧是DIO！我们不敢做的事，他毫不在乎地做了！真是佩服，真是我们的偶像！";
var currentPath = @".\";
var spamPath = Path.Combine(currentPath, @"Resources", @"spam.txt");
var hamPath = Path.Combine(currentPath, @"Resources", @"ham.txt");
var spamDetector = new SpamDetector();
spamDetector.Train(spamPath, hamPath);
Assert.AreEqual(spamDetector.IsSpam(corpus), CategorizationResult.Second);
}
public void Given_a_corpus_that_is_heavily_weighted_towards_being_spam()
{
var ham = @"做人了";
var spam = @"我不做人了！";
var detector = new SpamDetector();
detector.HamFound(ham);
detector.SpamFound(spam);
var result = detector.IsSpam(@"我不做人了！");
Assert.AreEqual(result, CategorizationResult.First);
}
private static void Main(string[] args)
{
var detector = new SpamDetector();
var tests = new Dictionary <string, bool>
{
{ "Sdf sdfsdfds fsff Dds fsf", true },
{ "wererwwrwerweerer rw rewrwerwererwr", true },
{ "hwajt the fdu eh ui futugt tge ufgb ets hewat ti sdo it", true },
{ "yrt asre the bertpera sojifsi hwrifle seen fie sue usertaje rwewererwer", true },
{ "EiybvisgfdsffbiFALFIFOBIOFUIBFRBsjdkdkdsjf", true },
{ "igish ssddf sdf", true },
{ "ioacdu uqeidiqdudaisdu a s diudsdi dusu doidu aso diuodsi uoiduaudsiaudiudasd aud", true },
{ "zvuoiyvci zoioizuoicu i udaou oiuoiuiweuiou we uoqu urqu ieuiuioqouwu we uuiowu", true },
{ "wert yuos", true },
{ "uadfa ffo sod fhfofdos fa jisjd efdqw wqddwdww", true },
{ "qrqrqr eri erer yqe ryry dfqeroworyqrqer rrqrere wrqo", true },
{
".snoitcurtsni eht yb deificeps snoitarepo )O/I( tuptuo/tupni dna lortnoc ,lacigol ,citemhtira cisab eht gnimrofrep yb margorp retupmoc a fo snoitcurtsni eht tuo seirrac taht retupmoc a nihtiw yrtiucric cinortcele eht si )UPC( tinu gnissecorp lartnec A",
true
},
{ "easd ojd eqorj daiweo jowfie qjo qwej jo w weji", true },
{ "nzcna di sidadoitna ffeha ef eq h hqq fqh erhe rr qrhre afe oqiwhi qweihuho f", true },
{ "rtesting sdf what sdf re foosdf", true },
{ "This is an example", false },
{ "Yes", false },
{ "for loop", false },
{ "Awesome", false },
{ "dictionary", false },
{ "apple bannana orange", false },
{ "Ok thanks", false },
{
"A central processing unit (CPU) is the electronic circuitry within a computer that carries out the instructions of a computer program by performing the basic arithmetic, logical, control and input/output (I/O) operations specified by the instructions.",
false
},
{ "This library detects if a phrase is spam or a legitimate sentence", false },
{ "Computer programming, often shortened to programming, sometimes called coding", false },
{
"In cryptography, encryption is the process of encoding messages or information in such a way that only authorized parties can read it.",
false
},
{ "keyboard spam detector", false },
};
foreach (var test in tests)
{
var result = detector.IsSpam(test.Key);
Console.ForegroundColor = ConsoleColor.Green;
if (result != test.Value)
{
Console.ForegroundColor = ConsoleColor.Red;
}
Console.Write(result.ToString().ToUpper());
Console.ForegroundColor = ConsoleColor.Gray;
Console.WriteLine(" " + test.Key);
}
Console.WriteLine();
while (true)
{
var phrase = Console.ReadLine();
var result = detector.IsSpam(phrase);
Console.ForegroundColor = ConsoleColor.Green;
if (result)
{
Console.ForegroundColor = ConsoleColor.Red;
}
Console.WriteLine(result ? "Spam" : "Not Spam");
Console.ForegroundColor = ConsoleColor.Gray;
}
}
