public async Task CompareAccuracyTest()
{
var highAccuracyTlsh = new TlshBuilder(BucketSize.Buckets256, ChecksumSize.Checksum3Bytes);
var lowAccuracyTlsh = new TlshBuilder(BucketSize.Buckets128, ChecksumSize.Checksum1Byte);
await highAccuracyTlsh.LoadFromStringAsync(data);
var highAccuracyHash1 = highAccuracyTlsh.GetHash(false);
await lowAccuracyTlsh.LoadFromStringAsync(data);
var lowAccuracyHash1 = lowAccuracyTlsh.GetHash(false);
highAccuracyTlsh.Reset();
lowAccuracyTlsh.Reset();
await highAccuracyTlsh.LoadFromStringAsync(dataSlightlyChanged);
var highAccuracyHash2 = highAccuracyTlsh.GetHash(false);
await lowAccuracyTlsh.LoadFromStringAsync(dataSlightlyChanged);
var lowAccuracyHash2 = lowAccuracyTlsh.GetHash(false);
var highAccuracyMatch = highAccuracyHash1.TotalDiff(highAccuracyHash2, true);
var lowAccuracyMatch = lowAccuracyHash1.TotalDiff(lowAccuracyHash2, true);
Assert.IsTrue(highAccuracyMatch < 30);
Assert.IsTrue(lowAccuracyMatch < 30);
Assert.IsTrue(lowAccuracyMatch < highAccuracyMatch);
}
public async Task CompareNotSimilarTest()
{
var dataVeryChanged = "The worst docs are the Windows source. After all, this is not what the "
+ "system uses for docs when it decides what to do next! The "
+ "manuals paraphrase the source code, often having been written at "
+ "many times and by the same people than who wrote the code. "
+ "Think of them as nothing like guidelines. Always are more like wishes... "
+ "Therefore, it is all too common to ignore sources and find "
+ "options and weird behaviors that are not put in the docs. Sometimes "
+ "you find options not described in the docs that are incomplete "
+ "and weird by the source.";
var tlsh = new TlshBuilder(BucketSize.Buckets256, ChecksumSize.Checksum3Bytes);
await tlsh.LoadFromStringAsync(data);
var hash1 = tlsh.GetHash(false);
tlsh.Reset();
await tlsh.LoadFromStringAsync(dataVeryChanged);
var hash2 = tlsh.GetHash(false);
var match = hash1.TotalDiff(hash2, true);
Assert.IsTrue(match > 200);
}
public async Task CompareHashTest()
{
var tlsh = new TlshBuilder(BucketSize.Buckets256, ChecksumSize.Checksum3Bytes);
await tlsh.LoadFromStringAsync(data);
var hash1 = tlsh.GetHash(false);
tlsh.Reset();
await tlsh.LoadFromStringAsync(dataSlightlyChanged);
var hash2 = tlsh.GetHash(false);
var match = hash1.TotalDiff(hash2, true);
Assert.IsTrue(match < 30);
}