private void SetParameters(string inputPath, ScanningMethod scanningMethod, DescriptorMethod descriptorMethod, MachineLearningMethod machineLearningMethod)
{
_symbolParameters = new Utility.SymbolParameters();
_hessianThreshould = _symbolParameters.HessianThresh;
_uniquenessThresh = _symbolParameters.UniquenessThress;
_histogramMatchingScore = _symbolParameters.HistogramMatchingScore;
_tm = _symbolParameters.TM;
_log = File.AppendText(inputPath + "/log.txt");
_machineLearningMethod = machineLearningMethod;
_scanningMethod = scanningMethod;
_descriptorType = descriptorMethod;
//check the FilePath
_inputPath = inputPath;
_inPath = inputPath + "/in";
_outPath = inputPath + "/out";
_intermediatePath = inputPath + "/Intermediate";
ValidateFilePath(_inPath, "in");
ValidateFilePath(_outPath, "out");
ValidateFilePath(_intermediatePath, "Intermediate");
ValidateFilePath(_inputPath, "");
_descriptor = new Descriptor();
_machineLearning = new MachineLearning();
_scanImage = new ScanImage();
}
// Start is called before the first frame update
void Start()
{
rb = GetComponent <Rigidbody2D>();
anim = GetComponent <Animator>();
stateController = stateControllerObject.GetComponent <StateController>();
player = GameObject.Find("RL AI").GetComponent <MachineLearning>();
}
/// <summary>
/// Callback for when the Train Model button is clicked. Disables all button and launches a new thread which asks extracts features (if not present) and then trains the model using
/// <see cref="MachineLearning.TrainModel{TFaceData}"/> (if not present), prompting the user if an existing version is found. The face data type passed in is dependent on <see cref="selectedFaceDataType"/>.
/// </summary>
/// <param name="sender"></param>
/// <param name="e"></param>
private async void TrainModelButton_Click(object sender, RoutedEventArgs e)
{
TrainModelButton.IsEnabled = false;
PredictEmotionButton.IsEnabled = false;
LoadImageToPredict.IsEnabled = false;
EvaluateModelButton.IsEnabled = false;
RunBenchmarksMenuItem.IsEnabled = false;
var trainButtonOriginalLabel = (string)TrainModelButton.Content;
TrainModelButton.Content = $"{trainButtonOriginalLabel} (Training...)";
await Task.Run(() =>
{
var existingFile = MachineLearning.GetFile(MachineLearning.GetFacialFeaturesFileName(selectedFaceDataType), false);
bool recalculatedFacialFeatures = false;
if (existingFile != null && IsExistingFileValid(existingFile))
{
if (MessageBox.Show("Existing facial landmarks file exists, extract again (will take extra time)?",
"Extracted Facial Feature File Found", MessageBoxButton.YesNo, MessageBoxImage.Question) ==
MessageBoxResult.Yes)
{
MachineLearning.ExtractFeatures(selectedFaceDataType);
recalculatedFacialFeatures = true;
}
}
else
{
MachineLearning.ExtractFeatures(selectedFaceDataType);
recalculatedFacialFeatures = true;
}
if (recalculatedFacialFeatures)
{
TrainModelAccordingToCurrentFaceMode();
}
else if (MachineLearning.GetFile(MachineLearning.GetModelZipFileName(selectedFaceDataType), false) !=
null)
{
if (MessageBox.Show("Trained model found, use the one stored on disk?", "Trained Model Found",
MessageBoxButton.YesNo, MessageBoxImage.Question) == MessageBoxResult.No)
{
TrainModelAccordingToCurrentFaceMode();
}
}
else
{
TrainModelAccordingToCurrentFaceMode();
}
});
TrainModelButton.Content = $"{trainButtonOriginalLabel} (Trained)";
TrainModelButton.IsEnabled = true;
PredictEmotionButton.IsEnabled = true;
LoadImageToPredict.IsEnabled = true;
EvaluateModelButton.IsEnabled = true;
RunBenchmarksMenuItem.IsEnabled = true;
}
public override async Task HandleAsync(IUpdateContext context, UpdateDelegate next, string[] args,
CancellationToken cancellationToken)
{
_telegramService = new TelegramService(context);
var message = _telegramService.Message;
var msgText = message.Text.Split(' ');
var param1 = "\t";
// var param1 = msgText.ValueOfIndex(1).Trim() ?? ",";
if (!_telegramService.IsSudoer())
{
Log.Information("This user is not sudoer");
return;
}
if (message.ReplyToMessage != null)
{
await _telegramService.SendTextAsync("Sedang mengimport dataset")
.ConfigureAwait(false);
var repMessage = message.ReplyToMessage;
if (repMessage.Document != null)
{
var document = repMessage.Document;
var chatId = message.Chat.Id.ToString().TrimStart('-');
var msgId = repMessage.MessageId;
var fileName = document.FileName;
var filePath = $"learn-dataset-{chatId}-{msgId}-{fileName}";
var savedFile = await _telegramService.DownloadFileAsync(filePath)
.ConfigureAwait(false);
await _telegramService.ImportCsv(savedFile, param1).ConfigureAwait(false);
await _telegramService.EditAsync("Sedang mempelajari dataset")
.ConfigureAwait(false);
await MachineLearning.SetupEngineAsync()
.ConfigureAwait(false);
await _telegramService.EditAsync("Import selesai")
.ConfigureAwait(false);
}
else
{
var typeHint = "File yang mau di import harus berupa dokumen bertipe csv, tsv atau sejenis";
await _telegramService.SendTextAsync(typeHint)
.ConfigureAwait(false);
}
}
else
{
await _telegramService.SendTextAsync("Balas file yang mau di import")
.ConfigureAwait(false);
}
}
private static void runExample(string title, string mlConfigPath)
{
var mlConfigFile2 = mlConfigPath;
Console.WriteLine(Environment.NewLine);
Console.WriteLine($"****{title}****");
Console.WriteLine(Environment.NewLine);
var token2 = new CancellationToken();
MachineLearning.Run(mlConfigFile2, DeviceDescriptor.UseDefaultDevice(), token2, trainingProgress, null);
}
public EvaluationMetrics(IDataView testSetDataView, Type TFaceData)
{
InitializeComponent();
// Test model and display results.
MachineLearning.TestModel(testSetDataView, TFaceData, out var metrics);
ConfusionMatrixTextBlock.Text = "";
ConfusionMatrixTextBlock.Text = metrics.ConfusionMatrix.GetFormattedConfusionTable();
MicroAccuracyTextBlock.Text = metrics.MicroAccuracy.ToString("#.###");
MacroAccuracyTextBlock.Text = metrics.MacroAccuracy.ToString("#.###");
LogLossTextBlock.Text = metrics.LogLoss.ToString("#.###");
LogLossReductionTextBlock.Text = metrics.LogLossReduction.ToString("#.###");
}
/// <summary>
/// Extract face features by calling <see cref="MachineLearning.ExtractFeatures"/> for the <paramref name="faceDataType"/>, if the <paramref name="checkIfPresent"/> is true.
/// </summary>
/// <param name="faceDataType"></param>
/// <param name="checkIfPresent"></param>
/// <seealso cref="MachineLearning.ExtractFeatures"/>
public static void ExtractFacialFeatures(Type faceDataType, bool checkIfPresent = true)
{
if (checkIfPresent)
{
var foundFile = MachineLearning.GetFile(MachineLearning.GetFacialFeaturesFileName(faceDataType), false);
if (foundFile != null)
{
return;
}
}
MachineLearning.ExtractFeatures(faceDataType);
}
/// <summary>
/// Calls <see cref="MachineLearning.TrainModel{TFaceData}"/> with the TFaceData set according to <see cref="selectedFaceDataType"/> as a workaround
/// as type variables cannot be cast to generic types. Shows a message box informing the user once its done training.
/// </summary>
private void TrainModelAccordingToCurrentFaceMode()
{
// Because you can't pass FaceDataType into a generic without serious workarounds due to the use of an out paramater.
if (selectedFaceDataType == typeof(FaceData1))
{
MachineLearning.TrainModel <FaceData1>(out testSetDataView, 100000);
}
else if (selectedFaceDataType == typeof(FaceData2))
{
MachineLearning.TrainModel <FaceData2>(out testSetDataView, 1000000);
}
else if (selectedFaceDataType == typeof(FaceData3))
{
MachineLearning.TrainModel <FaceData3>(out testSetDataView, 100000);
}
MessageBox.Show("Training Complete", "Training is complete.", MessageBoxButton.OK,
MessageBoxImage.Information);
}
public async Task <IActionResult> Index(IFormFile file)
{
if (file != null)
{
string path = Path.Combine(Path.GetTempPath(), "test-files");
long size = file.Length;
if (!Directory.Exists(path))
{
Directory.CreateDirectory(path);
}
var filePath = Path.Combine(path, file.FileName);
if (file.Length > 0)
{
using (var stream = new FileStream(filePath, FileMode.Create))
{
await file.CopyToAsync(stream);
}
}
//string sha256 = new StaticAnalysisProject.Helpers.Hash.SHA256().GetHash(filePath);
IFileReport report = new FileReport(filePath);
var ml = new MachineLearning().Predict(report);
object[] json = new object[] {
report.ToJson(),
ml.ToJson()
};
if (System.IO.File.Exists(filePath))
{
System.IO.File.Delete(filePath);
}
return(Content(JsonConvert.SerializeObject(json)));
}
Response.StatusCode = (int)HttpStatusCode.BadRequest;
return(Ok("{\"error\"}"));
}
public override async Task HandleAsync(IUpdateContext context, UpdateDelegate next, string[] args,
CancellationToken cancellationToken)
{
_telegramService = new TelegramService(context);
var message = _telegramService.Message;
if (message.ReplyToMessage != null)
{
var repMsg = message.ReplyToMessage;
var repMsgText = repMsg.Text;
await _telegramService.SendTextAsync("Sedang memprediksi pesan")
.ConfigureAwait(false);
var isSpam = MachineLearning.PredictMessage(repMsgText);
await _telegramService.EditAsync($"IsSpam: {isSpam}").ConfigureAwait(false);
return;
}
else
{
await _telegramService.SendTextAsync("Predicting message").ConfigureAwait(false);
}
}
public override async Task HandleAsync(IUpdateContext context, UpdateDelegate next, string[] args)
{
await _telegramService.AddUpdateContext(context);
var message = _telegramService.Message;
if (message.ReplyToMessage != null)
{
var repMsg = message.ReplyToMessage;
var repMsgText = repMsg.Text;
await _telegramService.SendTextMessageAsync("Sedang memprediksi pesan")
;
var isSpam = MachineLearning.PredictMessage(repMsgText);
await _telegramService.EditMessageTextAsync($"IsSpam: {isSpam}");
return;
}
else
{
await _telegramService.SendTextMessageAsync("Predicting message");
}
}
void Start()
{
mL = GetComponent <MachineLearning>();
weights = GetComponent <Weights>();
fitness = GetComponent <FitnessFunction>();
onLevelCreated = gameObject.GetComponent <OnLevelCreated>();
noiseMap = GetComponent <NoiseMap>();
chiseling = GetComponent <Chiseling>();
bnm = GetComponent <BuildNavMesh>();
gg = GetComponent <GGManager>();
modules.AddRange(moduleSO.list);
weights.AssignWeights(modules);
//startBuilding();
for (int i = 0; i < loadScreen.list[0].transform.childCount; i++)
{
loadScreen.list[0].transform.GetChild(i).gameObject.SetActive(true);
}
//InitializeSlots();
}
/// <summary>
/// Callback for when the <see cref="PredictEmotionButton"/> is clicked. This launches a new task which calls <see cref="MachineLearning.PredictEmotion"/> with the
/// <see cref="currentImageFileInfo"/> and <see cref="selectedFaceDataType"/>, which results are then used to update <see cref="PredictedEmotionMainLabel"/> and
/// <see cref="PredictedEmotionsListBox"/>.
/// </summary>
/// <param name="sender"></param>
/// <param name="e"></param>
private async void PredictEmotionButton_Click(object sender, RoutedEventArgs e)
{
var originalLabel = PredictEmotionButton.Content;
PredictedEmotionMainLabel.Content = "Predicting...";
PredictedEmotionsListBox.Items.Clear();
PredictEmotionButton.Content = (string)PredictEmotionButton.Content + " (Predicting...)";
PredictEmotionButton.IsEnabled = false;
FaceOutput predictedEmotion = null;
Dictionary <string, float> predictedEmotionsWithAllLabels = null;
MachineLearning.DrawPointsOfLandmarks(currentImageFileInfo);
await Task.Run(() =>
{
MachineLearning.PredictEmotion(currentImageFileInfo, selectedFaceDataType, out var prediction, out var predictionWithLabels);
predictedEmotion = prediction;
predictedEmotionsWithAllLabels = predictionWithLabels;
});
PredictedEmotionMainLabel.Content = $"{FirstLetterToUpper(predictedEmotion.PredictedEmotion)}" +
$" ({FloatToPercent(predictedEmotionsWithAllLabels[predictedEmotion.PredictedEmotion])})";
PredictedEmotionsListBox.Items.Clear();
predictedEmotionsWithAllLabels = predictedEmotionsWithAllLabels.OrderByDescending(c => c.Value).ToDictionary(i => i.Key, i => i.Value);
foreach (var predictedEmotionLabelAndScore in predictedEmotionsWithAllLabels)
{
PredictedEmotionsListBox.Items.Add(
$"{FirstLetterToUpper(predictedEmotionLabelAndScore.Key)}: {FloatToPercent(predictedEmotionLabelAndScore.Value)}");
}
PredictedEmotionsListBox.Items.RemoveAt(0);
PredictEmotionButton.Content = originalLabel;
PredictEmotionButton.IsEnabled = true;
}
/// <summary>
/// Run the benchmark on the designated <paramref name="faceDataType"/> saving the results in a csv.
/// </summary>
/// <param name="faceDataType"></param>
public static void RunMaximumIterationBenchmarks(Type faceDataType)
{
ExtractFacialFeatures(faceDataType, true);
const int maxRepeatsExclusive = 5;
const int maxIterationMultipliersInclusive = 6;
const int factorMultiplier = 10; // The 10 part of 10^n
Dictionary <int, BenchmarkStatistics> bsIterations = new Dictionary <int, BenchmarkStatistics>(maxIterationMultipliersInclusive * maxRepeatsExclusive);
for (int repeats = 0; repeats < maxRepeatsExclusive; ++repeats)
{
for (int iterationMultiplier = 1; iterationMultiplier <= maxIterationMultipliersInclusive; ++iterationMultiplier)
{
IDataView testSet = null;
int iterations = (int)Math.Pow(factorMultiplier, iterationMultiplier);
Stopwatch sw = new Stopwatch();
sw.Start();
if (faceDataType == typeof(FaceData1))
{
MachineLearning.TrainModel <FaceData1>(out testSet, iterations, false);
}
else if (faceDataType == typeof(FaceData2))
{
MachineLearning.TrainModel <FaceData2>(out testSet, iterations, false);
}
else if (faceDataType == typeof(FaceData3))
{
MachineLearning.TrainModel <FaceData3>(out testSet, iterations, false);
}
else
{
throw new ArgumentException("No corresponding logic for the specified faceDataType",
nameof(faceDataType));
}
sw.Stop();
MachineLearning.TestModel(testSet, faceDataType, out var testMetrics);
if (!bsIterations.ContainsKey(iterationMultiplier))
{
bsIterations.Add(iterationMultiplier, new BenchmarkStatistics(testMetrics.ConfusionMatrix.NumberOfClasses));
}
bsIterations[iterationMultiplier].AddData(testMetrics, (double)sw.ElapsedMilliseconds);
}
}
using (var writer = new StreamWriter($"Iterations Benchmark {faceDataType.ToString()}.csv", false))
{
// GENERAL STATS
writer.Write("Log Loss\n");
writer.Write($"Iterations, Median, Max, Min\n");
foreach (var keyValuePair in bsIterations)
{
writer.Write($"{Math.Pow(factorMultiplier, keyValuePair.Key)}, {keyValuePair.Value.LogLoss.Median}, {keyValuePair.Value.LogLoss.Max}, {keyValuePair.Value.LogLoss.Min},\n");
}
writer.Write("\n");
writer.Write("Log Loss Reduction\n");
writer.Write($"Iterations, Median, Max, Min\n");
foreach (var keyValuePair in bsIterations)
{
writer.Write($"{Math.Pow(factorMultiplier, keyValuePair.Key)}, {keyValuePair.Value.LogLossReduction.Median}, {keyValuePair.Value.LogLossReduction.Max}, {keyValuePair.Value.LogLossReduction.Min},\n");
}
writer.Write("\n");
writer.Write("Micro Accuracy\n");
writer.Write($"Iterations, Median, Max, Min\n");
foreach (var keyValuePair in bsIterations)
{
writer.Write($"{Math.Pow(factorMultiplier, keyValuePair.Key)}, {keyValuePair.Value.MicroAccuracy.Median}, {keyValuePair.Value.MicroAccuracy.Max}, {keyValuePair.Value.MicroAccuracy.Min},\n");
}
writer.Write("\n");
writer.Write("Macro Accuracy\n");
writer.Write($"Iterations, Median, Max, Min\n");
foreach (var keyValuePair in bsIterations)
{
writer.Write($"{Math.Pow(factorMultiplier, keyValuePair.Key)}, {keyValuePair.Value.MacroAccuracy.Median}, {keyValuePair.Value.MacroAccuracy.Max}, {keyValuePair.Value.MacroAccuracy.Min},\n");
}
writer.Write("\n");
writer.Write("Time Taken\n");
writer.Write($"Iterations, Median, Max, Min\n");
foreach (var keyValuePair in bsIterations)
{
writer.Write($"{Math.Pow(factorMultiplier, keyValuePair.Key)}, {keyValuePair.Value.TimeTakenMiliSeconds.Median}, {keyValuePair.Value.TimeTakenMiliSeconds.Max}, {keyValuePair.Value.TimeTakenMiliSeconds.Min},\n");
}
writer.Write("\n");
// CONFUSION MATRICES
foreach (var keyValuePair in bsIterations)
{
writer.Write("\n");
writer.Write(keyValuePair.Value.StatConfusionMatrix.GetString($"Iterations {Math.Pow(factorMultiplier, keyValuePair.Key)}"));
writer.Write("\n");
}
writer.Flush();
}
}
public override async Task HandleAsync(IUpdateContext context, UpdateDelegate next, string[] args)
{
await _telegramService.AddUpdateContext(context);
var message = _telegramService.Message;
if (!_telegramService.IsFromSudo)
{
Log.Information("This user is not sudoer");
return;
}
if (message.ReplyToMessage != null)
{
var repMessage = message.ReplyToMessage;
var repText = repMessage.Text ?? repMessage.Caption;
var param = message.Text.SplitText(" ").ToArray();
var mark = param.ValueOfIndex(1);
var opts = new List <string>
{
"spam", "ham"
};
if (!opts.Contains(mark))
{
await _telegramService.SendTextMessageAsync("Spesifikasikan spam atau ham (bukan spam)");
return;
}
await _telegramService.SendTextMessageAsync("Sedang memperlajari pesan");
var learnData = new LearnData
{
Message = repText.Replace("\n", " "),
Label = mark,
ChatId = _telegramService.ChatId,
FromId = _telegramService.FromId
};
if (_learningService.IsExist(learnData))
{
Log.Information("This message has learned");
await _telegramService.EditMessageTextAsync("Pesan ini mungkin sudah di tambahkan.");
return;
}
await _learningService.Save(learnData);
await _telegramService.EditMessageTextAsync("Memperbarui local dataset");
await _telegramService.EditMessageTextAsync("Sedang mempelajari dataset");
await MachineLearning.SetupEngineAsync();
await _telegramService.EditMessageTextAsync("Pesan berhasil di tambahkan ke Dataset");
}
else
{
await _telegramService.SendTextMessageAsync("Sedang mempelajari dataset");
await MachineLearning.SetupEngineAsync();
await _telegramService.EditMessageTextAsync("Training selesai");
}
}
public override async Task HandleAsync(IUpdateContext context, UpdateDelegate next, string[] args,
CancellationToken cancellationToken)
{
_telegramService = new TelegramService(context);
_rssService = new RssService(context.Update.Message);
var chatId = _telegramService.Message.Chat.Id;
var fromId = _telegramService.Message.From.Id;
var msg = _telegramService.Message;
if (fromId.IsSudoer())
{
Log.Information("Test started..");
await _telegramService.SendTextAsync("Sedang mengetes sesuatu")
.ConfigureAwait(false);
// var data = await new Query("rss_history")
// .Where("chat_id", chatId)
// .ExecForMysql()
// .GetAsync();
//
// var rssHistories = data
// .ToJson()
// .MapObject<List<RssHistory>>();
//
// ConsoleHelper.WriteLine(data.GetType());
// // ConsoleHelper.WriteLine(data.ToJson(true));
//
// ConsoleHelper.WriteLine(rssHistories.GetType());
// // ConsoleHelper.WriteLine(rssHistories.ToJson(true));
//
// ConsoleHelper.WriteLine("Test completed..");
// await "This test".LogToChannel();
// await RssHelper.SyncRssHistoryToCloud();
// await BotHelper.ClearLog();
// await SyncHelper.SyncGBanToLocalAsync();
// var greet = TimeHelper.GetTimeGreet();
var inlineKeyboard = new InlineKeyboardMarkup(new[]
{
// new[]
// {
// InlineKeyboardButton.WithCallbackData("Warn Username Limit", "info warn-username-limit"),
// InlineKeyboardButton.WithCallbackData("-", "callback-set warn_username_limit 3"),
// InlineKeyboardButton.WithCallbackData("4", "info setelah"),
// InlineKeyboardButton.WithCallbackData("+", "callback-set warn_username_limit 5")
// },
new[]
{
// InlineKeyboardButton.WithCallbackData("Warn Username Limit", "info warn-username-limit"),
InlineKeyboardButton.WithCallbackData("-", "callback-set warn_username_limit 3"),
InlineKeyboardButton.WithCallbackData("4", "info setelah"),
InlineKeyboardButton.WithCallbackData("+", "callback-set warn_username_limit 5")
}
});
// await _telegramService.EditAsync("Warn Username Limit", inlineKeyboard);
// LearningHelper.Setup2();
// LearningHelper.Predict();
if (msg.ReplyToMessage != null)
{
var repMsg = msg.ReplyToMessage;
var repMsgText = repMsg.Text;
Log.Information("Predicting message");
var isSpam = MachineLearning.PredictMessage(repMsgText);
await _telegramService.EditAsync($"IsSpam: {isSpam}")
.ConfigureAwait(false);
return;
}
await _telegramService.EditAsync("Complete")
.ConfigureAwait(false);
}
// else
// {
// await _requestProvider.SendTextAsync("Unauthorized.");
// }
}
private void Awake()
{
machineLearning = machineLearningObject.GetComponent <MachineLearning>();
}
public Painter(MachineLearning machineLearningWindow)
{
// this.MlWindow = machineLearningWindow;
// this.Canvas = this.MlWindow.Canvas;
// this.ResetCanvas();
}
private void UpdateDataGridView(MachineLearning.Learning.Regression.LearningRound lastRound)
{
string[] row = new string[cmd.exp.mlSettings.numberOfRounds * 2];
row[0] = lastRound.round.ToString();
row[1] = lastRound.learningError.ToString();
double relativeError = 0.0;
cmd.exp.learning.computeError(lastRound.FeatureSet, GlobalState.allMeasurements.Configurations, out relativeError);
row[2] = relativeError.ToString();
lastRound.learningError.ToString();
foreach (Feature f in lastRound.FeatureSet)
{
string name = f.ToString();
if (!termToIndex.ContainsKey(name))
{
perfInfGridView.Invoke((MethodInvoker)(() => perfInfGridView.Columns[termToIndex.Count + perfInfGrid_definedColumns].Name = name));
termToIndex.Add(name, termToIndex.Count + perfInfGrid_definedColumns);
}
row[termToIndex[name]] = Math.Round(f.Constant,2).ToString();
}
perfInfGridView.Invoke((MethodInvoker)(() => this.perfInfGridView.Rows.Add(row)));
}
public void SymbolRecognitionBySVM(string positivePath, string negativePath, string testPath)
{
double[][] inputs, testInputs;
int[] outputs;
Image <Bgr, Byte>[] Positives, Negatives;
Image <Bgr, Byte> test;
int total = 0;
int counter = 0;
_scanImage = new ScanImage();
_descriptor = new Descriptor();
_scanImage.SetImages(positivePath, negativePath, testPath, out Positives, out Negatives, out test);
if (Negatives != null && Negatives.Length != 0)
{
total += Negatives.Length;
}
if (Positives != null && Positives.Length != 0)
{
total += Positives.Length;
}
inputs = new double[total][];
outputs = new Int32[total];
if (Negatives != null || Negatives.Length != 0)
{
for (int i = 0; i < Negatives.Length; i++)
{
inputs[counter] = _descriptor.GetImageVector(Negatives[i], 100);
outputs[counter++] = 0;
}
}
if (Positives != null || Positives.Length != 0)
{
for (int i = 0; i < Positives.Length; i++)
{
inputs[counter] = _descriptor.GetImageVector(Positives[i], 100);
outputs[counter++] = 1;
}
}
_scanImage.ScanByPixel(@"C:\Users\simakmo\Documents\Sima\Symbol Reconition\data\", test, Positives[0], 50, 50);
string[] filePaths = Directory.GetFiles(@"C:\Users\simakmo\Documents\Sima\Symbol Reconition\data\Intermediate", "*.jpg");
testInputs = new double[filePaths.Length][];
counter = 0;
foreach (string path in filePaths)
{
testInputs[counter++] = _descriptor.GetImageVector(new Image <Bgr, Byte>(path), 150);
}
_machineLearning = new MachineLearning();
int dimention = Positives[0].Height * Positives[0].Width;
int[] testOutputs = _machineLearning.ApplySVMByGussianKernel(2, 2.5, 0.001, 0.2, inputs, outputs, dimention, testInputs);
List <Point> results = new List <Point>();
for (int j = 0; j < testOutputs.Length; j++)
{
int x, y, num;
if (testOutputs[j] == 1)
{
GetImageParameters(out x, out y, out num, filePaths[j]);
results.Add(new Point(x, y));
}
}
Visualization.DrawResults(results, Positives[0].Size, test, testPath);
}
public async Task <IActionResult> GetRelatedArticles(int?id)
{
if (id == null)
{
return(NotFound());
}
const int NUM_OF_RELATED_ARTICLES = 3;
var relatedArticles = new List <Article>();
try
{
int mlRelatedArticle = MachineLearning.GetRelatedArticle(_articlesTransitionDataPath, id.Value);
if (mlRelatedArticle != id.Value)
{
var article = await _context.Articles.FirstAsync(c => c.ArticleId == mlRelatedArticle);
if (article != null)
{
relatedArticles.Add(article);
// Try to find a related article to the related article
int mlRelatedToRelatedArticle = MachineLearning.GetRelatedArticle(_articlesTransitionDataPath, mlRelatedArticle);
if (mlRelatedToRelatedArticle != mlRelatedArticle && mlRelatedToRelatedArticle != id.Value)
{
article = await _context.Articles.FirstAsync(c => c.ArticleId == mlRelatedToRelatedArticle);
if (article != null)
{
relatedArticles.Add(article);
}
}
}
}
}
catch (Exception)
{
// It's just mean that there is no data yet for this article
}
int articlesCount = _context.Articles.Count();
Random random = new Random();
// Fill the rest of the related articles with random articles, in order to
// have the option for diversity (and by that make better predictions in the future)
while (relatedArticles.Count() < NUM_OF_RELATED_ARTICLES)
{
int randomArticleId = random.Next(1, articlesCount);
if (randomArticleId != id.Value && !relatedArticles.Exists(c => c.ArticleId == randomArticleId))
{
var randomArticle = await _context.Articles.FirstAsync(c => c.ArticleId == randomArticleId);
if (randomArticle != null)
{
relatedArticles.Add(randomArticle);
}
}
}
return(Json(relatedArticles));
}
static void Main(string[] args)
{
#region TEST
//var a = new AFile(@"C:\Users\Zelvar\source\repos\KeyLoggerVSB\KeyLoggerVSB\bin\Release\App.exe");
//Console.Write(new Hashes(File.ReadAllBytes(@"C:\Users\Zelvar\source\repos\KeyLoggerVSB\KeyLoggerVSB\bin\Release\App.exe")).ToString());
//Console.WriteLine();
//Console.Write(new PE(@"D:\SW\Arduino\Arduino.exe").ToString());
//Console.WriteLine();
//Console.Write(new PE(@"E:\SW\RadioDJv2\bass.dll").ToString());
/*HashSet<A> abc = new HashSet<A>();
* var a = new A();
* var b = new A();
* a.Name = "123";
* b.Name = "123";
*
* abc.Add(a);
* abc.Add(b);
*
* Console.WriteLine(abc.Count);
*
* foreach (var obj in abc) {
*
* Console.WriteLine("{0} {1}", obj.GetHashCode(), obj.Name);
*
* }*/
//////////////////////////////////////////////
//new Strings(@"D:\SW\Arduino\Arduino.exe");
//var a = new Strings(@"C:\Users\Zelvar\source\repos\KeyLoggerVSB\KeyLoggerVSB\bin\Release\App.exe");
//var a = new Strings(@"D:\SW\Arduino\Arduino.exe");
//Console.WriteLine(a.ContainsIP("10.0.0.1"));
//Console.WriteLine(a.ContainsIP("FFFF::"));
//Console.WriteLine(a.ContainsIP("10.A.0.1"));
//Console.WriteLine(a.ContainsIP("10.0.0.1"));
//Console.WriteLine(a.ContainsIP("FFFF::"));
//Console.WriteLine(a.ContainsURL("http://10.A.0.1"));
//Console.WriteLine(a.ContainsURL("http://blabla.com"));
//Console.WriteLine(a.ContainsEmail("tralala [email protected]"));
//Console.WriteLine(a.ContainsFileName("soubor.dll"));
//Console.WriteLine(a.ContainsFileName("Tralala soubor.rar"));
//Console.WriteLine(a.ContainsFileName("Tralala"));
/*
* foreach (var ip in a.ExtractIP())
* {
* Console.WriteLine("IP: {0}", ip.ToString());
* }
*
* foreach (var mail in a.ExtractMail())
* {
* Console.WriteLine("Mail: {0}", mail.ToString());
* }
*
* foreach (var url in a.ExtractURL())
* {
* Console.WriteLine("Url: {0}", url.ToString());
* }
*
* foreach (var file in a.ExtractFiles())
* {
* Console.WriteLine("File: {0}", file.ToString());
* }
*
* foreach (var file in a.ExtractKnownMethods())
* {
* Console.WriteLine("Known methods: {0}", file.ToString());
* }*/
//Console.WriteLine(new VirusTotal(@"C:\Users\Zelvar\source\repos\KeyLoggerVSB\KeyLoggerVSB\bin\Release\App.exe").ToString());
//Console.WriteLine(new Strings(@"C:\Users\Zelvar\source\repos\KeyLoggerVSB\KeyLoggerVSB\bin\Release\App.exe"));
//new DetectWithYara(@"E:\Steam\steamapps\common\Grand Theft Auto V\GTA5.exe");
//IFileReport a = new FileReport(@"E:\Steam\steamapps\common\Grand Theft Auto V\GTA5.exe");
//IFileReport a = new FileReport(@"E:\Steam\steamapps\common\Grand Theft Auto V\GTA5.exe");
//IFileReport a = new FileReport(@"E:\Steam\steamapps\common\Grand Theft Auto V\GTA5.exe");
//IFileReport a = new FileReport(@"C:\Users\Zelvar\source\repos\KeyLoggerVSB\KeyLoggerVSB\bin\Release\HookLibrary.dll", "malware");
//string json = JsonSerializer.Serialize(a);
//Console.WriteLine(json);
//byte[] a = File.ReadAllBytes(@"E:\Steam\steamapps\common\Grand Theft Auto V\GTA5.exe");
//Console.WriteLine(a.Entropy());
//Console.WriteLine("Val: {0} Efc: {1}", ent.Value, ent.Efficiency);
//Console.WriteLine(ent.ToString());
//byte[] b = new byte[] { (byte)'A',(byte)'h',(byte)'o',(byte)'j',(byte)' ',(byte)'j',(byte)'a',(byte)'k',(byte)' ',(byte)'s',(byte)'e',(byte)' ',(byte)'m',(byte)'a',(byte)'s'};
//Console.WriteLine(b.Entropy());
/*var hs = new HashSet<string>();
* IList<IFileReport> _fileReports = new List<IFileReport>();
*
* IList<FileReportRecovered> list = ExtensionHelpers.ListFromJson(File.ReadAllText("malware.json"));
* //IList<FileReportRecovered> list = ExtensionHelpers.ListFromJson(File.ReadAllText("software.json"));
* _fileReports = _fileReports.Concat(list).ToList();
* //_fileReports = _fileReports.Concat(list2).ToList();
*
* foreach (var a in _fileReports) {
*
* if(a.Behavior.Contains(""))
*
* foreach (var b in a.Behavior)
* {
* hs.Add(b.ToString());
* }
* }
*
* foreach(var a in hs)
* {
* Console.WriteLine(a);
* }*/
//Build model
#endregion
var ent = new MachineLearning();
var data = Directory.GetFiles(@"W:\DP-Vystup\malware\malware", "*", SearchOption.AllDirectories).ToArray();
foreach (var file in data)
{
var predicted = ent.Predict(file);
Console.WriteLine("{1} is malware: {0} {2}", predicted.IsMalware, file, predicted.Probability);
}
//Console.WriteLine("Is malware: {0}", ent.Predict(@"W:\DP-Vystup\malware\Adobe, Acrobat PDF Writer 3._exe").IsMalware);
/*var abc = ent.Predict(@"C:\Users\Zelvar\source\repos\KeyLoggerVSB\KeyLoggerVSB\bin\Release\HookLibrary.dll");
* Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);
* abc = ent.Predict(@"C:\Users\Zelvar\source\repos\KeyLoggerVSB\KeyLoggerVSB\bin\Release\App.exe");
* Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);
* //abc = ent.Predict(@"C:\Users\Zelvar\source\repos\FontCreator\FontCreator\bin\Debug\FontCreator.exe");
* //Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);
*
* //Console.WriteLine("Is malware: {0}", ent.Predict(@"E:\Steam\steamapps\common\Grand Theft Auto V\GTA5.exe").IsMalware);
* abc = ent.Predict(@"C:\Users\Zelvar\source\repos\DLLInjection\x64\Release\Dll-compresed.dll");
* Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);
* abc = ent.Predict(@"C:\Users\Zelvar\source\repos\DLLInjection\x64\Release\Dll.dll");
* Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);
* abc = ent.Predict(@"C:\Users\Zelvar\source\repos\KeyLoggerVSB\KeyLoggerVSB\bin\Release\HookLibrary.dll");
* Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);
*
* abc = ent.Predict(@"W:\DP-Vystup\SW\DP\MalwareSample.BotnetListener\bin\Release\netcoreapp3.1\MalwareSample.BotnetListener.exe");
* Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);
* abc = ent.Predict(@"W:\DP-Vystup\SW\DP\MalwareSample.BotnetListener\bin\Release\netcoreapp3.1\MalwareSample.BotnetListener.dll");
* Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);
*
* abc = ent.Predict(@"W:\DP-Vystup\SW\DP\MalwareSample.Keylogger\bin\Release\netcoreapp3.1\MalwareSample.Keylogger.exe");
* Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);
* abc = ent.Predict(@"W:\DP-Vystup\SW\DP\MalwareSample.Keylogger\bin\Release\netcoreapp3.1\MalwareSample.Keylogger.dll");
* Console.WriteLine("Is malware: {0} {1}", abc.IsMalware, abc.Probability);*/
}
private void SetParameters(string inputPath, ScanningMethod scanningMethod, DescriptorMethod descriptorMethod, MachineLearningMethod machineLearningMethod)
{
_symbolParameters = new Utility.SymbolParameters();
_hessianThreshould = _symbolParameters.HessianThresh;
_uniquenessThresh = _symbolParameters.UniquenessThress;
_histogramMatchingScore = _symbolParameters.HistogramMatchingScore;
_tm = _symbolParameters.TM;
_log = File.AppendText(inputPath + "/log.txt");
_machineLearningMethod = machineLearningMethod;
_scanningMethod = scanningMethod;
_descriptorType = descriptorMethod;
//check the FilePath
_inputPath = inputPath;
_inPath = inputPath + "/in";
_outPath = inputPath + "/out";
_intermediatePath = inputPath + "/Intermediate";
ValidateFilePath(_inPath, "in");
ValidateFilePath(_outPath, "out");
ValidateFilePath(_intermediatePath, "Intermediate");
ValidateFilePath(_inputPath, "");
_descriptor = new Descriptor();
_machineLearning = new MachineLearning();
_scanImage = new ScanImage();
}
// This function is called via javascript in the "Details" view of an article
public async Task <IActionResult> GetRelatedArticles(int?id)
{
if (id == null)
{
return(RedirectToAction("ArticleError"));
}
const int NUM_OF_RELATED_ARTICLES = 3;
List <Article> relatedArticles = new List <Article>();
try
{
// Get the related article to the viewed article
int mlRelatedArticle = MachineLearning.GetRelatedArticle(_articlesTransitionDataPath, id.Value);
// Make sure we don't recommend the same article as the viewed article
if (mlRelatedArticle != id.Value)
{
var article = await _context.Articles.FirstAsync(c => c.ArticleId == mlRelatedArticle);
if (article != null)
{
relatedArticles.Add(article);
// Get the related article to the related article
int mlRelatedToRelatedArticle = MachineLearning.GetRelatedArticle(_articlesTransitionDataPath, mlRelatedArticle);
// Check that the recommended article isn't itself or the previous recommended article
if (mlRelatedToRelatedArticle != mlRelatedArticle && mlRelatedToRelatedArticle != id.Value)
{
article = await _context.Articles.FirstAsync(c => c.ArticleId == mlRelatedToRelatedArticle);
if (article != null)
{
relatedArticles.Add(article);
}
}
}
}
}
catch (Exception)
{
// We don't do anything since it is OK for the algorithm to fail is there isn't any data for the viewed article
}
int articlesCount = _context.Articles.Count();
Random random = new Random();
// Fill the rest of the articles with random articles (in case the ML returned the same article and for better training of the model)
while (relatedArticles.Count() < NUM_OF_RELATED_ARTICLES)
{
int randomArticleId = random.Next(1, articlesCount);
if (randomArticleId != id.Value && !relatedArticles.Exists(c => c.ArticleId == randomArticleId))
{
var randomArticle = await _context.Articles.FirstAsync(c => c.ArticleId == randomArticleId);
if (randomArticle != null)
{
relatedArticles.Add(randomArticle);
}
}
}
return(Json(relatedArticles));
}
public async Task <IActionResult> GetRelatedItems(int?id)
{
if (id == null)
{
return(NotFound());
}
const int NUM_OF_RELATED_ITEMS = 3;
var relatedItems = new List <Item>();
try
{
int mlRelatedItem = MachineLearning.GetRelatedItem(_itemsTransitionDataPath, id.Value);
if (mlRelatedItem != id.Value)
{
var item = await _context.Items.FirstAsync(c => c.ItemId == mlRelatedItem);
if (item != null)
{
relatedItems.Add(item);
// Try to find a related item to the related item
int mlRelatedToRelatedItem = MachineLearning.GetRelatedItem(_itemsTransitionDataPath, mlRelatedItem);
if (mlRelatedToRelatedItem != mlRelatedItem && mlRelatedToRelatedItem != id.Value)
{
item = await _context.Items.FirstAsync(c => c.ItemId == mlRelatedToRelatedItem);
if (item != null)
{
relatedItems.Add(item);
}
}
}
}
}
catch (Exception)
{
}
int itemsCount = _context.Items.Count();
Random random = new Random();
// Fill the rest of the related items with random items, in order to
// have the option for diversity (and by that make better predictions in the future)
while (relatedItems.Count() < NUM_OF_RELATED_ITEMS)
{
int randomItemId = random.Next(1, itemsCount);
if (randomItemId != id.Value && !relatedItems.Exists(c => c.ItemId == randomItemId))
{
var randomItem = await _context.Items.FirstAsync(c => c.ItemId == randomItemId);
if (randomItem != null)
{
relatedItems.Add(randomItem);
}
}
}
return(Json(relatedItems));
}
void Start()
{
behaviour = GameObject.Find("RL AI").GetComponent <MachineLearning>();
player = GameObject.Find("Player").GetComponent <PlayerBehaviour>();
}
public void SymbolRecognitionBySVM(string positivePath, string negativePath, string testPath)
{
double[][] inputs, testInputs;
int[] outputs;
Image<Bgr, Byte>[] Positives, Negatives;
Image<Bgr, Byte> test;
int total = 0;
int counter = 0;
_scanImage = new ScanImage();
_descriptor = new Descriptor();
_scanImage.SetImages(positivePath, negativePath, testPath, out Positives, out Negatives, out test);
if (Negatives != null && Negatives.Length != 0)
total += Negatives.Length;
if (Positives != null && Positives.Length != 0)
total += Positives.Length;
inputs = new double[total][];
outputs = new Int32[total];
if (Negatives != null || Negatives.Length != 0)
for (int i = 0; i < Negatives.Length; i++)
{
inputs[counter] = _descriptor.GetImageVector(Negatives[i], 100);
outputs[counter++] = 0;
}
if (Positives != null || Positives.Length != 0)
for (int i = 0; i < Positives.Length; i++)
{
inputs[counter] = _descriptor.GetImageVector(Positives[i], 100);
outputs[counter++] = 1;
}
_scanImage.ScanByPixel(@"C:\Users\simakmo\Documents\Sima\Symbol Reconition\data\", test, Positives[0], 50, 50);
string[] filePaths = Directory.GetFiles(@"C:\Users\simakmo\Documents\Sima\Symbol Reconition\data\Intermediate", "*.jpg");
testInputs = new double[filePaths.Length][];
counter = 0;
foreach (string path in filePaths)
testInputs[counter++] = _descriptor.GetImageVector(new Image<Bgr, Byte>(path), 150);
_machineLearning = new MachineLearning();
int dimention = Positives[0].Height * Positives[0].Width;
int[] testOutputs = _machineLearning.ApplySVMByGussianKernel(2, 2.5, 0.001, 0.2, inputs, outputs, dimention, testInputs);
List<Point> results = new List<Point>();
for (int j = 0; j < testOutputs.Length; j++)
{
int x, y, num;
if (testOutputs[j] == 1)
{
GetImageParameters(out x, out y, out num, filePaths[j]);
results.Add(new Point(x, y));
}
}
Visualization.DrawResults(results, Positives[0].Size, test, testPath);
}
public override async Task HandleAsync(IUpdateContext context, UpdateDelegate next, string[] args,
CancellationToken cancellationToken)
{
_telegramService = new TelegramService(context);
var message = _telegramService.Message;
_learningService = new LearningService(message);
if (!_telegramService.IsSudoer())
{
Log.Information("This user is not sudoer");
return;
}
if (message.ReplyToMessage != null)
{
var repMessage = message.ReplyToMessage;
var repText = repMessage.Text ?? repMessage.Caption;
var param = message.Text.SplitText(" ").ToArray();
var mark = param.ValueOfIndex(1);
var opts = new List <string>()
{
"spam", "ham"
};
if (!opts.Contains(mark))
{
await _telegramService.SendTextAsync("Spesifikasikan spam atau ham (bukan spam)")
.ConfigureAwait(false);
return;
}
await _telegramService.SendTextAsync("Sedang memperlajari pesan")
.ConfigureAwait(false);
var learnData = new LearnData()
{
Message = repText.Replace("\n", " "),
Label = mark
};
if (LearningService.IsExist(learnData))
{
Log.Information("This message has learned");
await _telegramService.EditAsync("Pesan ini mungkin sudah di tambahkan.")
.ConfigureAwait(false);
return;
}
await _learningService.Save(learnData).ConfigureAwait(false);
await _telegramService.EditAsync("Memperbarui local dataset")
.ConfigureAwait(false);
// MachineLearning.WriteToCsv();
await _telegramService.EditAsync("Sedang mempelajari dataset")
.ConfigureAwait(false);
await MachineLearning.SetupEngineAsync().ConfigureAwait(false);
// BackgroundJob.Enqueue(() => LearningHelper.SetupEngine());
await _telegramService.EditAsync("Pesan berhasil di tambahkan ke Dataset")
.ConfigureAwait(false);
return;
}
else
{
await _telegramService.SendTextAsync("Sedang mempelajari dataset")
.ConfigureAwait(false);
await MachineLearning.SetupEngineAsync()
.ConfigureAwait(false);
await _telegramService.EditAsync("Training selesai")
.ConfigureAwait(false);
return;
}
await _telegramService.SendTextAsync("Balas pesan yang ingin di pelajari")
.ConfigureAwait(false);
}
