public void CreateDangerousEngine()
{
DangerousEngineCreated = true;
//var trainingFilePath = string.Format(@"Assets/StreamingAssets/{0}", Data.Instance.TraingDangerousDataPath);
//var testingFilePath = string.Format(@"Assets/StreamingAssets/{0}", Data.Instance.TestDangerousDataPath);
var trainingFilePath = UsefullUtils.GetPathToStreamingAssetsFile(Data.Instance.TraingDangerousDataPath);
var testingFilePath = UsefullUtils.GetPathToStreamingAssetsFile(Data.Instance.TestDangerousDataPath);
var trainingData = GetDataFromCsv(trainingFilePath);
var testData = GetDataFromCsv(testingFilePath);
IDataView dataView = mlContext.Data.LoadFromEnumerable(trainingData);
IDataView testDataView = mlContext.Data.LoadFromEnumerable(testData);
// Step 4 - we need to create the pipeline and define the workflows in it
// switch params: ("FeedBackText", "Features") to => ("Features", "FeedbackText")
var pipeline = mlContext.Transforms.Text.FeaturizeText("Features", "Features")
.Append(mlContext.BinaryClassification.Trainers
.FastTree(numberOfLeaves: 50, numberOfTrees: 50, minimumExampleCountPerLeaf: 1));
// Step 5 - traing the algorithm and get the model out
var model = pipeline.Fit(dataView);
var predictions = model.Transform(testDataView);
var metrics = mlContext.BinaryClassification.Evaluate(predictions, "Label");
// Step 7 - Use model
_dangerousPredictionsEngine = mlContext.Model.CreatePredictionEngine
<FeedBackTrainingData, FeedBackPrediction>
(model);
}
public void ReadString()
{
string path = UsefullUtils.GetPathToStreamingAssetsFile("Settings.txt");
//Read the text from directly from the test.txt file
StreamReader reader = new StreamReader(path);
IS_LOGGED_IN = reader.ReadToEnd() == "LOGGED_IN";
reader.Close();
}
public void UpdateCategoriesQuestions()
{
var savedCategories = DataService.GetAllCategories();
foreach (var category in savedCategories)
{
var filePath = UsefullUtils.GetPathToStreamingAssetsFile(category.File + ".html");
category.MaxLines = UsefullUtils.CountLinesInFile(filePath);
}
DataService.WriteCategoriesData(savedCategories, update: true);
}
public (IDataView training, IDataView test) LoadData(MLContext mlContext)
{
var trainingPath = Data.Instance.TraingDataPath;
var testingPath = Data.Instance.TraingDataPath;
//var trainingDataPath = string.Format(@"Assets/StreamingAssets/{0}", trainingPath);
//var testDataPath = string.Format(@"Assets/StreamingAssets/{0}", testingPath);
var trainingDataPath = UsefullUtils.GetPathToStreamingAssetsFile(trainingPath);
var testDataPath = UsefullUtils.GetPathToStreamingAssetsFile(testingPath);
IDataView trainingDataView = mlContext.Data.LoadFromTextFile <MedicRating>(trainingDataPath, hasHeader: true, separatorChar: ',');
IDataView testDataView = mlContext.Data.LoadFromTextFile <MedicRating>(testDataPath, hasHeader: true, separatorChar: ',');
return(trainingDataView, testDataView);
}
public void WriteString(string setting)
{
string path = UsefullUtils.GetPathToStreamingAssetsFile("Settings.txt");
File.WriteAllText(path, setting);
}