private static int ColumnCount(ref CSVLoaderSettings settings, ref string[] _row)
{
var row = _row;
int count = row.Length;
for (uint i = 0; i < row.Length; i++)
{
if (settings.filters.Exists(o => !o.Use(ref row, i)))
{
count--;
}
}
return(count);
}
// Prepare input data for prediction.
public static void PreparePredictor(ref NetworkContainer container, ref ImageNetworkPredictSettings settings)
{
TestForErrors(ref settings);
EncogWrapper.LoadNetworkFromFS(ref container, settings.trainedNetwork);
List <ICSVFilter> baseFilters = new List <ICSVFilter>(1);
ICSVFilter quaternions = new CSVEvenColumnFilter();
baseFilters.Add(quaternions);
// Setup loader.
CSVLoaderSettings CSVSettings = new CSVLoaderSettings
{
filePath = settings.predictData,
trimUp = 1,
trimDown = 0,
filters = baseFilters
};
var data = CSVLoader <Vector3> .LoadData(ref CSVSettings);
// Initialize image Transformer.
ImageTransformerSettings imageSettings = new ImageTransformerSettings
{
focusJoints = (LeapMotionJoint[])Enum.GetValues(typeof(LeapMotionJoint)),
samples = data,
size = settings.imgSize
};
ImageTransformer imageTransformer = new ImageTransformer();
if (settings.predictSettings.threshold.Equals(null))
{
settings.predictSettings = new EncogPredictSettings
{
threshold = 0.9
};
}
settings.predictSettings.data = imageTransformer.GetNeuralInput(imageSettings);
if (settings.predictSettings.data.Length != container.network.InputCount)
{
throw new NoNetworkMatchException("Sample count doesn't match network input count.");
}
}
// Convert data from a CSV file to the actual input array for the network.
private static void ComputeData(
uint sampleCount,
string inputDataDirectory,
ref double[][] outputData,
ref double[][] outputAnswers,
double outputValue,
int index)
{
DirectoryInfo inputDirectory = new DirectoryInfo(inputDataDirectory);
foreach (var file in inputDirectory.GetFiles("*.csv"))
{
List <ICSVFilter> baseFilters = new List <ICSVFilter>(1);
ICSVFilter quaternions = new CSVEvenColumnFilter();
baseFilters.Add(quaternions);
// Setup loader.
CSVLoaderSettings settings = new CSVLoaderSettings
{
filePath = file.FullName,
trimUp = 1,
trimDown = 0,
filters = baseFilters
};
var data = CSVLoader <Vector3> .LoadData(ref settings);
// Initialize image Transformer.
ImageTransformerSettings imageSettings = new ImageTransformerSettings
{
focusJoints = (LeapMotionJoint[])Enum.GetValues(typeof(LeapMotionJoint)),
samples = data,
size = sampleCount
};
ImageTransformer imageTransformer = new ImageTransformer();
BaseTrainHelper.Project1DInto2D(
imageTransformer.GetNeuralInput(imageSettings),
ref outputData,
index);
// Set answer to given value.
outputAnswers[index] = new double[] { outputValue };
index++;
}
}
public static void PreparePredictor(ref NetworkContainer container, ref CountNetworkPredictSettings settings)
{
TestForErrors(ref settings);
EncogWrapper.LoadNetworkFromFS(ref container, settings.trainedNetwork);
List <ICSVFilter> baseFilters = new List <ICSVFilter>(1);
ICSVFilter quaternions = new CSVEvenColumnFilter();
baseFilters.Add(quaternions);
// Setup loader.
CSVLoaderSettings CSVSettings = new CSVLoaderSettings
{
filePath = settings.predictData,
trimUp = 1,
trimDown = 0,
filters = baseFilters
};
var data = CSVLoader <Vector3> .LoadData(ref CSVSettings);
// Initialize CountBased Transformer settings.
IntervalBasedTransformerSettings countSettings = new IntervalBasedTransformerSettings
{
sampleList = data,
count = settings.sampleCount
};
CountBasedTransformer countTransformer = new CountBasedTransformer();
if (settings.predictSettings.threshold.Equals(null))
{
settings.predictSettings = new EncogPredictSettings
{
threshold = 0.9
};
}
settings.predictSettings.data = countTransformer.GetNeuralInput(countSettings);
if (settings.predictSettings.data.Length != container.network.InputCount)
{
throw new NoNetworkMatchException("Sample count doesn't match network input count.");
}
}
private static void ComputeData(
uint networkInputSize,
string inputDataDirectory,
ref double[][] outputData,
ref double[][] outputAnswers,
double outputValue,
int index)
{
DirectoryInfo inputDirectory = new DirectoryInfo(inputDataDirectory);
foreach (var file in inputDirectory.GetFiles("*.csv"))
{
List <ICSVFilter> baseFilters = new List <ICSVFilter>(1);
ICSVFilter quaternions = new CSVEvenColumnFilter();
baseFilters.Add(quaternions);
// Setup loader.
CSVLoaderSettings settings = new CSVLoaderSettings
{
filePath = file.FullName,
trimUp = 1,
trimDown = 0,
filters = baseFilters
};
var data = CSVLoader <Vector3> .LoadData(ref settings);
// Initialize CountBased Transformer settings.
IntervalBasedTransformerSettings countSettings = new IntervalBasedTransformerSettings
{
sampleList = data,
count = networkInputSize
};
CountBasedTransformer countTransformer = new CountBasedTransformer();
BaseTrainHelper.Project1DInto2D(
countTransformer.GetNeuralInput(countSettings),
ref outputData,
index);
// Set answer to given value.
outputAnswers[index] = new double[] { outputValue };
index++;
}
}
public static Sample <T>[] LoadData(ref CSVLoaderSettings settings)
{
if (!File.Exists(settings.filePath))
{
throw new FileNotFoundException(settings.filePath);
}
// Create a new Table.
var sampleList = new List <Sample <T> >();
// If the file has a header then skip it.
string[][] rows = File.ReadAllLines(settings.filePath)
.Select(line => line.Split(',')).ToArray();
if (settings.CSVHasHeader)
{
rows = rows.Skip(1).ToArray();
}
// Trim top or bottom rows if needed
if (settings.trimUp > 0 || settings.trimDown > 0)
{
rows = rows.Skip(settings.trimUp)
.Take(rows.Count() - settings.trimUp - settings.trimDown).ToArray();
}
if (rows.Count() == 0)
{
return(sampleList.ToArray());
}
int columnCount = ColumnCount(ref settings, ref rows[0]);
// For each row a sample is created.
for (uint rowIndex = 0; rowIndex < rows.Count(); rowIndex++)
{
// Check for unspecified end of file.
if (string.IsNullOrEmpty(rows[rowIndex][0]))
{
continue;
}
// Create new Sample.
Sample <T> sample = new Sample <T>();
// Parse required timestamp.
sample.timestamp = float.Parse(rows[rowIndex][0]);
sample.values = new T[columnCount];
uint valuesIndex = 0;
for (uint i = 0; i < rows[rowIndex].Count(); i++)
{
// Check if a filter is blocking the column.
if (settings.filters.Exists(o => !o.Use(ref rows[rowIndex], i)))
{
continue;
}
T vector = new T();
vector.parse(rows[rowIndex][i]);
sample.values[valuesIndex] = vector;
valuesIndex += 1;
}
sampleList.Add(sample);
}
return(sampleList.ToArray());
}
