Class SparseReader

Reader for data files containing samples in libsvm's sparse format.
Inheritance
System.Object
SparseReader
Implements
System.IDisposable
Inherited Members
System.Object.Equals(System.Object)
System.Object.Equals(System.Object, System.Object)
System.Object.GetHashCode()
System.Object.GetType()
System.Object.MemberwiseClone()
System.Object.ReferenceEquals(System.Object, System.Object)
System.Object.ToString()
Namespace: Mars.Common.IO
Assembly: Mars.IO.dll
Syntax
public sealed class SparseReader : IDisposable
Examples

The following example shows how to read all sparse samples from a file and retrieve them as a dense multidimensional vector. 

// Suppose we are going to read a sparse sample file containing
//  samples which have an actual dimension of 4. Since the samples
//  are in a sparse format, each entry in the file will probably
//  have a much lesser number of elements.
//
int sampleSize = 4;

// Create a new Sparse Sample Reader to read any given file,
//  passing the correct dense sample size in the constructor
//
SparseReader reader = new SparseReader(file, Encoding.Default, sampleSize);

// Declare a vector to obtain the label
//  of each of the samples in the file
//
int[] labels = null;

// Declare a vector to obtain the description (or comments)
//  about each of the samples in the file, if present.
//
string[] descriptions = null;

// Read the sparse samples and store them in a dense vector array
double[][] samples = reader.ReadToEnd(out labels, out descriptions);

Additionally, it is also possible to read each sample individually and sequentially. For this, we can use a while loop until we reach the end of the stream. 

// Suppose we are going to read a sparse sample file containing
//  samples which have an actual dimension of 4. Since the samples
//  are in a sparse format, each entry in the file will probably
//  have a much lesser number of elements.
//
int sampleSize = 4;

// Create a new Sparse Sample Reader to read any given file,
//  passing the correct dense sample size in the constructor
//
SparseReader reader = new SparseReader(file, Encoding.Default, sampleSize);

// Declare some variables to receive each sample
//
int label = 0;
string description;
double[] sample;

// Read a single sample from the file
sample = reader.ReadDense(out label, out description);

// Read all other samples from the file
while (!reader.EndOfStream)
{
  sample = reader.ReadDense(out label, out description);
}

Constructors

| Improve this Doc View Source

SparseReader(Stream, Encoding, Int32, CultureInfo)

Initializes a new instance of the SparseReader class.
Declaration
public SparseReader(Stream stream, Encoding encoding = null, int sampleSize = -1, CultureInfo culture = null)
Parameters
Type Name Description
System.IO.Stream stream The file stream to be read.
System.Text.Encoding encoding The character encoding to use.
System.Int32 sampleSize The size of the feature vectors stored in the file.
System.Globalization.CultureInfo culture The culture specification used to read files (e.g., de-DE for german comma decimal and time formats)
| Improve this Doc View Source

SparseReader(String, Encoding, Int32, CultureInfo)

Initializes a new instance of the SparseReader class.
Declaration
public SparseReader(string path, Encoding encoding = null, int sampleSize = -1, CultureInfo culture = null)
Parameters
Type Name Description
System.String path The complete file path to be read.
System.Text.Encoding encoding The character encoding to use.
System.Int32 sampleSize The size of the feature vectors stored in the file.
System.Globalization.CultureInfo culture Culture specific serialization options to read from e.g., number format point or comma.

Properties

| Improve this Doc View Source

BaseStream

Returns the underlying stream.
Declaration
public Stream BaseStream { get; }
Property Value
Type Description
System.IO.Stream
| Improve this Doc View Source

EndOfStream

Gets a value that indicates whether the current stream position is at the end of the stream.
Declaration
public bool EndOfStream { get; }
Property Value
Type Description
System.Boolean
| Improve this Doc View Source

Intercept

Gets or sets whether to include an intercept term (bias) value at the beginning of each new sample. Default is null (don't include anything).
Declaration
public double? Intercept { get; set; }
Property Value
Type Description
System.Nullable<System.Double>
| Improve this Doc View Source

NumberOfInputs

Gets the number of features present in this dataset. Please note that, when using the sparse representation, it is not strictly necessary to know this value.
Declaration
public int NumberOfInputs { get; }
Property Value
Type Description
System.Int32
| Improve this Doc View Source

SampleDescriptions

Gets the description associated with the last read values.
Declaration
public List<string> SampleDescriptions { get; }
Property Value
Type Description
System.Collections.Generic.List<System.String>

Methods

| Improve this Doc View Source

Dispose()

Performs application-defined tasks associated with freeing, releasing, or resetting unmanaged resources.
Declaration
public void Dispose()
| Improve this Doc View Source

Finalize()

Releases unmanaged resources and performs other cleanup operations before the SparseReader is reclaimed by garbage collection.
Declaration
protected void Finalize()
| Improve this Doc View Source

Read(Int32, out Sparse<Double>[], out Boolean[])

Reads count samples from the file and returns them as a Sparse<T> sparse vector, together with their associated output values.
Declaration
public void Read(int count, out Sparse<double>[] samples, out bool[] outputs)
Parameters
Type Name Description
System.Int32 count The number of samples to read.
Sparse<System.Double>[] samples The samples that have been read from the file.
System.Boolean[] outputs The output labels associated with each sample in samples.
| Improve this Doc View Source

ReadDense()

Reads a sample from the file and returns it as a dense vector, together with its associated output value.
Declaration
public Tuple<double[], double> ReadDense()
Returns
Type Description
System.Tuple<System.Double[], System.Double> A tuple containing the dense vector as the first item and its associated output value as the second item.
| Improve this Doc View Source

ReadDense(Int32)

Reads count samples from the file and returns them as a Sparse<T> sparse vector, together with their associated output values.
Declaration
public Tuple<double[][], double[]> ReadDense(int count)
Parameters
Type Name Description
System.Int32 count The number of samples to read.
Returns
Type Description
System.Tuple<System.Double[][], System.Double[]> A tuple containing the sparse vectors as the first item and their associated output values as the second item.
| Improve this Doc View Source

ReadDenseToEnd()

Reads all samples from the file and returns them as a dense vector, together with their associated output values.
Declaration
public Tuple<double[][], double[]> ReadDenseToEnd()
Returns
Type Description
System.Tuple<System.Double[][], System.Double[]> A tuple containing the dense vectors as the first item and their associated output values as the second item.
| Improve this Doc View Source

ReadSparse()

Reads a sample from the file and returns it as a Sparse<T> sparse vector, together with its associated output value.
Declaration
public Tuple<Sparse<double>, double> ReadSparse()
Returns
Type Description
System.Tuple<Sparse<System.Double>, System.Double> A tuple containing the sparse vector as the first item and its associated output value as the second item.
| Improve this Doc View Source

ReadToEnd(out Sparse<Double>[], out Boolean[])

Reads all samples from the file and returns them as a Sparse<T> sparse vector, together with their associated output values.
Declaration
public void ReadToEnd(out Sparse<double>[] samples, out bool[] outputs)
Parameters
Type Name Description
Sparse<System.Double>[] samples The samples that have been read from the file.
System.Boolean[] outputs The output labels associated with each sample in samples.

Implements

System.IDisposable

Extension Methods