public void InitMetaColumn(IEnumerable <string> mData)
{
try
{
var mcl = new List <MetaColumn>();
//parse meta data
int counter = 0;
foreach (var c in mData)
{
MetaColumn col = new MetaColumn();
//check if double point appear more than one time. In that case raise exception
if (c.Count(x => x == ':') > 1)
{
throw new Exception("Column data contains double point ':' which is reserved char. PLease remove double point from metadata.");
}
var strData = c.Substring(c.IndexOf(":") + 1);
var colValues = strData.Split(new char[] { ';' }, StringSplitOptions.RemoveEmptyEntries);
//
col.Name = colValues[0];
col.Id = counter;
col.Index = counter;
col.MissingValue = colValues[3];
col.Param = colValues[2];
//col.Scale = colValues[4];
col.Type = colValues[1];
col.Encoding = col.Type.Equals("category", StringComparison.OrdinalIgnoreCase) ? CategoryEncoding.OneHot.ToString() : CategoryEncoding.None.ToString();
counter++;
mcl.Add(col);
}
//
MetaData = mcl.ToArray();
}
catch (Exception)
{
throw;
}
}
private MetaColumn[] GetMetadata()
{
var cols = new List <MetaColumn>();
foreach (var c in m_trainData)
{
var m = new MetaColumn();
m.Encoding = c.Encoding.Description();
m.Id = c.Id;
m.Index = c.Index;
// m.IsIngored =c.is
m.MissingValue = c.MissingValue.ToString();
m.Name = c.Name;
m.Param = c.GetVariableType();
m.Scale = c.Scaling.ToString();
m.Type = c.ColumnDataType.ToString();
cols.Add(m);
}
return(cols.ToArray());
}
/// <summary>
/// Creates columns based of the columns properties argument
/// </summary>
/// <param name="colProp"></param>
/// <returns></returns>
private ColumnData CreateColumn(MetaColumn colProp)
{
//determine column type
ColumnType colType;
if (colProp.Type.Equals(ColumnType.Numeric.Description(), StringComparison.OrdinalIgnoreCase))
{
colType = ColumnType.Numeric;
}
else if (colProp.Type.Equals(ColumnType.Binary.Description(), StringComparison.OrdinalIgnoreCase))
{
colType = ColumnType.Binary;
}
else if (colProp.Type.Equals(ColumnType.Category.Description(), StringComparison.OrdinalIgnoreCase))
{
colType = ColumnType.Category;
}
else
{
colType = ColumnType.Unknown;
}
//determine encoding for category column type
CategoryEncoding colEncoding;
if (colProp.Encoding == null)
{
colEncoding = CategoryEncoding.None;
}
else if (colProp.Encoding.Equals(CategoryEncoding.Level.Description(), StringComparison.OrdinalIgnoreCase) ||
colProp.Encoding.Equals(CategoryEncoding.Level.ToString(), StringComparison.OrdinalIgnoreCase))
{
colEncoding = CategoryEncoding.Level;
}
else if (colProp.Encoding.Equals(CategoryEncoding.OneHot.Description(), StringComparison.OrdinalIgnoreCase) ||
colProp.Encoding.Equals(CategoryEncoding.OneHot.ToString(), StringComparison.OrdinalIgnoreCase))
{
colEncoding = CategoryEncoding.OneHot;
}
else if (colProp.Encoding.Equals(CategoryEncoding.Dummy1.Description(), StringComparison.OrdinalIgnoreCase) ||
colProp.Encoding.Equals(CategoryEncoding.Dummy1.Description(), StringComparison.OrdinalIgnoreCase))
{
colEncoding = CategoryEncoding.Dummy1;
}
else if (colProp.Encoding.Equals(CategoryEncoding.Dummy2.Description(), StringComparison.OrdinalIgnoreCase) ||
colProp.Encoding.Equals(CategoryEncoding.Dummy2.ToString(), StringComparison.OrdinalIgnoreCase))
{
colEncoding = CategoryEncoding.Dummy2;
}
else if (colProp.Encoding.Equals(CategoryEncoding.Binary1.Description(), StringComparison.OrdinalIgnoreCase) ||
colProp.Encoding.Equals(CategoryEncoding.Binary1.ToString(), StringComparison.OrdinalIgnoreCase))
{
colEncoding = CategoryEncoding.Binary1;
}
else if (colProp.Encoding.Equals(CategoryEncoding.Binary2.Description(), StringComparison.OrdinalIgnoreCase) ||
colProp.Encoding.Equals(CategoryEncoding.Binary2.ToString(), StringComparison.OrdinalIgnoreCase))
{
colEncoding = CategoryEncoding.Binary2;
}
else if (colProp.Encoding.Equals(CategoryEncoding.None.Description(), StringComparison.OrdinalIgnoreCase) ||
colProp.Encoding.Equals(CategoryEncoding.None.ToString(), StringComparison.OrdinalIgnoreCase))
{
colEncoding = CategoryEncoding.None;
}
else
{
throw new Exception($"Unknown encoding !");
}
//create column data type
var isOutput = colProp.Param.Equals(ParameterType.Output.Description(), StringComparison.OrdinalIgnoreCase) ||
colProp.Param.Equals(VariableType.Label.Description(), StringComparison.OrdinalIgnoreCase);
ColumnData col = new ColumnData(isOutput, colEncoding);
if (colProp.Scale == null)
{
col.SetNormalization(Scaling.None);
}
else if (colProp.Scale.Equals(Scaling.MinMax.Description(), StringComparison.OrdinalIgnoreCase))
{
col.SetNormalization(Scaling.MinMax);
}
else if (colProp.Scale.Equals(Scaling.Gauss.Description(), StringComparison.OrdinalIgnoreCase))
{
col.SetNormalization(Scaling.Gauss);
}
else if (colProp.Scale.Equals(Scaling.None.Description(), StringComparison.OrdinalIgnoreCase))
{
col.SetNormalization(Scaling.None);
}
else
{
throw new Exception($"Unknown scaling for column '{col.Name}'");
}
//set missing value action
if (colProp.MissingValue.Equals(MissingValue.Ignore.Description(), StringComparison.OrdinalIgnoreCase))
{
col.MissingValue = MissingValue.Ignore;
}
else if (colProp.MissingValue.Equals(MissingValue.Average.Description(), StringComparison.OrdinalIgnoreCase))
{
col.MissingValue = MissingValue.Average;
}
else if (colProp.MissingValue.Equals(MissingValue.Random.Description(), StringComparison.OrdinalIgnoreCase))
{
col.MissingValue = MissingValue.Random;
}
else if (colProp.MissingValue.Equals(MissingValue.Mode.Description(), StringComparison.OrdinalIgnoreCase))
{
col.MissingValue = MissingValue.Mode;
}
else if (colProp.MissingValue.Equals(MissingValue.Max.Description(), StringComparison.OrdinalIgnoreCase))
{
col.MissingValue = MissingValue.Max;
}
else if (colProp.MissingValue.Equals(MissingValue.Min.Description(), StringComparison.OrdinalIgnoreCase))
{
col.MissingValue = MissingValue.Min;
}
else
{
throw new Exception($"Unknown missing value for column '{col.Name}'");
}
//set column name and type
col.Name = colProp.Name;
col.ColumnDataType = colType;
col.Encoding = colEncoding;
col.Id = colProp.Id;
col.Index = colProp.Index;
return(col);
}
