/// <summary>
/// Computes the types (column and slotnames), the length reduction, categorical feature indices
/// and whether the column is suppressed.
/// The slotsMin and slotsMax arrays should be sorted and the intervals should not overlap.
/// </summary>
/// <param name="input">The input schema</param>
/// <param name="slotsMin">The beginning indices of the ranges of slots to be dropped</param>
/// <param name="slotsMax">The end indices of the ranges of slots to be dropped</param>
/// <param name="iinfo">The column index in Infos</param>
/// <param name="slotDropper">The slots to be dropped.</param>
/// <param name="suppressed">Whether the column is suppressed (all slots dropped)</param>
/// <param name="type">The column type</param>
/// <param name="categoricalRanges">Categorical feature indices.</param>
private void ComputeType(Schema input, int[] slotsMin, int[] slotsMax, int iinfo,
SlotDropper slotDropper, out bool suppressed, out ColumnType type, out int[] categoricalRanges)
{
Contracts.AssertValue(slotDropper);
Contracts.AssertValue(input);
Contracts.AssertNonEmpty(slotsMin);
Contracts.AssertNonEmpty(slotsMax);
Contracts.Assert(slotsMin.Length == slotsMax.Length);
Contracts.Assert(0 <= iinfo && iinfo < Infos.Length);
categoricalRanges = null;
// Register for metadata. Propagate the IsNormalized metadata.
using (var bldr = Metadata.BuildMetadata(iinfo, input, Infos[iinfo].Source,
MetadataUtils.Kinds.IsNormalized, MetadataUtils.Kinds.KeyValues))
{
var typeSrc = Infos[iinfo].TypeSrc;
if (!typeSrc.IsVector)
{
type = typeSrc;
suppressed = slotsMin.Length > 0 && slotsMin[0] == 0;
}
else if (!typeSrc.IsKnownSizeVector)
{
type = typeSrc;
suppressed = false;
}
else
{
Host.Assert(typeSrc.IsKnownSizeVector);
var dstLength = slotDropper.DstLength;
var hasSlotNames = input.HasSlotNames(Infos[iinfo].Source, Infos[iinfo].TypeSrc.VectorSize);
type = new VectorType(typeSrc.ItemType.AsPrimitive, Math.Max(dstLength, 1));
suppressed = dstLength == 0;
if (hasSlotNames && dstLength > 0)
{
// Add slot name metadata.
bldr.AddGetter <VBuffer <ReadOnlyMemory <char> > >(MetadataUtils.Kinds.SlotNames,
new VectorType(TextType.Instance, dstLength), GetSlotNames);
}
}
if (!suppressed)
{
if (MetadataUtils.TryGetCategoricalFeatureIndices(Source.Schema, Infos[iinfo].Source, out categoricalRanges))
{
VBuffer <int> dst = default(VBuffer <int>);
GetCategoricalSlotRangesCore(iinfo, slotDropper.SlotsMin, slotDropper.SlotsMax, categoricalRanges, ref dst);
// REVIEW: cache dst as opposed to caculating it again.
if (dst.Length > 0)
{
Contracts.Assert(dst.Length % 2 == 0);
bldr.AddGetter <VBuffer <int> >(MetadataUtils.Kinds.CategoricalSlotRanges,
MetadataUtils.GetCategoricalType(dst.Length / 2), GetCategoricalSlotRanges);
}
}
}
}
}
// Computes the column type and whether multiple indicator vectors need to be concatenated.
// Also populates the metadata.
private static void ComputeType(KeyToVectorTransform trans, ISchema input, int iinfo, ColInfo info, bool bag,
MetadataDispatcher md, out VectorType type, out bool concat)
{
Contracts.AssertValue(trans);
Contracts.AssertValue(input);
Contracts.AssertValue(info);
Contracts.Assert(info.TypeSrc.ItemType.IsKey);
Contracts.AssertValue(md);
int size = info.TypeSrc.ItemType.KeyCount;
Contracts.Assert(size > 0);
// See if the source has key names.
var typeNames = input.GetMetadataTypeOrNull(MetadataUtils.Kinds.KeyValues, info.Source);
if (typeNames == null || !typeNames.IsKnownSizeVector || !typeNames.ItemType.IsText ||
typeNames.VectorSize != size)
{
typeNames = null;
}
// Don't pass through any source column metadata.
using (var bldr = md.BuildMetadata(iinfo))
{
if (bag || info.TypeSrc.ValueCount == 1)
{
// Output is a single vector computed as the sum of the output indicator vectors.
concat = false;
type = new VectorType(NumberType.Float, size);
if (typeNames != null)
{
bldr.AddGetter <VBuffer <DvText> >(MetadataUtils.Kinds.SlotNames, typeNames, trans.GetKeyNames);
}
}
else
{
// Output is the concatenation of the multiple output indicator vectors.
concat = true;
type = new VectorType(NumberType.Float, info.TypeSrc.ValueCount, size);
if (typeNames != null && type.VectorSize > 0)
{
bldr.AddGetter <VBuffer <DvText> >(MetadataUtils.Kinds.SlotNames,
new VectorType(TextType.Instance, type), trans.GetSlotNames);
}
}
if (!bag && info.TypeSrc.ValueCount > 0)
{
bldr.AddGetter <VBuffer <DvInt4> >(MetadataUtils.Kinds.CategoricalSlotRanges,
MetadataUtils.GetCategoricalType(info.TypeSrc.ValueCount), trans.GetCategoricalSlotRanges);
}
if (!bag || info.TypeSrc.ValueCount == 1)
{
bldr.AddPrimitive(MetadataUtils.Kinds.IsNormalized, BoolType.Instance, DvBool.True);
}
}
}
