multicolumn mapping for some estimators

src/Microsoft.ML.Data/Transforms/ConversionsExtensionsCatalog.cs

@@ -65,6 +65,20 @@ public static TypeConvertingEstimator ConvertType(this TransformsCatalog.Convers
             DataKind outputKind = ConvertDefaults.DefaultOutputKind)
             => new TypeConvertingEstimator(CatalogUtils.GetEnvironment(catalog), new[] { new TypeConvertingEstimator.ColumnOptions(outputColumnName, outputKind, inputColumnName) });
 
+        /// <summary>
+        /// Changes column type of the input columns.
+        /// </summary>
+        /// <param name="catalog">The conversion transform's catalog.</param>
+        /// <param name="columns">Specifies the names of the columns on which to apply the transformation.</param>
+        /// <param name="outputKind">The expected kind of the output column.</param>
+        public static TypeConvertingEstimator ConvertType(this TransformsCatalog.ConversionTransforms catalog,
+            InputOutputColumnPair[] columns,
+            DataKind outputKind = ConvertDefaults.DefaultOutputKind)
+        {
+            var columnOptions = columns.Select(x => new TypeConvertingEstimator.ColumnOptions(x.OutputColumnName, outputKind, x.InputColumnName)).ToArray();
+            return new TypeConvertingEstimator(CatalogUtils.GetEnvironment(catalog), columnOptions);
+        }
+
         /// <summary>
         /// Changes column type of the input column.
         /// </summary>
@@ -89,6 +103,20 @@ internal static TypeConvertingEstimator ConvertType(this TransformsCatalog.Conve
         public static KeyToValueMappingEstimator MapKeyToValue(this TransformsCatalog.ConversionTransforms catalog, string outputColumnName, string inputColumnName = null)
             => new KeyToValueMappingEstimator(CatalogUtils.GetEnvironment(catalog), outputColumnName, inputColumnName);
 
+        /// <summary>
+        /// Convert the key types back to their original values.
+        /// </summary>
+        /// <param name="catalog">The conversion transform's catalog.</param>
+        /// <param name="columns">Specifies the names of the columns on which to apply the transformation.</param>
+        /// <example>
+        /// <format type="text/markdown">
+        /// <![CDATA[
+        ///  [!code-csharp[KeyToValueMappingEstimator](~/../docs/samples/docs/samples/Microsoft.ML.Samples/Dynamic/ValueMappingStringToKeyType.cs)]
+        /// ]]></format>
+        /// </example>
+        public static KeyToValueMappingEstimator MapKeyToValue(this TransformsCatalog.ConversionTransforms catalog, InputOutputColumnPair[] columns)
+            => new KeyToValueMappingEstimator(CatalogUtils.GetEnvironment(catalog), columns.Select(x => (x.OutputColumnName, x.InputColumnName)).ToArray());
+
         /// <summary>
         ///  Convert the key types (name of the column specified in the first item of the tuple) back to their original values
         ///  (named as specified in the second item of the tuple).
@@ -127,6 +155,21 @@ public static KeyToVectorMappingEstimator MapKeyToVector(this TransformsCatalog.
             string outputColumnName, string inputColumnName = null, bool outputCountVector = KeyToVectorMappingEstimator.Defaults.OutputCountVector)
             => new KeyToVectorMappingEstimator(CatalogUtils.GetEnvironment(catalog), outputColumnName, inputColumnName, outputCountVector);
 
+        /// <summary>
+        /// Maps columns of key types or key values into columns of floating point vectors.
+        /// </summary>
+        /// <param name="catalog">The conversion transform's catalog.</param>
+        /// <param name="columns">Specifies the names of the columns on which to apply the transformation.</param>
+        /// <param name="outputCountVector">Whether to combine multiple indicator vectors into a single vector of counts instead of concatenating them.
+        /// This is only relevant when the input column is a vector of keys.</param>
+        public static KeyToVectorMappingEstimator MapKeyToVector(this TransformsCatalog.ConversionTransforms catalog,
+            InputOutputColumnPair[] columns, bool outputCountVector = KeyToVectorMappingEstimator.Defaults.OutputCountVector)
+        {
+            var columnOptions = columns.Select(x => new KeyToVectorMappingEstimator.ColumnOptions(x.OutputColumnName, x.InputColumnName, outputCountVector)).ToArray();
+            return new KeyToVectorMappingEstimator(CatalogUtils.GetEnvironment(catalog), columnOptions);
+
+        }
+
         /// <summary>
         /// Converts value types into <see cref="KeyType"/>.
         /// </summary>
@@ -157,6 +200,36 @@ public static ValueToKeyMappingEstimator MapValueToKey(this TransformsCatalog.Co
            => new ValueToKeyMappingEstimator(CatalogUtils.GetEnvironment(catalog),
                new[] { new ValueToKeyMappingEstimator.ColumnOptions(outputColumnName, inputColumnName, maximumNumberOfKeys, keyOrdinality, addKeyValueAnnotationsAsText) }, keyData);
 
+        /// <summary>
+        /// Converts value types into <see cref="KeyType"/>.
+        /// </summary>
+        /// <param name="catalog">The conversion transform's catalog.</param>
+        /// <param name="columns">Specifies the names of the columns on which to apply the transformation.</param>
+        /// <param name="maximumNumberOfKeys">Maximum number of keys to keep per column when auto-training.</param>
+        /// <param name="keyOrdinality">How items should be ordered when vectorized. If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByOccurrence"/> choosen they will be in the order encountered.
+        /// If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByValue"/>, items are sorted according to their default comparison, for example, text sorting will be case sensitive (for example, 'A' then 'Z' then 'a').</param>
+        /// <param name="addKeyValueAnnotationsAsText">Whether key value annotations should be text, regardless of the actual input type.</param>
+        /// <param name="keyData">The data view containing the terms. If specified, this should be a single column data
+        /// view, and the key-values will be taken from that column. If unspecified, the key-values will be determined
+        /// from the input data upon fitting.</param>
+        /// <example>
+        /// <format type="text/markdown">
+        /// <![CDATA[
+        /// [!code-csharp[ValueToKey](~/../docs/samples/docs/samples/Microsoft.ML.Samples/Dynamic/KeyToValueValueToKey.cs)]
+        /// ]]>
+        /// </format>
+        /// </example>
+        public static ValueToKeyMappingEstimator MapValueToKey(this TransformsCatalog.ConversionTransforms catalog,
+            InputOutputColumnPair[] columns,
+            int maximumNumberOfKeys = ValueToKeyMappingEstimator.Defaults.MaximumNumberOfKeys,
+            ValueToKeyMappingEstimator.KeyOrdinality keyOrdinality = ValueToKeyMappingEstimator.Defaults.Ordinality,
+            bool addKeyValueAnnotationsAsText = ValueToKeyMappingEstimator.Defaults.AddKeyValueAnnotationsAsText,
+            IDataView keyData = null)
+        {
+            var columnOptions = columns.Select(x => new ValueToKeyMappingEstimator.ColumnOptions(x.OutputColumnName, x.InputColumnName, maximumNumberOfKeys, keyOrdinality, addKeyValueAnnotationsAsText)).ToArray();
+            return new ValueToKeyMappingEstimator(CatalogUtils.GetEnvironment(catalog), columnOptions, keyData);
+        }
+
         /// <summary>
         /// Converts value types into <see cref="KeyType"/>, optionally loading the keys to use from <paramref name="keyData"/>.
         /// </summary>

src/Microsoft.ML.Data/Transforms/ExtensionsCatalog.cs

@@ -8,6 +8,32 @@
 
 namespace Microsoft.ML
 {
+    /// <summary>
+    /// Specifies input and output column names for a transformation.
+    /// </summary>
+    public sealed class InputOutputColumnPair
+    {
+        /// <summary>
+        /// Name of the column to transform. If set to <see langword="null"/>, the value of the <see cref="OutputColumnName"/> will be used as source.
+        /// </summary>
+        public readonly string InputColumnName;
+        /// <summary>
+        /// Name of the column resulting from the transformation of <see cref="InputColumnName"/>.
+        /// </summary>
+        public readonly string OutputColumnName;
+
+        /// <summary>
+        /// Specifies input and output column names for a transformation.
+        /// </summary>
+        /// <param name="outputColumnName">Name of the column resulting from the transformation of <paramref name="inputColumnName"/>.</param>
+        /// <param name="inputColumnName">Name of the column to transform. If set to <see langword="null"/>, the value of the <paramref name="outputColumnName"/> will be used as source.</param>
+        public InputOutputColumnPair(string outputColumnName, string inputColumnName = null)
+        {
+            InputColumnName = inputColumnName;
+            OutputColumnName = outputColumnName;
+        }
+    }
+
     /// <summary>
     /// Specifies input and output column names for a transformation.
     /// </summary>

src/Microsoft.ML.Transforms/CategoricalCatalog.cs

@@ -2,6 +2,7 @@
 // The .NET Foundation licenses this file to you under the MIT license.
 // See the LICENSE file in the project root for more information.
 
+using System.Linq;
 using Microsoft.ML.Data;
 using Microsoft.ML.Transforms;
 
@@ -40,6 +41,34 @@ public static OneHotEncodingEstimator OneHotEncoding(this TransformsCatalog.Cate
             => new OneHotEncodingEstimator(CatalogUtils.GetEnvironment(catalog),
                 new[] { new OneHotEncodingEstimator.ColumnOptions(outputColumnName, inputColumnName, outputKind, maximumNumberOfKeys, keyOrdinality) }, keyData);
 
+        /// <summary>
+        /// Convert text columns into one-hot encoded vectors.
+        /// </summary>
+        /// <param name="catalog">The transform catalog</param>
+        /// <param name="columns">Specifies the names of the columns on which to apply the transformation.</param>
+        /// <param name="outputKind">Output kind: Bag (multi-set vector), Ind (indicator vector), Key (index), or Binary encoded indicator vector.</param>
+        /// <param name="maximumNumberOfKeys">Maximum number of terms to keep per column when auto-training.</param>
+        /// <param name="keyOrdinality">How items should be ordered when vectorized. If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByOccurrence"/> choosen they will be in the order encountered.
+        /// If <see cref="ValueToKeyMappingEstimator.KeyOrdinality.ByValue"/>, items are sorted according to their default comparison, for example, text sorting will be case sensitive (for example, 'A' then 'Z' then 'a').</param>
+        /// <param name="keyData">Specifies an ordering for the encoding. If specified, this should be a single column data view,
+        /// and the key-values will be taken from that column. If unspecified, the ordering will be determined from the input data upon fitting.</param>
+        /// <example>
+        /// <format type="text/markdown">
+        /// <![CDATA[
+        ///  [!code-csharp[RPCA](~/../docs/samples/docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Categorical/OneHotEncoding.cs)]
+        /// ]]></format>
+        /// </example>
+        public static OneHotEncodingEstimator OneHotEncoding(this TransformsCatalog.CategoricalTransforms catalog,
+                InputOutputColumnPair[] columns,
+                OneHotEncodingEstimator.OutputKind outputKind = OneHotEncodingEstimator.Defaults.OutKind,
+                int maximumNumberOfKeys = ValueToKeyMappingEstimator.Defaults.MaximumNumberOfKeys,
+                ValueToKeyMappingEstimator.KeyOrdinality keyOrdinality = ValueToKeyMappingEstimator.Defaults.Ordinality,
+                IDataView keyData = null)
+        {
+            var columnOptions = columns.Select(x => new OneHotEncodingEstimator.ColumnOptions(x.OutputColumnName, x.InputColumnName, outputKind, maximumNumberOfKeys, keyOrdinality)).ToArray();
+            return new OneHotEncodingEstimator(CatalogUtils.GetEnvironment(catalog), columnOptions, keyData);
+        }
+
         /// <summary>
         /// Convert several text column into one-hot encoded vectors.
         /// </summary>
@@ -88,6 +117,31 @@ public static OneHotHashEncodingEstimator OneHotHashEncoding(this TransformsCata
             => new OneHotHashEncodingEstimator(CatalogUtils.GetEnvironment(catalog),
                 new[] { new OneHotHashEncodingEstimator.ColumnOptions(outputColumnName, inputColumnName, outputKind, numberOfBits, seed, useOrderedHashing, maximumNumberOfInverts) });
 
+        /// <summary>
+        /// Convert text columns into hash-based one-hot encoded vector columns.
+        /// </summary>
+        /// <param name="catalog">The transform catalog</param>
+        /// <param name="columns">Specifies the names of the columns on which to apply the transformation.</param>
+        /// <param name="outputKind">The conversion mode.</param>
+        /// <param name="numberOfBits">Number of bits to hash into. Must be between 1 and 30, inclusive.</param>
+        /// <param name="seed">Hashing seed.</param>
+        /// <param name="useOrderedHashing">Whether the position of each term should be included in the hash.</param>
+        /// <param name="maximumNumberOfInverts">During hashing we constuct mappings between original values and the produced hash values.
+        /// Text representation of original values are stored in the slot names of the  metadata for the new column.Hashing, as such, can map many initial values to one.
+        /// <paramref name="maximumNumberOfInverts"/> specifies the upper bound of the number of distinct input values mapping to a hash that should be retained.
+        /// <value>0</value> does not retain any input values. <value>-1</value> retains all input values mapping to each hash.</param>
+        public static OneHotHashEncodingEstimator OneHotHashEncoding(this TransformsCatalog.CategoricalTransforms catalog,
+                InputOutputColumnPair[] columns,
+                OneHotEncodingEstimator.OutputKind outputKind = OneHotEncodingEstimator.OutputKind.Indicator,
+                int numberOfBits = OneHotHashEncodingEstimator.Defaults.NumberOfBits,
+                uint seed = OneHotHashEncodingEstimator.Defaults.Seed,
+                bool useOrderedHashing = OneHotHashEncodingEstimator.Defaults.UseOrderedHashing,
+                int maximumNumberOfInverts = OneHotHashEncodingEstimator.Defaults.MaximumNumberOfInverts)
+        {
+            var columnOptions = columns.Select(x => new OneHotHashEncodingEstimator.ColumnOptions(x.OutputColumnName, x.InputColumnName, outputKind, numberOfBits, seed, useOrderedHashing, maximumNumberOfInverts)).ToArray();
+            return new OneHotHashEncodingEstimator(CatalogUtils.GetEnvironment(catalog), columnOptions);
+        }
+
         /// <summary>
         /// Convert several text column into hash-based one-hot encoded vectors.
         /// </summary>

src/Microsoft.ML.Transforms/ExtensionsCatalog.cs

@@ -2,24 +2,14 @@
 // The .NET Foundation licenses this file to you under the MIT license.
 // See the LICENSE file in the project root for more information.
 
+using System.Linq;
 using Microsoft.ML.Data;
 using Microsoft.ML.Transforms;
 
 namespace Microsoft.ML
 {
     public static class ExtensionsCatalog
     {
-        /// <summary>
-        /// Creates a new output column, of boolean type, with the same number of slots as the input column. The value in the output column
-        /// is true if the value in the input column is missing.
-        /// </summary>
-        /// <param name="catalog">The transform extensions' catalog.</param>
-        /// <param name="columns">The names of the input columns of the transformation and the corresponding names for the output columns.</param>
-        [BestFriend]
-        internal static MissingValueIndicatorEstimator IndicateMissingValues(this TransformsCatalog catalog,
-            params ColumnOptions[] columns)
-            => new MissingValueIndicatorEstimator(CatalogUtils.GetEnvironment(catalog), ColumnOptions.ConvertToValueTuples(columns));
-
         /// <summary>
         /// Creates a new output column, or replaces the source with a new column
         /// (depending on whether the <paramref name="inputColumnName"/> is given a value, or left to null)
@@ -41,6 +31,21 @@ public static MissingValueIndicatorEstimator IndicateMissingValues(this Transfor
             string inputColumnName = null)
             => new MissingValueIndicatorEstimator(CatalogUtils.GetEnvironment(catalog), outputColumnName, inputColumnName);
 
+        /// <summary>
+        /// Creates a new output column, of boolean type, with the same number of slots as the input column. The value in the output column
+        /// is true if the value in the input column is missing.
+        /// </summary>
+        /// <param name="catalog">The transform extensions' catalog.</param>
+        /// <param name="columns">Specifies the names of the columns on which to apply the transformation.</param>
+        /// <example>
+        /// <format type="text/markdown">
+        /// <![CDATA[
+        ///  [!code-csharp[RPCA](~/../docs/samples/docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/IndicateMissingValues.cs)]
+        /// ]]></format>
+        /// </example>
+        public static MissingValueIndicatorEstimator IndicateMissingValues(this TransformsCatalog catalog, InputOutputColumnPair[] columns)
+            => new MissingValueIndicatorEstimator(CatalogUtils.GetEnvironment(catalog), columns.Select(x => (x.OutputColumnName, x.InputColumnName)).ToArray());
+
         /// <summary>
         /// Creates a new output column, or replaces the source with a new column
         /// (depending on whether the <paramref name="outputColumnName"/> is given a value, or left to null)
@@ -69,6 +74,31 @@ public static MissingValueReplacingEstimator ReplaceMissingValues(this Transform
             bool imputeBySlot = MissingValueReplacingEstimator.Defaults.ImputeBySlot)
         => new MissingValueReplacingEstimator(CatalogUtils.GetEnvironment(catalog), new[] { new MissingValueReplacingEstimator.ColumnOptions(outputColumnName, inputColumnName, replacementMode, imputeBySlot) });
 
+        /// <summary>
+        /// Creates a new output column, identical to the input column for everything but the missing values.
+        /// The missing values of the input column, in this new column are replaced with <see cref="MissingValueReplacingEstimator.ReplacementMode.DefaultValue"/>.
+        /// </summary>
+        /// <param name="catalog">The transform extensions' catalog.</param>
+        /// <param name="columns">Specifies the names of the columns on which to apply the transformation.</param>
+        /// <param name="replacementMode">The type of replacement to use as specified in <see cref="MissingValueReplacingEstimator.ReplacementMode"/></param>
+        /// <param name="imputeBySlot">If true, per-slot imputation of replacement is performed.
+        /// Otherwise, replacement value is imputed for the entire vector column. This setting is ignored for scalars and variable vectors,
+        /// where imputation is always for the entire column.</param>
+        /// <example>
+        /// <format type="text/markdown">
+        /// <![CDATA[
+        ///  [!code-csharp[RPCA](~/../docs/samples/docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/ReplaceMissingValues.cs)]
+        /// ]]></format>
+        /// </example>
+        public static MissingValueReplacingEstimator ReplaceMissingValues(this TransformsCatalog catalog,
+            InputOutputColumnPair[] columns,
+            MissingValueReplacingEstimator.ReplacementMode replacementMode = MissingValueReplacingEstimator.Defaults.Mode,
+            bool imputeBySlot = MissingValueReplacingEstimator.Defaults.ImputeBySlot)
+        {
+            var columnOptions = columns.Select(x => new MissingValueReplacingEstimator.ColumnOptions(x.OutputColumnName, x.InputColumnName, replacementMode, imputeBySlot)).ToArray();
+            return new MissingValueReplacingEstimator(CatalogUtils.GetEnvironment(catalog), columnOptions);
+        }
+
         /// <summary>
         /// Creates a new output column, identical to the input column for everything but the missing values.
         /// The missing values of the input column, in this new column are replaced with <see cref="MissingValueReplacingEstimator.ReplacementMode.DefaultValue"/>.