Multi column MapKeyToValue and MapValueToKey

docs/samples/Microsoft.ML.Samples/Dynamic/Trainers/MulticlassClassification/StochasticDualCoordinateAscent.cs

@@ -1,5 +1,4 @@
-using Microsoft.ML.Data;
-using Microsoft.ML.SamplesUtils;
+using Microsoft.ML.SamplesUtils;
 
 namespace Microsoft.ML.Samples.Dynamic.Trainers.MulticlassClassification
 {

docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Conversion/ConvertType.cs

@@ -1,7 +1,8 @@
 using System;
+using Microsoft.ML;
 using Microsoft.ML.Data;
 
-namespace Microsoft.ML.Samples.Dynamic
+namespace Samples.Dynamic
 {
     public static class ConvertType
     {
@@ -39,10 +40,12 @@ public static void Example()
             // A: False    Aconv:0
             // A: False    Aconv:0
         }
+
         private class InputData
         {
             public bool Survived;
         }
+
         private sealed class TransformedData : InputData
         {
             public Int32 SurvivedInt32 { get; set; }

docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Conversion/ConvertTypeMultiColumn.cs

@@ -1,7 +1,8 @@
 using System;
+using Microsoft.ML;
 using Microsoft.ML.Data;
 
-namespace Microsoft.ML.Samples.Dynamic
+namespace Samples.Dynamic
 {
     // This example illustrates how to convert multiple columns of different types to one type, in this case System.Single. 
     // This is often a useful data transformation before concatenating the features together and passing them to a particular estimator.
@@ -58,6 +59,7 @@ public static void Example()
             //      1       8904        6.368924E+17        8.09
 
         }
+
         // The initial data type
         private class InputData
         {
@@ -66,6 +68,7 @@ private class InputData
             public DateTime Feature3;
             public double Feature4;
         }
+
         // The resulting data type after the transformation
         private class TransformedData : InputData
         {

docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Conversion/Hash.cs

@@ -1,7 +1,8 @@
 using System;
+using Microsoft.ML;
 using Microsoft.ML.Data;
 
-namespace Microsoft.ML.Samples.Dynamic
+namespace Samples.Dynamic
 {
     // This example demonstrates hashing of categorical string and integer data types.
     public static class Hash

docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Conversion/KeyToValueToKey.cs

@@ -1,11 +1,13 @@
 using System;
 using System.Collections.Generic;
+using Microsoft.ML;
 using Microsoft.ML.Data;
+using Microsoft.ML.SamplesUtils;
 using Microsoft.ML.Transforms;
 
-namespace Microsoft.ML.Samples.Dynamic
+namespace Samples.Dynamic
 {
-    public class KeyToValueValueToKey
+    public class KeyToValueToKey
     {
         public static void Example()
         {
@@ -14,7 +16,7 @@ public static void Example()
             var ml = new MLContext();
 
             // Get a small dataset as an IEnumerable and load it into ML.NET data set.
-            IEnumerable<SamplesUtils.DatasetUtils.SampleTopicsData> data = SamplesUtils.DatasetUtils.GetTopicsData();
+            IEnumerable<DatasetUtils.SampleTopicsData> data = DatasetUtils.GetTopicsData();
             var trainData = ml.Data.LoadFromEnumerable(data);
 
             // Preview of one of the columns of the the topics data. 

docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Conversion/MapKeyToValueMultiColumn.cs

@@ -0,0 +1,76 @@
+using System;
+using Microsoft.ML;
+using Microsoft.ML.SamplesUtils;
+
+namespace Samples.Dynamic
+{
+    public class MapKeyToValueMultiColumn
+    {
+        public static void Example()
+        {
+            // Create a new context for ML.NET operations. It can be used for exception tracking and logging, 
+            // as a catalog of available operations and as the source of randomness.
+            // Setting the seed to a fixed number in this example to make outputs deterministic.
+            var mlContext = new MLContext(seed: 0);
+
+            // Create a list of data examples.
+            var examples = DatasetUtils.GenerateRandomMulticlassClassificationExamples(1000);
+
+            // Convert the examples list to an IDataView object, which is consumable by ML.NET API.
+            var dataView = mlContext.Data.LoadFromEnumerable(examples);
+
+            //////////////////// Data Preview ////////////////////
+            // Label    Features
+            // AA       0.7262433,0.8173254,0.7680227,0.5581612,0.2060332,0.5588848,0.9060271,0.4421779,0.9775497,0.2737045
+            // BB       0.4919063,0.6673147,0.8326591,0.6695119,1.182151,0.230367,1.06237,1.195347,0.8771811,0.5145918
+            // CC       1.216908,1.248052,1.391902,0.4326252,1.099942,0.9262842,1.334019,1.08762,0.9468155,0.4811099
+            // DD       0.7871246,1.053327,0.8971719,1.588544,1.242697,1.362964,0.6303943,0.9810045,0.9431419,1.557455
+
+            // Create a pipeline. 
+            var pipeline =
+                    // Convert the string labels into key types.
+                    mlContext.Transforms.Conversion.MapValueToKey("Label")
+                    // Apply StochasticDualCoordinateAscent multiclass trainer.
+                    .Append(mlContext.MulticlassClassification.Trainers.SdcaMaximumEntropy());
+
+            // Train the model and do predictions on same data set. 
+            // Typically predictions would be in a different, validation set. 
+            var dataWithPredictions = pipeline.Fit(dataView).Transform(dataView);
+
+            // at this point, the Label colum is tranformed from strings, to DataViewKeyType and
+            // the transformation has added the PredictedLabel column, with 
+            var newPipeline = mlContext.Transforms.Conversion.MapKeyToValue(new[] 
+            {
+                new InputOutputColumnPair("LabelOriginalValue","Label"),
+                new InputOutputColumnPair("PredictedLabelOriginalValue","PredictedLabel")
+            });
+
+            var transformedData = newPipeline.Fit(dataWithPredictions).Transform(dataWithPredictions);
+
+            var values = mlContext.Data.CreateEnumerable<TransformedData>(transformedData, reuseRowObject: false);
+
+            // Printing the columns of the transformed data. 
+            Console.WriteLine($" Label   LabelOriginalValue   PredictedLabel   PredictedLabelOriginalValue");
+            foreach (var row in values)
+                Console.WriteLine($"{row.Label}\t\t{row.LabelOriginalValue}\t\t\t{row.PredictedLabel}\t\t\t{row.PredictedLabelOriginalValue}");
+
+            // Label LabelOriginalValue   PredictedLabel   PredictedLabelOriginalValue
+            //  1           AA                  2                   BB
+            //  1           AA                  1                   AA
+            //  4           DD                  4                   DD
+            //  2           BB                  2                   BB
+            //  1           AA                  1                   AA
+            //  1           AA                  1                   AA
+            //  1           AA                  1                   AA
+            //  2           BB                  2                   BB
+
+        }
+        private class TransformedData
+        {
+            public uint Label { get; set; }
+            public uint PredictedLabel { get; set; }
+            public string LabelOriginalValue { get; set; }
+            public string PredictedLabelOriginalValue { get; set; }
+        }
+    }
+}

docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Conversion/MapValue.cs

@@ -1,9 +1,8 @@
 using System;
 using System.Collections.Generic;
-using Microsoft.ML.Data;
+using Microsoft.ML;
 
-
-namespace Microsoft.ML.Samples.Dynamic
+namespace Samples.Dynamic
 {
     public static class MapValue
     {
@@ -79,11 +78,13 @@ public static void Example()
             // 12-25yrs       Long               3       5       High
             // 25+yrs         Long               3       5       High
         }
+
         private class DataPoint
         {
             public string Timeframe { get; set; }
             public int Score { get; set; }
         }
+
         private class TransformedData : DataPoint
         {
             public string TimeframeCategory { get; set; }

docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Conversion/MapValueIdvLookup.cs

@@ -1,7 +1,8 @@
 using System;
 using System.Collections.Generic;
+using Microsoft.ML;
 
-namespace Microsoft.ML.Samples.Dynamic
+namespace Samples.Dynamic
 {
     public static class MapValueIdvLookup
     {
@@ -68,10 +69,12 @@ private class LookupMap
             public float Value { get; set; }
             public string Category { get; set; }
         }
+
         private class DataPoint
         {
             public float Price { get; set; }
         }
+
         private class TransformedData : DataPoint
         {
             public string PriceCategory { get; set; }

docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Conversion/MapValueToArray.cs

@@ -1,6 +1,8 @@
 using System;
 using System.Collections.Generic;
-namespace Microsoft.ML.Samples.Dynamic
+using Microsoft.ML;
+
+namespace Samples.Dynamic
 {
     public static class MapValueToArray
     { 
@@ -55,10 +57,12 @@ public static void Example()
             // 12-25yrs     12, 50,300
             // 25+yrs       12, 50, 300
         }
+
         public class DataPoint
         {
             public string Timeframe { get; set; }
         }
+
         public class TransformedData : DataPoint
         {
             public int[] Features { get; set; }

docs/samples/Microsoft.ML.Samples/Dynamic/Transforms/Conversion/MapValueToKeyMultiColumn.cs

@@ -0,0 +1,115 @@
+using System;
+using System.Collections.Generic;
+using Microsoft.ML;
+
+namespace Samples.Dynamic
+{
+    public static class MapValueToKeyMultiColumn
+    {
+        /// This example demonstrates the use of the ValueMappingEstimator by mapping strings to other string values, or floats to strings. 
+        /// This is useful to map types to a grouping. 
+        /// It is possible to have multiple values map to the same category.
+        public static void Example()
+        {
+            // Create a new ML context, for ML.NET operations. It can be used for exception tracking and logging, 
+            // as well as the source of randomness.
+            var mlContext = new MLContext();
+
+            // Get a small dataset as an IEnumerable.
+            var rawData = new[] {
+                new DataPoint() { StudyTime = "0-4yrs" , Course = "CS" },
+                new DataPoint() { StudyTime = "6-11yrs" , Course = "CS" },
+                new DataPoint() { StudyTime = "12-25yrs" , Course = "LA" },
+                new DataPoint() { StudyTime = "0-5yrs" , Course = "DS" }
+            };
+
+            var data = mlContext.Data.LoadFromEnumerable(rawData);
+
+            // Constructs the ML.net pipeline
+            var pipeline = mlContext.Transforms.Conversion.MapValueToKey(new[] {
+                new  InputOutputColumnPair("StudyTimeCategory", "StudyTime"),
+                new  InputOutputColumnPair("CourseCategory", "Course")
+                },
+                keyOrdinality: Microsoft.ML.Transforms.ValueToKeyMappingEstimator.KeyOrdinality.ByValue,
+                addKeyValueAnnotationsAsText: true);
+
+            // Fits the pipeline to the data.
+            IDataView transformedData = pipeline.Fit(data).Transform(data);
+
+            // Getting the resulting data as an IEnumerable.
+            // This will contain the newly created columns.
+            IEnumerable<TransformedData> features = mlContext.Data.CreateEnumerable<TransformedData>(transformedData, reuseRowObject: false);
+
+            Console.WriteLine($" StudyTime   StudyTimeCategory   Course    CourseCategory");
+            foreach (var featureRow in features)
+                Console.WriteLine($"{featureRow.StudyTime}\t\t{featureRow.StudyTimeCategory}\t\t\t{featureRow.Course}\t\t{featureRow.CourseCategory}");
+
+            // TransformedData obtained post-transformation.
+            //
+            //  StudyTime   StudyTimeCategory   DevelopmentTime    DevelopmentTimeCategory
+            // 0-4yrs          1                6-11yrs                3
+            // 6-11yrs         4                6-11yrs                3
+            // 12-25yrs        3                25+yrs                 2
+            // 0-5yrs          2                0-5yrs                 1
+
+            // If we wanted to provide the mapping, rather than letting the transform create it, 
+            // we could do so by creating an IDataView one column containing the values to map to. 
+            // If the values in the dataset are not found in the lookup IDataView they will get mapped to the mising value, 0.
+            // The keyData are shared among the columns, therefore the keys are not contiguous for the column. 
+            // Create the lookup map data IEnumerable.  
+            var lookupData = new[] {
+                new LookupMap { Key = "0-4yrs" },
+                new LookupMap { Key = "6-11yrs" },
+                new LookupMap { Key = "25+yrs"  },
+                new LookupMap { Key = "CS" },
+                new LookupMap { Key = "DS" },
+                new LookupMap { Key = "LA"  }
+            };
+
+            // Convert to IDataView
+            var lookupIdvMap = mlContext.Data.LoadFromEnumerable(lookupData);
+
+            // Constructs the ML.net pipeline
+            var pipelineWithLookupMap = mlContext.Transforms.Conversion.MapValueToKey(new[] {
+                new  InputOutputColumnPair("StudyTimeCategory", "StudyTime"),
+                new  InputOutputColumnPair("CourseCategory", "Course")
+                },
+                keyData: lookupIdvMap);
+
+            // Fits the pipeline to the data.
+            transformedData = pipelineWithLookupMap.Fit(data).Transform(data);
+
+            // Getting the resulting data as an IEnumerable.
+            // This will contain the newly created columns.
+            features = mlContext.Data.CreateEnumerable<TransformedData>(transformedData, reuseRowObject: false);
+
+            Console.WriteLine($" StudyTime   StudyTimeCategory   DevelopmentTime    DevelopmentTimeCategory");
+            foreach (var featureRow in features)
+                Console.WriteLine($"{featureRow.StudyTime}\t\t{featureRow.StudyTimeCategory}\t\t\t{featureRow.Course}\t\t{featureRow.CourseCategory}");
+
+            // StudyTime    StudyTimeCategory   DevelopmentTime     DevelopmentTimeCategory
+            // 0 - 4yrs          1                  6 - 11yrs         2
+            // 6 - 11yrs         2                  6 - 11yrs         2
+            // 12 - 25yrs        0                  25 + yrs          3
+            // 0 - 5yrs          0                  0 - 5yrs          0
+        }
+
+        private class DataPoint
+        {
+            public string StudyTime { get; set; }
+            public string Course { get; set; }
+        }
+
+        private class TransformedData : DataPoint
+        {
+            public uint StudyTimeCategory { get; set; }
+            public uint CourseCategory { get; set; }
+        }
+
+        // Type for the IDataView that will be serving as the map
+        private class LookupMap
+        {
+            public string Key { get; set; }
+        }
+    }
+}