some refactor in MatrixFactorization, plus fix small bugs

Directory.Build.props

@@ -19,6 +19,7 @@
       https://dotnetfeed.blob.core.windows.net/dotnet-core/index.json;
       https://dotnet.myget.org/F/dotnet-core/api/v3/index.json;
       https://dotnet.myget.org/F/roslyn-analyzers/api/v3/index.json;
+      C:\Users\xiaoyuz\source\repos\image\machinelearning\bin\packages;
     </RestoreSources>
   </PropertyGroup>
 

docs/samples/Microsoft.ML.AutoML.Samples/DataStructures/Movie.cs

@@ -0,0 +1,22 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+using Microsoft.ML.Data;
+
+namespace Microsoft.ML.AutoML.Samples.DataStructures
+{
+    public class Movie
+    {
+        [LoadColumn(0)]
+        public string userId;
+
+        [LoadColumn(1)]
+        public string movieId;
+
+        [LoadColumn(2)]
+        public float rating;
+
+        [LoadColumn(3)]
+        public float timestamp;
+    }
+}

docs/samples/Microsoft.ML.AutoML.Samples/RecommendationExperiment.cs

@@ -2,6 +2,7 @@
 using System.IO;
 using System.Linq;
 using Microsoft.ML.AutoML;
+using Microsoft.ML.AutoML.Samples.DataStructures;
 using Microsoft.ML.Data;
 
 namespace Microsoft.ML.AutoML.Samples
@@ -19,8 +20,8 @@ public static void Run()
             MLContext mlContext = new MLContext();
 
             // STEP 1: Load data
-            IDataView trainDataView = mlContext.Data.LoadFromTextFile<TaxiTrip>(TrainDataPath, hasHeader: true, separatorChar: ',');
-            IDataView testDataView = mlContext.Data.LoadFromTextFile<TaxiTrip>(TestDataPath, hasHeader: true, separatorChar: ',');
+            IDataView trainDataView = mlContext.Data.LoadFromTextFile<Movie>(TrainDataPath, hasHeader: true, separatorChar: ',');
+            IDataView testDataView = mlContext.Data.LoadFromTextFile<Movie>(TestDataPath, hasHeader: true, separatorChar: ',');
 
             var settings = new RecommendationExperimentSettings(RecommendationExperimentScenario.MF, "userId", "movieId");
             // STEP 2: Run AutoML experiment
@@ -47,17 +48,13 @@ public static void Run()
                 mlContext.Model.Save(bestRun.Model, trainDataView.Schema, fs);
 
             // STEP 7: Create prediction engine from the best trained model
-            var predictionEngine = mlContext.Model.CreatePredictionEngine<TaxiTrip, TaxiTripFarePrediction>(bestRun.Model);
+            var predictionEngine = mlContext.Model.CreatePredictionEngine<Movie, TaxiTripFarePrediction>(bestRun.Model);
 
             // STEP 8: Initialize a new test taxi trip, and get the predicted fare
-            var testTaxiTrip = new TaxiTrip
+            var testTaxiTrip = new Movie
             {
-                VendorId = "VTS",
-                RateCode = 1,
-                PassengerCount = 1,
-                TripTimeInSeconds = 1140,
-                TripDistance = 3.75f,
-                PaymentType = "CRD"
+                userId="1",
+                movieId = "1097",
             };
             var prediction = predictionEngine.Predict(testTaxiTrip);
             Console.WriteLine($"Predicted fare for test taxi trip: {prediction.FareAmount}");

src/Microsoft.ML.AutoML/API/AutoCatalog.cs

@@ -17,7 +17,7 @@ public sealed class AutoCatalog
         /// <summary>
         /// save some intermediate value
         /// </summary>
-        public static Dictionary<string, object> ValuePairs { get; set; }
+        public static Dictionary<string, object> ValuePairs { get; set; } = new Dictionary<string, object>();
 
         internal AutoCatalog(MLContext context)
         {

src/Microsoft.ML.AutoML/Microsoft.ML.AutoML.csproj

@@ -11,9 +11,9 @@
   </PropertyGroup>
 
   <ItemGroup>
-    <PackageReference Include="Microsoft.ML" Version="$(MlDotNetPackageVersion)" />
     <PackageReference Include="Microsoft.ML.LightGBM" Version="$(MlDotNetPackageVersion)" />
     <PackageReference Include="Microsoft.ML.Mkl.Components" Version="$(MlDotNetPackageVersion)" />
+    <PackageReference Include="Microsoft.ML.Recommender" Version="0.16.0-preview2-28122-0" />
   </ItemGroup>
 
   <PropertyGroup>

src/Microsoft.ML.AutoML/TrainerExtensions/TrainerExtensionCatalog.cs

@@ -39,7 +39,8 @@ internal class TrainerExtensionCatalog
                 { TrainerName.SgdCalibratedBinary, typeof(SgdCalibratedBinaryExtension) },
                 { TrainerName.SgdCalibratedOva, typeof(SgdCalibratedOvaExtension) },
                 { TrainerName.SymbolicSgdLogisticRegressionBinary, typeof(SymbolicSgdLogisticRegressionBinaryExtension) },
-                { TrainerName.SymbolicSgdLogisticRegressionOva, typeof(SymbolicSgdLogisticRegressionOvaExtension) }
+                { TrainerName.SymbolicSgdLogisticRegressionOva, typeof(SymbolicSgdLogisticRegressionOvaExtension) },
+                { TrainerName.MatrixFactorization, typeof(MatrixFactorizationExtension) },
             };
 
         private static readonly IDictionary<Type, TrainerName> _extensionTypesToTrainerNames =
@@ -72,6 +73,10 @@ public static IEnumerable<ITrainerExtension> GetTrainers(TaskKind task,
             {
                 trainers = GetRegressionLearners();
             }
+            else if (task == TaskKind.Recommendation)
+            {
+                trainers = GetRecommendationLearners();
+            }
             else
             {
                 // should not be possible to reach here
@@ -134,5 +139,13 @@ private static IEnumerable<ITrainerExtension> GetRegressionLearners()
                 new OlsRegressionExtension(),
             };
         }
+
+        private static IEnumerable<ITrainerExtension> GetRecommendationLearners()
+        {
+            return new ITrainerExtension[]
+            {
+                new MatrixFactorizationExtension(),
+            };
+        }
     }
 }

src/Microsoft.ML.AutoML/Utils/UserInputValidationUtil.cs

@@ -262,7 +262,7 @@ private static IEnumerable<DataViewType> GetAllowedLabelTypes(TaskKind task)
                 // (if input label is not already a key) before invoking the trainer.
                 case TaskKind.MulticlassClassification:
                     return null;
-                case TaskKind.Regression:
+                case TaskKind.Regression | TaskKind.Recommendation:
                     return new DataViewType[] { NumberDataViewType.Single };
                 default:
                     throw new NotSupportedException($"Unsupported task type: {task}");

src/Microsoft.ML.Recommender/MatrixFactorizationPredictor.cs

@@ -422,14 +422,18 @@ DataViewRow ISchemaBoundRowMapper.GetRow(DataViewRow input, IEnumerable<DataView
     /// <summary>
     /// Trains a <see cref="MatrixFactorizationModelParameters"/>. It factorizes the training matrix into the product of two low-rank matrices.
     /// </summary>
-    public sealed class MatrixFactorizationPredictionTransformer : PredictionTransformerBase<MatrixFactorizationModelParameters>
+    public sealed class MatrixFactorizationPredictionTransformer : PredictionTransformerBase<MatrixFactorizationModelParameters>, ISingleFeaturePredictionTransformer<MatrixFactorizationModelParameters>
     {
         internal const string LoaderSignature = "MaFactPredXf";
         internal string MatrixColumnIndexColumnName { get; }
         internal string MatrixRowIndexColumnName { get; }
         internal DataViewType MatrixColumnIndexColumnType { get; }
         internal DataViewType MatrixRowIndexColumnType { get; }
 
+        public string FeatureColumnName => throw new NotImplementedException();
+
+        public DataViewType FeatureColumnType => throw new NotImplementedException();
+
         /// <summary>
         /// Build a transformer based on matrix factorization predictor (model) and the input schema (trainSchema). The created
         /// transformer can only transform IDataView objects compatible to the input schema; that is, that IDataView must contain

src/Microsoft.ML.Recommender/MatrixFactorizationTrainer.cs

@@ -84,7 +84,7 @@ namespace Microsoft.ML.Trainers
     /// </format>
     /// </example>
     public sealed class MatrixFactorizationTrainer : ITrainer<MatrixFactorizationModelParameters>,
-        IEstimator<MatrixFactorizationPredictionTransformer>
+        IEstimator<MatrixFactorizationPredictionTransformer>, ITrainerEstimator<MatrixFactorizationPredictionTransformer, MatrixFactorizationModelParameters>
     {
         /// <summary>
         /// Type of loss function.
@@ -111,7 +111,7 @@ public enum LossFunctionType
         /// <summary>
         /// Advanced options for the <see cref="MatrixFactorizationTrainer"/>.
         /// </summary>
-        public sealed class Options
+        public sealed class Options : TrainerInputBaseWithLabel
         {
             /// <summary>
             /// The name of variable (i.e., Column in a <see cref="IDataView"/> type system) used as matrix's column index.
@@ -123,11 +123,6 @@ public sealed class Options
             /// </summary>
             public string MatrixRowIndexColumnName;
 
-            /// <summary>
-            /// The name variable (i.e., column in a <see cref="IDataView"/> type system) used as matrix's element value.
-            /// </summary>
-            public string LabelColumnName;
-
             /// <summary>
             /// Loss function minimized for finding factor matrices.
             /// </summary>
@@ -302,9 +297,8 @@ internal static class Defaults
         /// <summary>
         /// The <see cref="TrainerInfo"/> contains general parameters for this trainer.
         /// </summary>
-        TrainerInfo ITrainer.Info => _info;
 
-        private readonly TrainerInfo _info;
+        public TrainerInfo Info { get; set; }
 
         /// <summary>
         /// Initializes a new instance of <see cref="MatrixFactorizationTrainer"/> through the <see cref="Options"/> class.
@@ -337,7 +331,7 @@ internal MatrixFactorizationTrainer(IHostEnvironment env, Options options)
             _quiet = options.Quiet;
             _doNmf = options.NonNegative;
 
-            _info = new TrainerInfo(normalization: false, caching: false);
+            Info = new TrainerInfo(normalization: false, caching: false);
 
             LabelName = options.LabelColumnName;
             MatrixColumnIndexName = options.MatrixColumnIndexColumnName;
@@ -379,7 +373,7 @@ internal MatrixFactorizationTrainer(IHostEnvironment env,
             _quiet = args.Quiet;
             _doNmf = args.NonNegative;
 
-            _info = new TrainerInfo(normalization: false, caching: false);
+            Info = new TrainerInfo(normalization: false, caching: false);
 
             LabelName = labelColumnName;
             MatrixColumnIndexName = matrixColumnIndexColumnName;

src/Microsoft.ML.TensorFlow/TensorFlow/TensorGeneric.cs

@@ -3,7 +3,6 @@
 
 using System;
 using System.Text;
-using Microsoft.ML.Runtime;
 
 namespace Microsoft.ML.Transforms.TensorFlow
 {