Address comments.

1. Rename classes.
2. Use IReadOnlyList instead of ReadOnlySpan.
3. Remove new namespace, Representation.

Author: wschin

src/Microsoft.ML.FastTree/FastTree.cs

@@ -56,7 +56,7 @@ public abstract class FastTreeTrainerBase<TArgs, TTransformer, TModel> :
         protected readonly TArgs Args;
         protected readonly bool AllowGC;
         protected int FeatureCount;
-        private protected TreeEnsemble TrainedEnsemble;
+        private protected InternalTreeEnsemble TrainedEnsemble;
         private protected RoleMappedData ValidData;
         /// <summary>
         /// If not null, it's a test data set passed in from training context. It will be converted to one element in
@@ -88,7 +88,7 @@ public abstract class FastTreeTrainerBase<TArgs, TTransformer, TModel> :
         protected double[] InitValidScores;
         protected double[][] InitTestScores;
         //protected int Iteration;
-        private protected TreeEnsemble Ensemble;
+        private protected InternalTreeEnsemble Ensemble;
 
         protected bool HasValidSet => ValidSet != null;
 
@@ -488,7 +488,7 @@ protected bool AreSamplesWeighted(IChannel ch)
 
         private void InitializeEnsemble()
         {
-            Ensemble = new TreeEnsemble();
+            Ensemble = new InternalTreeEnsemble();
         }
 
         /// <summary>
@@ -793,7 +793,7 @@ private float GetMachineAvailableBytes()
 
         // This method is called at the end of each training iteration, with the tree that was learnt on that iteration.
         // Note that this tree can be null if no tree was learnt this iteration.
-        private protected virtual void CustomizedTrainingIteration(RegressionTree tree)
+        private protected virtual void CustomizedTrainingIteration(InternalRegressionTree tree)
         {
         }
 
@@ -924,7 +924,7 @@ internal abstract class DataConverter
         /// of features we actually trained on. This can be null in the event that no filtering
         /// occurred.
         /// </summary>
-        /// <seealso cref="TreeEnsemble.RemapFeatures"/>
+        /// <seealso cref="InternalTreeEnsemble.RemapFeatures"/>
         public int[] FeatureMap;
 
         protected readonly IHost Host;
@@ -2811,13 +2811,13 @@ public abstract class TreeEnsembleModelParameters :
         ISingleCanSaveOnnx
     {
         /// <summary>
-        /// An ensemble of trees exposed to users. It is a wrapper on the <see langword="internal"/> <see cref="TreeEnsemble"/> in <see cref="ML.FastTree.Representation.TreeRegressorCollection"/>.
+        /// An ensemble of trees exposed to users. It is a wrapper on the <see langword="internal"/> <see cref="InternalTreeEnsemble"/> in <see cref="ML.FastTree.Representation.TreeEnsemble"/>.
         /// </summary>
-        public ML.FastTree.Representation.TreeRegressorCollection TrainedTreeCollection { get; }
+        public ML.FastTree.Representation.TreeEnsemble TrainedTreeCollection { get; }
 
         // The below two properties are necessary for tree Visualizer
         [BestFriend]
-        internal TreeEnsemble TrainedEnsemble => TrainedTreeCollection.TreeEnsemble;
+        internal InternalTreeEnsemble TrainedEnsemble => TrainedTreeCollection.UnderlyingTreeEnsemble;
         int ITreeEnsemble.NumTrees => TrainedEnsemble.NumTrees;
 
         // Inner args is used only for documentation purposes when saving comments to INI files.
@@ -2861,7 +2861,7 @@ public abstract class TreeEnsembleModelParameters :
 
         /// The following function is used in both FastTree and LightGBM so <see cref="BestFriendAttribute"/> is required.
         [BestFriend]
-        internal TreeEnsembleModelParameters(IHostEnvironment env, string name, TreeEnsemble trainedEnsemble, int numFeatures, string innerArgs)
+        internal TreeEnsembleModelParameters(IHostEnvironment env, string name, InternalTreeEnsemble trainedEnsemble, int numFeatures, string innerArgs)
             : base(env, name)
         {
             Host.CheckValue(trainedEnsemble, nameof(trainedEnsemble));
@@ -2871,7 +2871,7 @@ internal TreeEnsembleModelParameters(IHostEnvironment env, string name, TreeEnse
             // REVIEW: When we make the predictor wrapper, we may want to further "optimize"
             // the trained ensemble to, for instance, resize arrays so that they are of the length
             // the actual number of leaves/nodes, or remove unnecessary arrays, and so forth.
-            TrainedTreeCollection = new ML.FastTree.Representation.TreeRegressorCollection(trainedEnsemble);
+            TrainedTreeCollection = new ML.FastTree.Representation.TreeEnsemble(trainedEnsemble);
             InnerArgs = innerArgs;
             NumFeatures = numFeatures;
 
@@ -2899,7 +2899,7 @@ protected TreeEnsembleModelParameters(IHostEnvironment env, string name, ModelLo
             if (ctx.Header.ModelVerWritten >= VerCategoricalSplitSerialized)
                 categoricalSplits = true;
 
-            TrainedTreeCollection = new ML.FastTree.Representation.TreeRegressorCollection(new TreeEnsemble(ctx, usingDefaultValues, categoricalSplits));
+            TrainedTreeCollection = new ML.FastTree.Representation.TreeEnsemble(new InternalTreeEnsemble(ctx, usingDefaultValues, categoricalSplits));
             MaxSplitFeatIdx = TrainedEnsemble.GetMaxFeatureIndex();
 
             InnerArgs = ctx.LoadStringOrNull();
@@ -3202,7 +3202,7 @@ private void SaveEnsembleAsCode(TextWriter writer, RoleMappedSchema schema)
             MetadataUtils.GetSlotNames(schema, RoleMappedSchema.ColumnRole.Feature, NumFeatures, ref names);
 
             int i = 0;
-            foreach (RegressionTree tree in TrainedEnsemble.Trees)
+            foreach (InternalRegressionTree tree in TrainedEnsemble.Trees)
             {
                 writer.Write("double treeOutput{0}=", i);
                 SaveTreeAsCode(tree, writer, in names);
@@ -3218,13 +3218,13 @@ private void SaveEnsembleAsCode(TextWriter writer, RoleMappedSchema schema)
         /// <summary>
         /// Convert a single tree to code, called recursively
         /// </summary>
-        private void SaveTreeAsCode(RegressionTree tree, TextWriter writer, in VBuffer<ReadOnlyMemory<char>> names)
+        private void SaveTreeAsCode(InternalRegressionTree tree, TextWriter writer, in VBuffer<ReadOnlyMemory<char>> names)
         {
             ToCSharp(tree, writer, 0, in names);
         }
 
         // converts a subtree into a C# expression
-        private void ToCSharp(RegressionTree tree, TextWriter writer, int node, in VBuffer<ReadOnlyMemory<char>> names)
+        private void ToCSharp(InternalRegressionTree tree, TextWriter writer, int node, in VBuffer<ReadOnlyMemory<char>> names)
         {
             if (node < 0)
             {
@@ -3308,9 +3308,9 @@ Row ICanGetSummaryAsIRow.GetStatsIRowOrNull(RoleMappedSchema schema)
 
         private sealed class Tree : ITree<VBuffer<Float>>
         {
-            private readonly RegressionTree _regTree;
+            private readonly InternalRegressionTree _regTree;
 
-            public Tree(RegressionTree regTree)
+            public Tree(InternalRegressionTree regTree)
             {
                 _regTree = regTree;
             }

src/Microsoft.ML.FastTree/FastTreeClassification.cs

@@ -70,7 +70,7 @@ private static VersionInfo GetVersionInfo()
 
         protected override uint VerCategoricalSplitSerialized => 0x00010005;
 
-        internal FastTreeBinaryModelParameters(IHostEnvironment env, TreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
+        internal FastTreeBinaryModelParameters(IHostEnvironment env, InternalTreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
             : base(env, RegistrationName, trainedEnsemble, featureCount, innerArgs)
         {
         }
@@ -365,7 +365,7 @@ protected override void GetGradientInOneQuery(int query, int threadIndex)
                 }
             }
 
-            public void AdjustTreeOutputs(IChannel ch, RegressionTree tree,
+            public void AdjustTreeOutputs(IChannel ch, InternalRegressionTree tree,
                 DocumentPartitioning partitioning, ScoreTracker trainingScores)
             {
                 const double epsilon = 1.4e-45;

src/Microsoft.ML.FastTree/FastTreeRanking.cs

@@ -374,7 +374,7 @@ protected override void Train(IChannel ch)
             PrintTestGraph(ch);
         }
 
-        private protected override void CustomizedTrainingIteration(RegressionTree tree)
+        private protected override void CustomizedTrainingIteration(InternalRegressionTree tree)
         {
             Contracts.AssertValueOrNull(tree);
             if (tree != null && Args.CompressEnsemble)
@@ -992,7 +992,7 @@ protected override void GetGradientInOneQuery(int query, int threadIndex)
                 }
             }
 
-            void IStepSearch.AdjustTreeOutputs(IChannel ch, RegressionTree tree, DocumentPartitioning partitioning,
+            void IStepSearch.AdjustTreeOutputs(IChannel ch, InternalRegressionTree tree, DocumentPartitioning partitioning,
                                             ScoreTracker trainingScores)
             {
                 const double epsilon = 1.4e-45;
@@ -1131,7 +1131,7 @@ private static VersionInfo GetVersionInfo()
 
         protected override uint VerCategoricalSplitSerialized => 0x00010005;
 
-        internal FastTreeRankingModelParameters(IHostEnvironment env, TreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
+        internal FastTreeRankingModelParameters(IHostEnvironment env, InternalTreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
             : base(env, RegistrationName, trainedEnsemble, featureCount, innerArgs)
         {
         }

src/Microsoft.ML.FastTree/FastTreeRegression.cs

@@ -416,7 +416,7 @@ public ObjectiveImpl(Dataset trainData, Options options)
                 _labels = GetDatasetRegressionLabels(trainData);
             }
 
-            public void AdjustTreeOutputs(IChannel ch, RegressionTree tree, DocumentPartitioning partitioning, ScoreTracker trainingScores)
+            public void AdjustTreeOutputs(IChannel ch, InternalRegressionTree tree, DocumentPartitioning partitioning, ScoreTracker trainingScores)
             {
                 double shrinkage = LearningRate * Shrinkage;
                 for (int l = 0; l < tree.NumLeaves; ++l)
@@ -467,7 +467,7 @@ private static VersionInfo GetVersionInfo()
 
         protected override uint VerCategoricalSplitSerialized => 0x00010005;
 
-        internal FastTreeRegressionModelParameters(IHostEnvironment env, TreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
+        internal FastTreeRegressionModelParameters(IHostEnvironment env, InternalTreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
             : base(env, RegistrationName, trainedEnsemble, featureCount, innerArgs)
         {
         }

src/Microsoft.ML.FastTree/FastTreeTweedie.cs

@@ -363,7 +363,7 @@ public ObjectiveImpl(Dataset trainData, Options options)
                 _maxClamp = Math.Abs(options.MaxTreeOutput);
             }
 
-            public void AdjustTreeOutputs(IChannel ch, RegressionTree tree, DocumentPartitioning partitioning, ScoreTracker trainingScores)
+            public void AdjustTreeOutputs(IChannel ch, InternalRegressionTree tree, DocumentPartitioning partitioning, ScoreTracker trainingScores)
             {
                 double shrinkage = LearningRate * Shrinkage;
                 var scores = trainingScores.Scores;
@@ -470,7 +470,7 @@ private static VersionInfo GetVersionInfo()
 
         protected override uint VerCategoricalSplitSerialized => 0x00010003;
 
-        internal FastTreeTweedieModelParameters(IHostEnvironment env, TreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
+        internal FastTreeTweedieModelParameters(IHostEnvironment env, InternalTreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
             : base(env, RegistrationName, trainedEnsemble, featureCount, innerArgs)
         {
         }

src/Microsoft.ML.FastTree/GamModelParameters.cs

@@ -437,7 +437,7 @@ private void GetFeatureContributions(in VBuffer<float> features, ref VBuffer<flo
         void ICanSaveInIniFormat.SaveAsIni(TextWriter writer, RoleMappedSchema schema, ICalibrator calibrator)
         {
             Host.CheckValue(writer, nameof(writer), "writer must not be null");
-            var ensemble = new TreeEnsemble();
+            var ensemble = new InternalTreeEnsemble();
 
             for (int featureIndex = 0; featureIndex < NumShapeFunctions; featureIndex++)
             {
@@ -525,11 +525,11 @@ private int CreateBalancedTreeRecursive(int lower, int upper,
             }
         }
 
-        private static RegressionTree CreateRegressionTree(
+        private static InternalRegressionTree CreateRegressionTree(
             int numLeaves, int[] splitFeatures, float[] rawThresholds, int[] lteChild, int[] gtChild, double[] leafValues)
         {
             var numInternalNodes = numLeaves - 1;
-            return RegressionTree.Create(
+            return InternalRegressionTree.Create(
                 numLeaves: numLeaves,
                 splitFeatures: splitFeatures,
                 rawThresholds: rawThresholds,

src/Microsoft.ML.FastTree/RandomForestClassification.cs

@@ -80,7 +80,7 @@ private static VersionInfo GetVersionInfo()
         /// </summary>
         public override PredictionKind PredictionKind => PredictionKind.BinaryClassification;
 
-        internal FastForestClassificationModelParameters(IHostEnvironment env, TreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
+        internal FastForestClassificationModelParameters(IHostEnvironment env, InternalTreeEnsemble trainedEnsemble, int featureCount, string innerArgs)
             : base(env, RegistrationName, trainedEnsemble, featureCount, innerArgs)
         { }
 

src/Microsoft.ML.FastTree/RandomForestRegression.cs

@@ -60,7 +60,7 @@ private static VersionInfo GetVersionInfo()
 
         protected override uint VerCategoricalSplitSerialized => 0x00010006;
 
-        internal FastForestRegressionModelParameters(IHostEnvironment env, TreeEnsemble trainedEnsemble, int featureCount, string innerArgs, int samplesCount)
+        internal FastForestRegressionModelParameters(IHostEnvironment env, InternalTreeEnsemble trainedEnsemble, int featureCount, string innerArgs, int samplesCount)
             : base(env, RegistrationName, trainedEnsemble, featureCount, innerArgs)
         {
             _quantileSampleCount = samplesCount;