Lockdown Microsoft.ML.Data Dataview folder. (#2608)

* Lockdown Microsoft.ML.Data Dataview folder.

* PR feedback.

Author: codemzs

src/Microsoft.ML.Data/DataView/AppendRowsDataView.cs

@@ -27,7 +27,8 @@ namespace Microsoft.ML.Data
     ///
     /// An AppendRowsDataView instance is shuffleable iff all of its sources are shuffleable and their row counts are known.
     /// </summary>
-    public sealed class AppendRowsDataView : IDataView
+    [BestFriend]
+    internal sealed class AppendRowsDataView : IDataView
     {
         public const string RegistrationName = "AppendRowsDataView";
 

src/Microsoft.ML.Data/DataView/CompositeRowToRowMapper.cs

@@ -11,7 +11,8 @@ namespace Microsoft.ML.Data
     /// <summary>
     /// A row-to-row mapper that is the result of a chained application of multiple mappers.
     /// </summary>
-    public sealed class CompositeRowToRowMapper : IRowToRowMapper
+    [BestFriend]
+    internal sealed class CompositeRowToRowMapper : IRowToRowMapper
     {
         [BestFriend]
         internal IRowToRowMapper[] InnerMappers { get; }

src/Microsoft.ML.Data/DataView/EmptyDataView.cs

@@ -12,7 +12,8 @@ namespace Microsoft.ML.Data
     /// <summary>
     /// This implements a data view that has a schema, but no rows.
     /// </summary>
-    public sealed class EmptyDataView : IDataView
+    [BestFriend]
+    internal sealed class EmptyDataView : IDataView
     {
         private readonly IHost _host;
 

src/Microsoft.ML.Data/DataView/LambdaFilter.cs

@@ -16,7 +16,7 @@ namespace Microsoft.ML.Data
     /// This applies the user provided RefPredicate to a column and drops rows that map to false. It automatically
     /// injects a standard conversion from the actual type of the source column to typeSrc (if needed).
     /// </summary>
-    public static class LambdaFilter
+    internal static class LambdaFilter
     {
         public static IDataView Create<TSrc>(IHostEnvironment env, string name, IDataView input,
             string src, DataViewType typeSrc, InPredicate<TSrc> predicate)

src/Microsoft.ML.Data/DataView/OpaqueDataView.cs

@@ -12,7 +12,8 @@ namespace Microsoft.ML.Data
     /// Opaque IDataView implementation to provide a barrier for data pipe optimizations.
     /// Used in cross validatation to generate the train/test pipelines for each fold.
     /// </summary>
-    public sealed class OpaqueDataView : IDataView
+    [BestFriend]
+    internal sealed class OpaqueDataView : IDataView
     {
         private readonly IDataView _source;
         public bool CanShuffle => _source.CanShuffle;

src/Microsoft.ML.Data/DataView/SimpleRow.cs

@@ -69,17 +69,4 @@ public override bool IsColumnActive(int col)
             return _getters[col] != null;
         }
     }
-
-    public static class SimpleSchemaUtils
-    {
-        public static DataViewSchema Create(IExceptionContext ectx, params KeyValuePair<string, DataViewType>[] columns)
-        {
-            Contracts.CheckValueOrNull(ectx);
-            ectx.CheckValue(columns, nameof(columns));
-
-            var builder = new SchemaBuilder();
-            builder.AddColumns(columns.Select(kvp => new DataViewSchema.DetachedColumn(kvp.Key, kvp.Value)));
-            return builder.GetSchema();
-        }
-    }
 }

src/Microsoft.ML.Data/DataView/TypedCursor.cs

@@ -32,7 +32,8 @@ internal interface IRowReadableAs<TRow> : IDisposable
     /// It can populate the user-supplied object's fields with the values of the current row.
     /// </summary>
     /// <typeparam name="TRow">The user-defined type that is being populated while cursoring.</typeparam>
-    public abstract class RowCursor<TRow> : DataViewRowCursor, IRowReadableAs<TRow>
+    [BestFriend]
+    internal abstract class RowCursor<TRow> : DataViewRowCursor, IRowReadableAs<TRow>
         where TRow : class
     {
         public abstract void FillValues(TRow row);
@@ -42,7 +43,8 @@ public abstract class RowCursor<TRow> : DataViewRowCursor, IRowReadableAs<TRow>
     /// This interface allows to create strongly typed cursors over a <see cref="IDataView"/>.
     /// </summary>
     /// <typeparam name="TRow">The user-defined type that is being populated while cursoring.</typeparam>
-    public interface ICursorable<TRow>
+    [BestFriend]
+    internal interface ICursorable<TRow>
         where TRow : class
     {
         /// <summary>

src/Microsoft.ML.Data/DataView/ZipDataView.cs

@@ -14,7 +14,8 @@ namespace Microsoft.ML.Data
     /// This is a data view that is a 'zip' of several data views.
     /// The length of the zipped data view is equal to the shortest of the lengths of the components.
     /// </summary>
-    public sealed class ZipDataView : IDataView
+    [BestFriend]
+    internal sealed class ZipDataView : IDataView
     {
         // REVIEW: there are other potential 'zip modes' that can be implemented:
         // * 'zip longest', iterate until all sources finish, and return the 'sensible missing values' for sources that ended

src/Microsoft.ML.Data/Evaluators/BinaryClassifierEvaluator.cs

@@ -1559,8 +1559,9 @@ private static IDataView ExtractWarnings(IHost host, Dictionary<string, IDataVie
             IDataView warnings;
             if (!metrics.TryGetValue(MetricKinds.Warnings, out warnings))
             {
-                warnings = new EmptyDataView(host, SimpleSchemaUtils.Create(host,
-                    new KeyValuePair<string, DataViewType>(MetricKinds.ColumnNames.WarningText, TextDataViewType.Instance)));
+                var schemaBuilder = new SchemaBuilder();
+                schemaBuilder.AddColumn(MetricKinds.ColumnNames.WarningText, TextDataViewType.Instance);
+                warnings = new EmptyDataView(host, schemaBuilder.GetSchema());
             }
 
             return warnings;
@@ -1571,11 +1572,11 @@ private static IDataView ExtractOverallMetrics(IHost host, Dictionary<string, ID
             IDataView overallMetrics;
             if (!metrics.TryGetValue(MetricKinds.OverallMetrics, out overallMetrics))
             {
-                overallMetrics = new EmptyDataView(host,
-                    SimpleSchemaUtils.Create(host,
-                        evaluator.GetOverallMetricColumns()
-                            .Select(mc => new KeyValuePair<string, DataViewType>(mc.LoadName, NumberDataViewType.Double))
-                            .ToArray()));
+                var schemaBuilder = new SchemaBuilder();
+                foreach (var mc in evaluator.GetOverallMetricColumns())
+                    schemaBuilder.AddColumn(mc.LoadName, NumberDataViewType.Double);
+
+                overallMetrics = new EmptyDataView(host, schemaBuilder.GetSchema());
             }
 
             return overallMetrics;
@@ -1586,8 +1587,9 @@ private static IDataView ExtractConfusionMatrix(IHost host, Dictionary<string, I
             IDataView confusionMatrix;
             if (!metrics.TryGetValue(MetricKinds.ConfusionMatrix, out confusionMatrix))
             {
-                confusionMatrix = new EmptyDataView(host,
-                    SimpleSchemaUtils.Create(host, new KeyValuePair<string, DataViewType>(MetricKinds.ColumnNames.Count, NumberDataViewType.Double)));
+                var schemaBuilder = new SchemaBuilder();
+                schemaBuilder.AddColumn(MetricKinds.ColumnNames.Count, NumberDataViewType.Double);
+                confusionMatrix = new EmptyDataView(host, schemaBuilder.GetSchema());
             }
 
             return confusionMatrix;

test/Microsoft.ML.StaticPipelineTesting/StaticPipeTests.cs

@@ -207,11 +207,11 @@ private static KeyValuePair<string, DataViewType> P(string name, DataViewType ty
         public void AssertStaticSimple()
         {
             var env = new MLContext(0);
-            var schema = SimpleSchemaUtils.Create(env,
-                P("hello", TextDataViewType.Instance),
-                P("my", new VectorType(NumberDataViewType.Int64, 5)),
-                P("friend", new KeyType(typeof(uint), 3)));
-            var view = new EmptyDataView(env, schema);
+            var schemaBuilder = new SchemaBuilder();
+            schemaBuilder.AddColumn("hello", TextDataViewType.Instance);
+            schemaBuilder.AddColumn("my", new VectorType(NumberDataViewType.Int64, 5));
+            schemaBuilder.AddColumn("friend", new KeyType(typeof(uint), 3));
+            var view = new EmptyDataView(env, schemaBuilder.GetSchema());
 
             view.AssertStatic(env, c => new
             {
@@ -231,11 +231,12 @@ public void AssertStaticSimple()
         public void AssertStaticSimpleFailure()
         {
             var env = new MLContext(0);
-            var schema = SimpleSchemaUtils.Create(env,
-                P("hello", TextDataViewType.Instance),
-                P("my", new VectorType(NumberDataViewType.Int64, 5)),
-                P("friend", new KeyType(typeof(uint), 3)));
-            var view = new EmptyDataView(env, schema);
+            var schemaBuilder = new SchemaBuilder();
+            schemaBuilder.AddColumn("hello", TextDataViewType.Instance);
+            schemaBuilder.AddColumn("my", new VectorType(NumberDataViewType.Int64, 5));
+            schemaBuilder.AddColumn("friend", new KeyType(typeof(uint), 3));
+
+            var view = new EmptyDataView(env, schemaBuilder.GetSchema());
 
             Assert.ThrowsAny<Exception>(() =>
                 view.AssertStatic(env, c => new