Internalize IDataLoader

src/Microsoft.ML.Data/Commands/DataCommand.cs

@@ -397,7 +397,8 @@ private IDataLoader LoadTransformChain(IDataLoader srcData)
         }
     }
 
-    public static class LoaderUtils
+    [BestFriend]
+    internal static class LoaderUtils
     {
         /// <summary>
         /// Saves <paramref name="loader"/> to the specified <paramref name="file"/>.

src/Microsoft.ML.Data/Data/IDataLoader.cs

@@ -10,9 +10,10 @@ namespace Microsoft.ML.Data
 {
     /// <summary>
     /// An interface for exposing some number of items that can be opened for reading.
+    /// </summary>
     /// REVIEW: Reconcile this with the functionality exposed by IHostEnvironment. For example,
     /// we could simply replace this with an array of IFileHandle.
-    /// </summary>
+
     public interface IMultiStreamSource
     {
         /// <summary>
@@ -32,26 +33,29 @@ public interface IMultiStreamSource
 
         /// <summary>
         /// Opens the indicated item and returns a text stream reader on it.
-        /// REVIEW: Consider making this an extension method.
         /// </summary>
+        /// REVIEW: Consider making this an extension method.
         TextReader OpenTextReader(int index);
     }
 
     /// <summary>
     /// Signature for creating an <see cref="IDataLoader"/>.
     /// </summary>
-    public delegate void SignatureDataLoader(IMultiStreamSource data);
+    [BestFriend]
+    internal delegate void SignatureDataLoader(IMultiStreamSource data);
 
     /// <summary>
     /// Signature for loading an <see cref="IDataLoader"/>.
     /// </summary>
-    public delegate void SignatureLoadDataLoader(ModelLoadContext ctx, IMultiStreamSource data);
+    [BestFriend]
+    internal delegate void SignatureLoadDataLoader(ModelLoadContext ctx, IMultiStreamSource data);
 
     /// <summary>
     /// Interface for a data loader. An <see cref="IDataLoader"/> can save its model information
     /// and is instantiatable from arguments and an <see cref="IMultiStreamSource"/> .
     /// </summary>
-    public interface IDataLoader : IDataView, ICanSaveModel
+    [BestFriend]
+    internal interface IDataLoader : IDataView, ICanSaveModel
     {
     }
 

src/Microsoft.ML.Data/DataLoadSave/Binary/BinaryLoader.cs

@@ -36,7 +36,8 @@
 
 namespace Microsoft.ML.Data.IO
 {
-    public sealed class BinaryLoader : IDataLoader, IDisposable
+    [BestFriend]
+    internal sealed class BinaryLoader : IDataLoader, IDisposable
     {
         public sealed class Arguments
         {

src/Microsoft.ML.Data/DataLoadSave/CompositeDataLoader.cs

@@ -29,7 +29,8 @@ namespace Microsoft.ML.Data
     /// The family of <c>Create</c> methods only instantiate <see cref="CompositeDataLoader"/>'s
     /// when there are transforms to keep, otherwise they just return underlying loaders.
     /// </summary>
-    public sealed class CompositeDataLoader : IDataLoader, ITransposeDataView
+    [BestFriend]
+    internal sealed class CompositeDataLoader : IDataLoader, ITransposeDataView
     {
         public sealed class Arguments
         {

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

@@ -12,6 +12,7 @@ namespace Microsoft.ML.Data
     /// A convenience class for concatenating several schemas together.
     /// This would be necessary when combining IDataViews through any type of combining operation, for example, zip.
     /// </summary>
+    [BestFriend]
     internal sealed class ZipBinding
     {
         private readonly Schema[] _sources;

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

@@ -362,7 +362,7 @@ internal bool TryUnparse(StringBuilder sb)
             }
         }
 
-        public sealed class Arguments
+        internal sealed class Arguments
         {
             [Argument(ArgumentType.Multiple | ArgumentType.Required, HelpText = "New column definition(s) (optional form: name:src)", ShortName = "col", SortOrder = 1)]
             public Column[] Column;

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

@@ -123,7 +123,8 @@ public abstract class ArgumentsBase : TransformInputBase
             public string DataFile;
 
             [Argument(ArgumentType.Multiple, HelpText = "Data loader", NullName = "<Auto>", SortOrder = 111, Visibility = ArgumentAttribute.VisibilityType.CmdLineOnly, SignatureType = typeof(SignatureDataLoader))]
-            public IComponentFactory<IMultiStreamSource, IDataLoader> Loader;
+            [BestFriend]
+            internal IComponentFactory<IMultiStreamSource, IDataLoader> Loader;
 
             [Argument(ArgumentType.AtMostOnce, HelpText = "Name of the text column containing the terms", ShortName = "termCol", SortOrder = 112, Visibility = ArgumentAttribute.VisibilityType.CmdLineOnly)]
             public string TermsColumn;

src/Microsoft.ML.Data/Utilities/ComponentCreation.cs

@@ -126,7 +126,8 @@ public static IDataView LoadTransforms(this IHostEnvironment env, Stream modelSt
         /// <summary>
         /// Creates a data loader from the arguments object.
         /// </summary>
-        public static IDataLoader CreateLoader<TArgs>(this IHostEnvironment env, TArgs arguments, IMultiStreamSource files)
+        [BestFriend]
+        internal static IDataLoader CreateLoader<TArgs>(this IHostEnvironment env, TArgs arguments, IMultiStreamSource files)
             where TArgs : class, new()
         {
             Contracts.CheckValue(env, nameof(env));
@@ -137,7 +138,8 @@ public static IDataLoader CreateLoader<TArgs>(this IHostEnvironment env, TArgs a
         /// <summary>
         /// Creates a data loader from the 'LoadName{settings}' string.
         /// </summary>
-        public static IDataLoader CreateLoader(this IHostEnvironment env, string settings, IMultiStreamSource files)
+        [BestFriend]
+        internal static IDataLoader CreateLoader(this IHostEnvironment env, string settings, IMultiStreamSource files)
         {
             Contracts.CheckValue(env, nameof(env));
             Contracts.CheckValue(files, nameof(files));

src/Microsoft.ML.Parquet/Microsoft.ML.Parquet.csproj

@@ -1,4 +1,4 @@
-<Project Sdk="Microsoft.NET.Sdk">
+<Project Sdk="Microsoft.NET.Sdk">
 
   <PropertyGroup>
     <TargetFramework>netstandard2.0</TargetFramework>

src/Microsoft.ML.Parquet/ParquetLoader.cs

@@ -32,7 +32,8 @@ namespace Microsoft.ML.Data
     /// <summary>
     /// Loads a parquet file into an IDataView. Supports basic mapping from Parquet input column data types to framework data types.
     /// </summary>
-    public sealed class ParquetLoader : IDataLoader, IDisposable
+    [BestFriend]
+    internal sealed class ParquetLoader : IDataLoader, IDisposable
     {
         /// <summary>
         /// A Column is a singular representation that consolidates all the related column chunks in the

src/Microsoft.ML.Data/DataLoadSave/PartitionedFileLoader.cs → src/Microsoft.ML.Parquet/PartitionedFileLoader.cs

@@ -47,7 +47,8 @@ namespace Microsoft.ML.Data
     ///             data1.parquet
     ///             data1.parquet
     /// </example>
-    public sealed class PartitionedFileLoader : IDataLoader
+    [BestFriend]
+    internal sealed class PartitionedFileLoader : IDataLoader
     {
         internal const string Summary = "Loads a horizontally partitioned file set.";
         internal const string UserName = "Partitioned Loader";

src/Microsoft.ML.Data/DataLoadSave/PartitionedPathParser.cs → src/Microsoft.ML.Parquet/PartitionedPathParser.cs

@@ -33,12 +33,14 @@ namespace Microsoft.ML.Data
     /// <summary>
     /// Delegate signature for a partitioned path parser.
     /// </summary>
-    public delegate void PartitionedPathParser();
+    [BestFriend]
+    internal delegate void PartitionedPathParser();
 
     /// <summary>
     /// Supports extracting column names and values from a path string.
     /// </summary>
-    public interface IPartitionedPathParser
+    [BestFriend]
+    internal interface IPartitionedPathParser
     {
         /// <summary>
         /// Extract the column definitions from a file path.
@@ -58,12 +60,13 @@ public interface IPartitionedPathParser
     }
 
     [TlcModule.ComponentKind("PartitionedPathParser")]
-    public interface IPartitionedPathParserFactory : IComponentFactory<IPartitionedPathParser>
+    [BestFriend]
+    internal interface IPartitionedPathParserFactory : IComponentFactory<IPartitionedPathParser>
     {
         new IPartitionedPathParser CreateComponent(IHostEnvironment env);
     }
 
-    public sealed class SimplePartitionedPathParser : IPartitionedPathParser, ICanSaveModel
+    internal sealed class SimplePartitionedPathParser : IPartitionedPathParser, ICanSaveModel
     {
         internal const string Summary = "A simple parser that extracts directory names as column values. Column names are defined as arguments.";
         internal const string UserName = "Simple Partitioned Path Parser";
@@ -193,12 +196,13 @@ public IEnumerable<string> ParseValues(string path)
 
     [TlcModule.Component(Name = ParquetPartitionedPathParser.LoadName, FriendlyName = ParquetPartitionedPathParser.UserName,
         Desc = ParquetPartitionedPathParser.Summary, Alias = ParquetPartitionedPathParser.ShortName)]
-    public class ParquetPartitionedPathParserFactory : IPartitionedPathParserFactory
+    internal class ParquetPartitionedPathParserFactory : IPartitionedPathParserFactory
     {
         public IPartitionedPathParser CreateComponent(IHostEnvironment env) => new ParquetPartitionedPathParser();
     }
 
-    public sealed class ParquetPartitionedPathParser : IPartitionedPathParser, ICanSaveModel
+    [BestFriend]
+    internal sealed class ParquetPartitionedPathParser : IPartitionedPathParser, ICanSaveModel
     {
         internal const string Summary = "Extract name/value pairs from Parquet formatted directory names. Example path: Year=2018/Month=12/data1.parquet";
         internal const string UserName = "Parquet Partitioned Path Parser";
@@ -276,8 +280,7 @@ public void Save(ModelSaveContext ctx)
                 ctx.SaveString(sb.ToString());
             };
         }
-
-        public IEnumerable<PartitionedFileLoader.Column> ParseColumns(string path)
+        IEnumerable<PartitionedFileLoader.Column> IPartitionedPathParser.ParseColumns(string path)
         {
             if (!TryParseNames(path, out List<string> names))
             {

src/Microsoft.ML.Parquet/Properties/AssemblyInfo.cs

@@ -0,0 +1,9 @@
+// Licensed to the .NET Foundation under one or more agreements.
+// 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.Runtime.CompilerServices;
+using Microsoft.ML;
+
+[assembly: InternalsVisibleTo(assemblyName: "Microsoft.ML.TestFramework" + PublicKey.TestValue)]
+[assembly: WantsToBeBestFriends]

src/Microsoft.ML.Transforms/Text/StopWordsRemovingTransformer.cs

@@ -628,7 +628,7 @@ public abstract class ArgumentsBase
             public string DataFile;
 
             [Argument(ArgumentType.Multiple, HelpText = "Data loader", NullName = "<Auto>", SortOrder = 3, Visibility = ArgumentAttribute.VisibilityType.CmdLineOnly, SignatureType = typeof(SignatureDataLoader))]
-            public IComponentFactory<IMultiStreamSource, IDataLoader> Loader;
+            internal IComponentFactory<IMultiStreamSource, IDataLoader> Loader;
 
             [Argument(ArgumentType.AtMostOnce, HelpText = "Name of the text column containing the stopwords", ShortName = "stopwordsCol", SortOrder = 4, Visibility = ArgumentAttribute.VisibilityType.CmdLineOnly)]
             public string StopwordsColumn;

src/Microsoft.ML.Transforms/Text/WordBagTransform.cs

@@ -424,7 +424,7 @@ public sealed class TermLoaderArguments
         public string DataFile;
 
         [Argument(ArgumentType.Multiple, HelpText = "Data loader", NullName = "<Auto>", SortOrder = 3, Visibility = ArgumentAttribute.VisibilityType.CmdLineOnly, SignatureType = typeof(SignatureDataLoader))]
-        public IComponentFactory<IMultiStreamSource, IDataLoader> Loader;
+        internal IComponentFactory<IMultiStreamSource, IDataLoader> Loader;
 
         [Argument(ArgumentType.AtMostOnce, HelpText = "Name of the text column containing the terms", ShortName = "termCol", SortOrder = 4, Visibility = ArgumentAttribute.VisibilityType.CmdLineOnly)]
         public string TermsColumn;

test/Microsoft.ML.TestFramework/DataPipe/TestDataPipeBase.cs

@@ -169,7 +169,7 @@ private void CheckSameSchemaShape(SchemaShape promised, SchemaShape delivered)
         /// * pathData defaults to breast-cancer.txt.
         /// * actLoader is invoked for extra validation (if non-null).
         /// </summary>
-        protected IDataLoader TestCore(string pathData, bool keepHidden, string[] argsPipe,
+        internal IDataLoader TestCore(string pathData, bool keepHidden, string[] argsPipe,
             Action<IDataLoader> actLoader = null, string suffix = "", string suffixBase = null, bool checkBaseline = true,
             bool forceDense = false, bool logCurs = false, bool roundTripText = true,
             bool checkTranspose = false, bool checkId = true, bool baselineSchema = true, int digitsOfPrecision = DigitsOfPrecision)
@@ -301,7 +301,7 @@ protected IDataLoader TestCore(string pathData, bool keepHidden, string[] argsPi
             return pipe1;
         }
 
-        protected IDataLoader CreatePipeDataLoader(IHostEnvironment env, string pathData, string[] argsPipe, out MultiFileSource files)
+        private IDataLoader CreatePipeDataLoader(IHostEnvironment env, string pathData, string[] argsPipe, out MultiFileSource files)
         {
             VerifyArgParsing(env, argsPipe);
 
@@ -318,30 +318,6 @@ protected IDataLoader CreatePipeDataLoader(IHostEnvironment env, string pathData
             return pipe;
         }
 
-        /// <summary>
-        /// Apply pipe's transforms and optionally ChooseColumns transform to newView, 
-        /// and test if pipe and newPipe have the same schema and values.
-        /// </summary>
-        protected void TestApplyTransformsToData(IHostEnvironment env, IDataLoader pipe, IDataView newView, string chooseArgs = null)
-        {
-            Contracts.AssertValue(pipe);
-            Contracts.AssertValue(newView);
-
-            IDataView view = pipe;
-            newView = ApplyTransformUtils.ApplyAllTransformsToData(env, view, newView);
-            if (!string.IsNullOrWhiteSpace(chooseArgs))
-            {
-                var component = new SubComponent<IDataTransform, SignatureDataTransform>("Choose", chooseArgs);
-                view = component.CreateInstance(env, view);
-                newView = component.CreateInstance(env, newView);
-            }
-
-            if (!CheckSameSchemas(view.Schema, newView.Schema))
-                Failed();
-            else if (!CheckSameValues(view, newView))
-                Failed();
-        }
-
         protected void VerifyArgParsing(IHostEnvironment env, string[] strs)
         {
             string str = CmdParser.CombineSettings(strs);
@@ -448,7 +424,7 @@ protected bool SaveLoadText(IDataView view, IHostEnvironment env,
             return true;
         }
 
-        protected string SavePipe(IDataLoader pipe, string suffix = "", string dir = "Pipeline")
+        protected private string SavePipe(IDataLoader pipe, string suffix = "", string dir = "Pipeline")
         {
             string name = TestName + suffix + ".zip";
             string pathModel = DeleteOutputPath("SavePipe", name);
@@ -471,7 +447,7 @@ protected string[] Concat(string[] args1, string[] args2)
             return res;
         }
 
-        protected IDataLoader LoadPipe(string pathModel, IHostEnvironment env, IMultiStreamSource files)
+        private IDataLoader LoadPipe(string pathModel, IHostEnvironment env, IMultiStreamSource files)
         {
             using (var file = Env.OpenInputFile(pathModel))
             using (var strm = file.OpenReadStream())

test/Microsoft.ML.TestFramework/Properties/AssemblyInfo.cs

@@ -6,3 +6,4 @@
 
 [assembly: InternalsVisibleTo(assemblyName: "Microsoft.ML.Predictor.Tests" + PublicKey.TestValue)]
 [assembly: InternalsVisibleTo(assemblyName: "Microsoft.ML.TimeSeries.Tests" + PublicKey.TestValue)]
+[assembly: InternalsVisibleTo(assemblyName: "Microsoft.ML.Tests" + PublicKey.TestValue)]

test/Microsoft.ML.Tests/Transformers/PcaTests.cs

@@ -17,16 +17,14 @@ namespace Microsoft.ML.Tests.Transformers
 {
     public sealed class PcaTests : TestDataPipeBase
     {
-        private readonly IHostEnvironment _env;
         private readonly string _dataSource;
         private readonly TextSaver _saver;
 
         public PcaTests(ITestOutputHelper helper)
             : base(helper)
         {
-            _env = new MLContext(seed: 1);
             _dataSource = GetDataPath("generated_regression_dataset.csv");
-            _saver = new TextSaver(_env, new TextSaver.Arguments { Silent = true, OutputHeader = false });
+            _saver = new TextSaver(ML, new TextSaver.Arguments { Silent = true, OutputHeader = false });
         }
 
         [Fact]