Fix CodeGen command and add a unit test

src/Microsoft.ML.Api/GenerateCodeCommand.cs

@@ -121,7 +121,6 @@ public void Run()
                     bool isScoreColumn = scorer.Schema.GetMetadataTypeOrNull(MetadataUtils.Kinds.ScoreColumnSetId, i) != null;
 
                     var sb = isScoreColumn ? scoreSb : nonScoreSb;
-
                     if (sb.Length > 0)
                         sb.AppendLine();
 

src/Microsoft.ML.Api/GeneratedCodeTemplate.csresource

@@ -1,6 +1,7 @@
 using System;
 using System.Collections.Generic;
-using Microsoft.ML;
+using Microsoft.ML.Legacy;
+using Microsoft.ML.Runtime.Api;
 
 namespace MLGeneratedCode
 {
@@ -31,6 +32,14 @@ public class ScoredOutput
 {
 /*#SCORE_CLASS_DECL#*/
 /*#/SCORE_CLASS_DECL#*/
+
+            // These are all remaining available columns, either supplied as the input, or intermediate
+            // columns generated by the transforms. Materializing these columns has a performance cost,
+            // so they are commented out. Feel free to uncomment any column that is useful for your scenario.
+#if false
+/*#SCORED_EXAMPLE_CLASS_DECL#*/
+/*#/SCORED_EXAMPLE_CLASS_DECL#*/
+#endif
 }
 
 /*public static void Main(string[] args)
@@ -45,7 +54,7 @@ PredictAsync(modelPath);
 /// This method demonstrates how to run prediction.
 ///
 ///</summary>
-public static void Predict(string modelPath)
+public static async void PredictAsync(string modelPath)
 {
     var model = await PredictionModel.ReadAsync<InputData, ScoredOutput>(modelPath);
 

test/BaselineOutput/Common/Command/codegen-out.cs

@@ -0,0 +1,88 @@
+using System;
+using System.Collections.Generic;
+using Microsoft.ML.Legacy;
+using Microsoft.ML.Runtime.Api;
+
+namespace MLGeneratedCode
+{
+public class Program
+{
+/// <summary>
+/// This is the input to the trained model.
+///
+/// In most pipelines, not all columns that are used in training are also used in scoring. Namely, the label
+/// and weight columns are almost never required at scoring time. Since we don't know which columns
+/// are 'optional' in this sense, all the columns are listed below.
+///
+/// You are free to remove any fields from the below class. If the fields are not required for scoring, the model
+/// will continue to work. Otherwise, the exception will be thrown when a prediction engine is created.
+///
+/// </summary>
+public class InputData
+{
+            public Single Label;
+
+            [VectorType(5)]
+            [ColumnName("F!1")]
+            public Single[] Column1 = new Single[5];
+
+            [VectorType(4)]
+            public Single[] F2 = new Single[4];
+}
+
+/// <summary>
+/// This is the output of the scored model, the prediction.
+///
+///</summary>
+public class ScoredOutput
+{
+            public Boolean PredictedLabel;
+
+            public Single Score;
+
+            public Single Probability;
+
+            // These are all remaining available columns, either supplied as the input, or intermediate
+            // columns generated by the transforms. Materializing these columns has a performance cost,
+            // so they are commented out. Feel free to uncomment any column that is useful for your scenario.
+#if false
+            public Single Label;
+
+            [VectorType(5)]
+            [ColumnName("F!1")]
+            public Single[] Column1;
+
+            [VectorType(4)]
+            public Single[] F2;
+
+            [VectorType(9)]
+            public Single[] Features;
+#endif
+}
+
+/*public static void Main(string[] args)
+{
+string modelPath;
+modelPath = "model.zip";
+PredictAsync(modelPath);
+}*/
+
+/// <summary>
+/// This method demonstrates how to run prediction.
+///
+///</summary>
+public static async void PredictAsync(string modelPath)
+{
+    var model = await PredictionModel.ReadAsync<InputData, ScoredOutput>(modelPath);
+
+    var inputData = new InputData();
+    // TODO: populate the example's features.
+
+    var score = model.Predict(inputData);
+    // TODO: consume the resulting score.
+
+    var scores = model.Predict(new List<InputData> { inputData, inputData });
+    // TODO: consume the resulting scores.
+  }
+} 
+}

test/Microsoft.ML.Predictor.Tests/TestParallelFasttreeInterface.cs

@@ -173,7 +173,7 @@ public TestParallelFasttreeInterface(ITestOutputHelper helper)
         {
         }
 
-        [Fact]
+        [Fact(Skip = "'checker' is not a valid value for the 'parag' argument in FastTree")]
         [TestCategory("ParallelFasttree")]
         public void CheckFastTreeParallelInterface()
         {
@@ -185,7 +185,8 @@ public void CheckFastTreeParallelInterface()
             var trainArgs = string.Format(
                 "train data={{{0}}} loader=Text{{col=Label:0 col=F!1:1-5 col=F2:6-9}} xf=Concat{{col=Features:F!1,F2}}  tr=FastTreeBinaryClassification{{lr=0.1 nl=12 mil=10 iter=1 parag=checker}} out={{{1}}}",
                 dataPath, modelOutPath);
-            MainForTest(trainArgs);
+            var res = MainForTest(trainArgs);
+            Assert.Equal(0, res);
         }
     }
 }

test/Microsoft.ML.TestFramework/BaseTestBaseline.cs

@@ -2,6 +2,7 @@
 // The .NET Foundation licenses this file to you under the MIT license.
 // See the LICENSE file in the project root for more information.
 
+using Microsoft.ML.Runtime.Api;
 using Microsoft.ML.Runtime.Data;
 using Microsoft.ML.Runtime.Internal.Utilities;
 using Microsoft.ML.Runtime.Tools;
@@ -100,6 +101,7 @@ protected override void Initialize()
             _env = new ConsoleEnvironment(42, outWriter: LogWriter, errWriter: LogWriter)
                 .AddStandardComponents();
             ML = new MLContext(42);
+            ML.AddStandardComponents();
         }
 
         // This method is used by subclass to dispose of disposable objects
@@ -815,10 +817,9 @@ protected static StreamReader OpenReader(string path)
         /// This method is used in unit tests when the output is not baselined.
         /// If the output is to be baselined and compared, the other overload should be used.
         /// </summary>
-        protected static int MainForTest(string args)
+        protected int MainForTest(string args)
         {
-            var env = new MLContext();
-            return Maml.MainCore(env, args, false);
+            return Maml.MainCore(ML, args, false);
         }
     }
 

test/Microsoft.ML.TestFramework/EnvironmentExtensions.cs

@@ -3,6 +3,7 @@
 // See the LICENSE file in the project root for more information.
 
 using Microsoft.ML.Runtime;
+using Microsoft.ML.Runtime.Api;
 using Microsoft.ML.Runtime.Data;
 using Microsoft.ML.Runtime.Ensemble;
 using Microsoft.ML.Runtime.Learners;
@@ -28,6 +29,7 @@ public static TEnvironment AddStandardComponents<TEnvironment>(this TEnvironment
 #pragma warning disable 612
             env.ComponentCatalog.RegisterAssembly(typeof(Experiment).Assembly); // ML.Legacy
 #pragma warning restore 612
+            env.ComponentCatalog.RegisterAssembly(typeof(ComponentCreation).Assembly); // ML.Api
             return env;
         }
     }

test/Microsoft.ML.TestFramework/TestCommandBase.cs

@@ -672,16 +672,15 @@ public void CommandShowSchema()
         }
 
         [TestCategory(Cat)]
-        [Fact(Skip = "Need CoreTLC specific baseline update")]
+        [Fact]
         public void CommandShowSchemaModel()
         {
-            string trainDataPath = GetDataPath(@"..\UCI", "adult.test.tiny");
+            string trainDataPath = GetDataPath("adult.tiny.with-schema.txt");
             string modelPath = ModelPath().Path;
             string args =
                 string.Format(
                     @"train data={{{0}}}
                      loader=Text{{
-                        sep=, 
                         header=+ 
                         col=NumFeatures:Num:9-14 
                         col=CatFeaturesText:TX:0~* 
@@ -922,29 +921,27 @@ public void CommandCrossValidationAndSave()
         // multiple different FastTree (Ranking and Classification for example) instances in different threads.
         // FastTree internally fails if we try to run it simultaneously and if this happens we wouldn't get model file for training.
         [TestCategory(Cat)]
-        [Fact(Skip = "Need CoreTLC specific baseline update")]
+        [Fact]
         public void CommandTrainFastTreeInDifferentThreads()
         {
-            var dataPath = GetDataPath("vw.dat");
-            var firstModelOutPath = CreateOutputPath("TreeTransform-model2.zip");
-            var secondModelOutPath = CreateOutputPath("TreeTransform-model1.zip");
-            var trainArgs = "Train tr=SDCA loader=TextLoader{sep=space col=Label:R4:0 col=Features:R4:1 col=Name:TX:2,5-17 col=Cat:TX:3 col=Cat01:TX:4}" + "xf=CategoricalTransform{col=Cat col=Cat01} xf=Concat{col=Features:Features,Cat,Cat01} xf=TreeFeat{tr=FastTreeBinaryClassification} xf=TreeFeat" + "{tr=FastTreeRanking} xf=Concat{col=Features:Features,Leaves,Paths,Trees}";
+            var dataPath = GetDataPath(TestDatasets.adult.testFilename);
+            var firstModelOutPath = DeleteOutputPath("TreeTransform-model2.zip");
+            var secondModelOutPath = DeleteOutputPath("TreeTransform-model1.zip");
+            var trainArgs = $"Train tr=SDCA {TestDatasets.adult.loaderSettings} {TestDatasets.adult.mamlExtraSettings[0]} {TestDatasets.adult.mamlExtraSettings[1]}" +
+                " xf=TreeFeat{tr=FastTreeBinaryClassification} xf=TreeFeat{tr=FastTreeRanking} xf=Concat{col=Features:Features,Leaves,Paths,Trees}";
 
-            var firsttrainArgs = string.Format("{0} data={1} out={2}", trainArgs, dataPath, firstModelOutPath.Path);
-            var secondTrainArgs = string.Format("{0} data={1} out={2}", trainArgs, dataPath, secondModelOutPath.Path);
+            var firsttrainArgs = $"{trainArgs} data={dataPath} out={firstModelOutPath}";
+            var secondTrainArgs = $"{trainArgs} data={dataPath} out={secondModelOutPath}";
 
-            var t = new Task[2];
-            t[0] = new Task(() => { MainForTest(firsttrainArgs); });
-            t[1] = new Task(() => { MainForTest(secondTrainArgs); });
+            var t = new Task<int>[2];
+            t[0] = new Task<int>(() => MainForTest(firsttrainArgs));
+            t[1] = new Task<int>(() => MainForTest(secondTrainArgs));
             t[0].Start();
             t[1].Start();
             Task.WaitAll(t);
 
-            if (!File.Exists(firstModelOutPath.Path))
-                Fail("First model doesn't exist");
-            if (!File.Exists(secondModelOutPath.Path))
-                Fail("Second model doesn't exist");
-            Done();
+            Assert.Equal(0, t[0].Result);
+            Assert.Equal(0, t[1].Result);
         }
 
         [TestCategory(Cat), TestCategory("FastTree")]
@@ -2096,5 +2093,32 @@ public void Datatypes()
             TestCore("savedata", intermediateData.Path, "loader=binary", "saver=text", textOutputPath.Arg("dout"));
             Done();
         }
+
+        [Fact]
+        public void CommandCodeGen()
+        {
+            if (!RuntimeInformation.IsOSPlatform(OSPlatform.Windows))
+                return;
+
+            // REVIEW: this tests that the generated output matches the baseline. This does NOT baseline
+            // the console output. Currently, there's no console output either, but if some is added, a baseline test
+            // will be in order.
+
+            // First, train a model on breast-cancer.
+            var dataPath = GetDataPath("breast-cancer.txt");
+            var modelOutPath = DeleteOutputPath("Command", "codegen-model.zip");
+            var csOutPath = DeleteOutputPath("Command", "codegen-out.cs");
+
+            var trainArgs = string.Format(
+                "train data={{{0}}} loader=Text{{col=Label:0 col=F!1:1-5 col=F2:6-9}} xf=Concat{{col=Features:F!1,F2}} tr=lr out={{{1}}}",
+                dataPath, modelOutPath);
+            MainForTest(trainArgs);
+
+            // Now, generate the prediction code.
+            MainForTest(string.Format("codegen in={{{0}}} cs={{{1}}} modelNameOverride=model.zip", modelOutPath, csOutPath));
+            CheckEquality("Command", "codegen-out.cs");
+
+            Done();
+        }
     }
 }