Role mapped improvements (#496)

* RoleMappedSchema/Data change to use constructors
* Nuke all create methods, pointless "no-roles" constructor.
* Nuke TrainUtils.CreateExamples/CreateExamplesOpt
* Opportunistically improve code quality and reporting of Kmeans++

src/Microsoft.ML.Api/ComponentCreation.cs

@@ -52,7 +52,7 @@ public static RoleMappedData CreateExamples(this IHostEnvironment env, IDataView
             env.CheckValueOrNull(weight);
             env.CheckValueOrNull(custom);
 
-            return TrainUtils.CreateExamples(data, label, features, group, weight, name: null, custom: custom);
+            return new RoleMappedData(data, label, features, group, weight, name: null, custom: custom);
         }
 
         /// <summary>

src/Microsoft.ML.Api/GenerateCodeCommand.cs

@@ -108,8 +108,8 @@ public void Run()
                 {
                     var roles = ModelFileUtils.LoadRoleMappingsOrNull(_host, fs);
                     scorer = roles != null
-                        ? _host.CreateDefaultScorer(RoleMappedData.CreateOpt(transformPipe, roles), pred)
-                        : _host.CreateDefaultScorer(_host.CreateExamples(transformPipe, "Features"), pred);
+                        ? _host.CreateDefaultScorer(new RoleMappedData(transformPipe, roles, opt: true), pred)
+                        : _host.CreateDefaultScorer(new RoleMappedData(transformPipe, label: null, "Features"), pred);
                 }
 
                 var nonScoreSb = new StringBuilder();

src/Microsoft.ML.Api/PredictionEngine.cs

@@ -49,8 +49,8 @@ internal BatchPredictionEngine(IHostEnvironment env, Stream modelStream, bool ig
             {
                 var roles = ModelFileUtils.LoadRoleMappingsOrNull(env, modelStream);
                 pipe = roles != null
-                    ? env.CreateDefaultScorer(RoleMappedData.CreateOpt(pipe, roles), predictor)
-                    : env.CreateDefaultScorer(env.CreateExamples(pipe, "Features"), predictor);
+                    ? env.CreateDefaultScorer(new RoleMappedData(pipe, roles, opt: true), predictor)
+                    : env.CreateDefaultScorer(new RoleMappedData(pipe, label: null, "Features"), predictor);
             }
 
             _pipeEngine = new PipeEngine<TDst>(env, pipe, ignoreMissingColumns, outputSchemaDefinition);

src/Microsoft.ML.Core/Data/MetadataUtils.cs

@@ -312,7 +312,6 @@ public static bool HasSlotNames(this ISchema schema, int col, int vectorSize)
         public static void GetSlotNames(RoleMappedSchema schema, RoleMappedSchema.ColumnRole role, int vectorSize, ref VBuffer<DvText> slotNames)
         {
             Contracts.CheckValueOrNull(schema);
-            Contracts.CheckValue(role.Value, nameof(role));
             Contracts.CheckParam(vectorSize >= 0, nameof(vectorSize));
 
             IReadOnlyList<ColumnInfo> list;

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

@@ -254,7 +254,7 @@ private RoleMappedData ApplyAllTransformsToData(IHostEnvironment env, IChannel c
             RoleMappedData srcData, IDataView marker)
         {
             var pipe = ApplyTransformUtils.ApplyAllTransformsToData(env, srcData.Data, dstData, marker);
-            return RoleMappedData.Create(pipe, srcData.Schema.GetColumnRoleNames());
+            return new RoleMappedData(pipe, srcData.Schema.GetColumnRoleNames());
         }
 
         /// <summary>
@@ -277,7 +277,7 @@ private RoleMappedData CreateRoleMappedData(IHostEnvironment env, IChannel ch, I
             // Training pipe and examples.
             var customCols = TrainUtils.CheckAndGenerateCustomColumns(ch, Args.CustomColumn);
 
-            return TrainUtils.CreateExamples(data, label, features, group, weight, name, customCols);
+            return new RoleMappedData(data, label, features, group, weight, name, customCols);
         }
 
         private string GetSplitColumn(IChannel ch, IDataView input, ref IDataView output)
@@ -568,7 +568,7 @@ private FoldResult RunFold(int fold)
                     {
                         using (var file = host.CreateOutputFile(modelFileName))
                         {
-                            var rmd = RoleMappedData.Create(
+                            var rmd = new RoleMappedData(
                                 CompositeDataLoader.ApplyTransform(host, _loader, null, null,
                                 (e, newSource) => ApplyTransformUtils.ApplyAllTransformsToData(e, trainData.Data, newSource)),
                                 trainData.Schema.GetColumnRoleNames());
@@ -581,17 +581,17 @@ private FoldResult RunFold(int fold)
                     if (!evalComp.IsGood())
                         evalComp = EvaluateUtils.GetEvaluatorType(ch, scorePipe.Schema);
                     var eval = evalComp.CreateInstance(host);
-                    // Note that this doesn't require the provided columns to exist (because of "Opt").
+                    // Note that this doesn't require the provided columns to exist (because of the "opt" parameter).
                     // We don't normally expect the scorer to drop columns, but if it does, we should not require
                     // all the columns in the test pipeline to still be present.
-                    var dataEval = RoleMappedData.CreateOpt(scorePipe, testData.Schema.GetColumnRoleNames());
+                    var dataEval = new RoleMappedData(scorePipe, testData.Schema.GetColumnRoleNames(), opt: true);
 
                     var dict = eval.Evaluate(dataEval);
                     RoleMappedData perInstance = null;
                     if (_savePerInstance)
                     {
                         var perInst = eval.GetPerInstanceMetrics(dataEval);
-                        perInstance = RoleMappedData.CreateOpt(perInst, dataEval.Schema.GetColumnRoleNames());
+                        perInstance = new RoleMappedData(perInst, dataEval.Schema.GetColumnRoleNames(), opt: true);
                     }
                     ch.Done();
                     return new FoldResult(dict, dataEval.Schema.Schema, perInstance, trainData.Schema);

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

@@ -305,7 +305,7 @@ protected void LoadModelObjects(
                             // can be loaded with no data at all, to get their schemas.
                             if (trainPipe == null)
                                 trainPipe = ModelFileUtils.LoadLoader(Host, rep, new MultiFileSource(null), loadTransforms: true);
-                            trainSchema = RoleMappedSchema.Create(trainPipe.Schema, trainRoleMappings);
+                            trainSchema = new RoleMappedSchema(trainPipe.Schema, trainRoleMappings);
                         }
                         // If the role mappings are null, an alternative would be to fail. However the idea
                         // is that the scorer should always still succeed, although perhaps with reduced

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

@@ -158,7 +158,7 @@ public static IDataTransform Create(IHostEnvironment env, Arguments args, IDataV
                     evalComp = EvaluateUtils.GetEvaluatorType(ch, input.Schema);
 
                 var eval = evalComp.CreateInstance(env);
-                var data = TrainUtils.CreateExamples(input, label, null, group, weight, null, customCols);
+                var data = new RoleMappedData(input, label, null, group, weight, null, customCols);
                 return eval.GetPerInstanceMetrics(data);
             }
         }
@@ -236,7 +236,7 @@ private void RunCore(IChannel ch)
             if (!evalComp.IsGood())
                 evalComp = EvaluateUtils.GetEvaluatorType(ch, view.Schema);
             var evaluator = evalComp.CreateInstance(Host);
-            var data = TrainUtils.CreateExamples(view, label, null, group, weight, name, customCols);
+            var data = new RoleMappedData(view, label, null, group, weight, name, customCols);
             var metrics = evaluator.Evaluate(data);
             MetricWriter.PrintWarnings(ch, metrics);
             evaluator.PrintFoldResults(ch, metrics);
@@ -248,7 +248,7 @@ private void RunCore(IChannel ch)
             if (!string.IsNullOrWhiteSpace(Args.OutputDataFile))
             {
                 var perInst = evaluator.GetPerInstanceMetrics(data);
-                var perInstData = TrainUtils.CreateExamples(perInst, label, null, group, weight, name, customCols);
+                var perInstData = new RoleMappedData(perInst, label, null, group, weight, name, customCols);
                 var idv = evaluator.GetPerInstanceDataViewToSave(perInstData);
                 MetricWriter.SavePerInstance(Host, ch, Args.OutputDataFile, idv);
             }

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

@@ -219,7 +219,7 @@ public static void LoadModel(IHostEnvironment env, Stream modelStream, bool load
                     if (roles != null)
                     {
                         var emptyView = ModelFileUtils.LoadPipeline(env, rep, new MultiFileSource(null));
-                        schema = RoleMappedSchema.CreateOpt(emptyView.Schema, roles);
+                        schema = new RoleMappedSchema(emptyView.Schema, roles, opt: true);
                     }
                     else
                     {

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

@@ -97,10 +97,7 @@ private void RunCore(IChannel ch)
 
             ch.Trace("Creating loader");
 
-            IPredictor predictor;
-            IDataLoader loader;
-            RoleMappedSchema trainSchema;
-            LoadModelObjects(ch, true, out predictor, true, out trainSchema, out loader);
+            LoadModelObjects(ch, true, out var predictor, true, out var trainSchema, out var loader);
             ch.AssertValue(predictor);
             ch.AssertValueOrNull(trainSchema);
             ch.AssertValue(loader);
@@ -116,7 +113,7 @@ private void RunCore(IChannel ch)
             string group = TrainUtils.MatchNameOrDefaultOrNull(ch, loader.Schema,
                 nameof(Args.GroupColumn), Args.GroupColumn, DefaultColumnNames.GroupId);
             var customCols = TrainUtils.CheckAndGenerateCustomColumns(ch, Args.CustomColumn);
-            var schema = TrainUtils.CreateRoleMappedSchemaOpt(loader.Schema, feat, group, customCols);
+            var schema = new RoleMappedSchema(loader.Schema, label: null, feature: feat, group: group, custom: customCols, opt: true);
             var mapper = bindable.Bind(Host, schema);
 
             if (!scorer.IsGood())
@@ -153,22 +150,20 @@ private void RunCore(IChannel ch)
                 Args.OutputAllColumns == true || Utils.Size(Args.OutputColumn) == 0;
 
             if (Args.OutputAllColumns == true && Utils.Size(Args.OutputColumn) != 0)
-                ch.Warning("outputAllColumns=+ always writes all columns irrespective of outputColumn specified.");
+                ch.Warning(nameof(Args.OutputAllColumns) + "=+ always writes all columns irrespective of " + nameof(Args.OutputColumn) + " specified.");
 
             if (!outputAllColumns && Utils.Size(Args.OutputColumn) != 0)
             {
                 foreach (var outCol in Args.OutputColumn)
                 {
-                    int dummyColIndex;
-                    if (!loader.Schema.TryGetColumnIndex(outCol, out dummyColIndex))
+                    if (!loader.Schema.TryGetColumnIndex(outCol, out int dummyColIndex))
                         throw ch.ExceptUserArg(nameof(Arguments.OutputColumn), "Column '{0}' not found.", outCol);
                 }
             }
 
-            int colMax;
             uint maxScoreId = 0;
             if (!outputAllColumns)
-                maxScoreId = loader.Schema.GetMaxMetadataKind(out colMax, MetadataUtils.Kinds.ScoreColumnSetId);
+                maxScoreId = loader.Schema.GetMaxMetadataKind(out int colMax, MetadataUtils.Kinds.ScoreColumnSetId);
             ch.Assert(outputAllColumns || maxScoreId > 0); // score set IDs are one-based
             var cols = new List<int>();
             for (int i = 0; i < loader.Schema.ColumnCount; i++)
@@ -211,12 +206,12 @@ private bool ShouldAddColumn(ISchema schema, int i, uint scoreSet, bool outputNa
             {
                 switch (schema.GetColumnName(i))
                 {
-                case "Label":
-                case "Name":
-                case "Names":
-                    return true;
-                default:
-                    break;
+                    case "Label":
+                    case "Name":
+                    case "Names":
+                        return true;
+                    default:
+                        break;
                 }
             }
             if (Args.OutputColumn != null && Array.FindIndex(Args.OutputColumn, schema.GetColumnName(i).Equals) >= 0)
@@ -229,8 +224,7 @@ public static class ScoreUtils
     {
         public static IDataScorerTransform GetScorer(IPredictor predictor, RoleMappedData data, IHostEnvironment env, RoleMappedSchema trainSchema)
         {
-            ISchemaBoundMapper mapper;
-            var sc = GetScorerComponentAndMapper(predictor, null, data.Schema, env, out mapper);
+            var sc = GetScorerComponentAndMapper(predictor, null, data.Schema, env, out var mapper);
             return sc.CreateInstance(env, data.Data, mapper, trainSchema);
         }
 
@@ -247,9 +241,8 @@ public static IDataScorerTransform GetScorer(SubComponent<IDataScorerTransform,
             env.CheckValueOrNull(customColumns);
             env.CheckValueOrNull(trainSchema);
 
-            var schema = TrainUtils.CreateRoleMappedSchemaOpt(input.Schema, featureColName, groupColName, customColumns);
-            ISchemaBoundMapper mapper;
-            var sc = GetScorerComponentAndMapper(predictor, scorer, schema, env, out mapper);
+            var schema = new RoleMappedSchema(input.Schema, label: null, feature: featureColName, group: groupColName, custom: customColumns, opt: true);
+            var sc = GetScorerComponentAndMapper(predictor, scorer, schema, env, out var mapper);
             return sc.CreateInstance(env, input, mapper, trainSchema);
         }
 
@@ -280,7 +273,7 @@ public static SubComponent<IDataScorerTransform, SignatureDataScorer> GetScorerC
             Contracts.AssertValue(mapper);
 
             string loadName = null;
-            DvText scoreKind = default(DvText);
+            DvText scoreKind = default;
             if (mapper.OutputSchema.ColumnCount > 0 &&
                 mapper.OutputSchema.TryGetMetadata(TextType.Instance, MetadataUtils.Kinds.ScoreColumnKind, 0, ref scoreKind) &&
                 scoreKind.HasChars)
@@ -311,10 +304,8 @@ public static ISchemaBindableMapper GetSchemaBindableMapper(IHostEnvironment env
             env.CheckValue(predictor, nameof(predictor));
             env.CheckValueOrNull(scorerSettings);
 
-            ISchemaBindableMapper bindable;
-
             // See if we can instantiate a mapper using scorer arguments.
-            if (scorerSettings.IsGood() && TryCreateBindableFromScorer(env, predictor, scorerSettings, out bindable))
+            if (scorerSettings.IsGood() && TryCreateBindableFromScorer(env, predictor, scorerSettings, out var bindable))
                 return bindable;
 
             // The easy case is that the predictor implements the interface.

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

@@ -114,7 +114,7 @@ private void RunCore(IChannel ch)
             if (!evalComp.IsGood())
                 evalComp = EvaluateUtils.GetEvaluatorType(ch, scorePipe.Schema);
             var evaluator = evalComp.CreateInstance(Host);
-            var data = TrainUtils.CreateExamples(scorePipe, label, null, group, weight, name, customCols);
+            var data = new RoleMappedData(scorePipe, label, null, group, weight, name, customCols);
             var metrics = evaluator.Evaluate(data);
             MetricWriter.PrintWarnings(ch, metrics);
             evaluator.PrintFoldResults(ch, metrics);
@@ -128,7 +128,7 @@ private void RunCore(IChannel ch)
             if (!string.IsNullOrWhiteSpace(Args.OutputDataFile))
             {
                 var perInst = evaluator.GetPerInstanceMetrics(data);
-                var perInstData = TrainUtils.CreateExamples(perInst, label, null, group, weight, name, customCols);
+                var perInstData = new RoleMappedData(perInst, label, null, group, weight, name, customCols);
                 var idv = evaluator.GetPerInstanceDataViewToSave(perInstData);
                 MetricWriter.SavePerInstance(Host, ch, Args.OutputDataFile, idv);
             }