implement auto featurizer

src/Microsoft.ML.AutoML/API/AutoCatalog.cs

@@ -4,8 +4,11 @@
 
 using System;
 using System.Collections.Generic;
+using System.Diagnostics.Contracts;
+using System.Linq;
 using Microsoft.ML.AutoML.CodeGen;
 using Microsoft.ML.Data;
+using Microsoft.ML.Runtime;
 using Microsoft.ML.SearchSpace;
 using Microsoft.ML.Trainers.FastTree;
 
@@ -538,55 +541,164 @@ public SweepableEstimator[] Regression(string labelColumnName = DefaultColumnNam
  /// <param name="inputColumnName">input column name.</param>
  internal SweepableEstimator[] TextFeaturizer(string outputColumnName, string inputColumnName)
  {
- throw new NotImplementedException();
+ var option = new FeaturizeTextOption
+ {
+ InputColumnName = inputColumnName,
+ OutputColumnName = outputColumnName,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateFeaturizeText(option) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing numeric columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] NumericFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] NumericFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(inputColumnNames, nameof(inputColumnNames));
+ Contracts.CheckValue(outputColumnNames, nameof(outputColumnNames));
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+ var replaceMissingValueOption = new ReplaceMissingValueOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateReplaceMissingValues(replaceMissingValueOption) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing catalog columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] CatalogFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] CatalogFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+
+ var option = new OneHotOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new SweepableEstimator[] { SweepableEstimatorFactory.CreateOneHotEncoding(option), SweepableEstimatorFactory.CreateOneHotHashEncoding(option) };
  }
 
  /// <summary>
  /// Create a single featurize pipeline according to <paramref name="data"/>. This function will collect all columns in <paramref name="data"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
  /// <param name="data">input data.</param>
  /// <param name="catalogColumns">columns that should be treated as catalog. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="numericColumns">columns that should be treated as numeric. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="textColumns">columns that should be treated as text. If not specified, it will automatically infer if a column is catalog or not.</param>
  /// <param name="outputColumnName">output feature column.</param>
  /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
-internal MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] excludeColumns = null)
+public MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] numericColumns = null, string[] textColumns = null, string[] excludeColumns = null)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+
+ // validate if there's overlapping among catalogColumns, numericColumns, textColumns and excludeColumns
+ var overallColumns = new string[][] { catalogColumns, numericColumns, textColumns, excludeColumns }
+ .Where(c => c != null)
+ .SelectMany(c => c);
+
+ if (overallColumns != null)
+ {
+ Contracts.Assert(overallColumns.Count() == overallColumns.Distinct().Count(), "detect overlapping among catalogColumns, numericColumns, textColumns and excludedColumns");
+ }
+
+ var columnInfo = new ColumnInformation();
+
+ if (excludeColumns != null)
+ {
+ foreach (var ignoreColumn in excludeColumns)
+ {
+ columnInfo.IgnoredColumnNames.Add(ignoreColumn);
+ }
+ }
+
+ if (catalogColumns != null)
+ {
+ foreach (var catalogColumn in catalogColumns)
+ {
+ columnInfo.CategoricalColumnNames.Add(catalogColumn);
+ }
+ }
+
+ if (numericColumns != null)
+ {
+ foreach (var column in numericColumns)
+ {
+ columnInfo.NumericColumnNames.Add(column);
+ }
+ }
+
+ if (textColumns != null)
+ {
+ foreach (var column in textColumns)
+ {
+ columnInfo.TextColumnNames.Add(column);
+ }
+ }
+
+ return this.Featurizer(data, columnInfo, outputColumnName);
  }
 
  /// <summary>
- /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/>,
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
+ /// <param name="data">input data.</param>
  /// <param name="columnInformation">column information.</param>
  /// <param name="outputColumnName">output feature column.</param>
- /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
- /// <returns></returns>
- internal MultiModelPipeline Featurizer(ColumnInformation columnInformation, string outputColumnName = "Features", string[] excludeColumns = null)
+ /// <returns>A <see cref="MultiModelPipeline"/> for featurization.</returns>
+ public MultiModelPipeline Featurizer(IDataView data, ColumnInformation columnInformation, string outputColumnName = "Features")
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+ Contracts.CheckValue(columnInformation, nameof(columnInformation));
+
+ var columnPurposes = PurposeInference.InferPurposes(this._context, data, columnInformation);
+ var textFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.TextFeature);
+ var numericFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.NumericFeature);
+ var catalogFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.CategoricalFeature);
+ var textFeatureColumnNames = textFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var numericFeatureColumnNames = numericFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var catalogFeatureColumnNames = catalogFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+
+ var pipeline = new MultiModelPipeline();
+ if (numericFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.NumericFeaturizer(numericFeatureColumnNames, numericFeatureColumnNames));
+ }
+
+ if (catalogFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.CatalogFeaturizer(catalogFeatureColumnNames, catalogFeatureColumnNames));
+ }
+
+ foreach (var textColumn in textFeatureColumnNames)
+ {
+ pipeline = pipeline.Append(this.TextFeaturizer(textColumn, textColumn));
+ }
+
+ var option = new ConcatOption
+ {
+ InputColumnNames = textFeatureColumnNames.Concat(numericFeatureColumnNames).Concat(catalogFeatureColumnNames).ToArray(),
+ OutputColumnName = outputColumnName,
+ };
+
+ if (option.InputColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(SweepableEstimatorFactory.CreateConcatenate(option));
+ }
+
+ return pipeline;
  }
  }
 }

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

@@ -182,7 +182,7 @@ internal HashingTransformer(IHostEnvironment env, params HashingEstimator.Column
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
 
  if (column.Combine && column.UseOrderedHashing)
  throw Host.ExceptParam(nameof(HashingEstimator.ColumnOptions.Combine), "When the 'Combine' option is specified, ordered hashing is not supported.");

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

@@ -184,7 +184,7 @@ internal NgramHashingTransformer(IHostEnvironment env, params NgramHashingEstima
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
  }
  }
 

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_iris_test.approved.txt

@@ -0,0 +1,34 @@
+{
+ "schema": "e0 * e1",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "OutputColumnName": "Features"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_uci_adult_test.approved.txt

@@ -0,0 +1,83 @@
+{
+ "schema": "e0 * (e1 \u002B e2) * e3",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "Features"
+ ],
+ "InputColumnNames": [
+ "Features"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "OneHotEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e2": {
+ "estimatorType": "OneHotHashEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e3": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "Features",
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "OutputColumnName": "OutputFeature"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/AutoFeaturizerTests.cs

@@ -0,0 +1,65 @@
+// 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;
+using System.Collections.Generic;
+using System.Text;
+using System.Text.Json;
+using Microsoft.ML.TestFramework;
+using Xunit;
+using Xunit.Abstractions;
+using ApprovalTests;
+using ApprovalTests.Namers;
+using ApprovalTests.Reporters;
+using System.Text.Json.Serialization;
+
+namespace Microsoft.ML.AutoML.Test
+{
+ public class AutoFeaturizerTests : BaseTestClass
+ {
+ private readonly JsonSerializerOptions _jsonSerializerOptions;
+
+ public AutoFeaturizerTests(ITestOutputHelper output)
+ : base(output)
+ {
+ _jsonSerializerOptions = new JsonSerializerOptions()
+ {
+ WriteIndented = true,
+ Converters =
+ {
+ new JsonStringEnumConverter(), new DoubleToDecimalConverter(), new FloatToDecimalConverter(),
+ },
+ };
+
+ if (Environment.GetEnvironmentVariable("HELIX_CORRELATION_ID") != null)
+ {
+ Approvals.UseAssemblyLocationForApprovedFiles();
+ }
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_uci_adult_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetUciAdultDataView();
+ var pipeline = context.Auto().Featurizer(dataset, outputColumnName: "OutputFeature", excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_iris_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetIrisDataView();
+ var pipeline = context.Auto().Featurizer(dataset, excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+ }
+}

src/Microsoft.ML.AutoML/API/AutoCatalog.cs

@@ -4,8 +4,11 @@
 
 using System;
 using System.Collections.Generic;
+using System.Diagnostics.Contracts;
+using System.Linq;
 using Microsoft.ML.AutoML.CodeGen;
 using Microsoft.ML.Data;
+using Microsoft.ML.Runtime;
 using Microsoft.ML.SearchSpace;
 using Microsoft.ML.Trainers.FastTree;
 
@@ -538,55 +541,164 @@ public SweepableEstimator[] Regression(string labelColumnName = DefaultColumnNam
  /// <param name="inputColumnName">input column name.</param>
  internal SweepableEstimator[] TextFeaturizer(string outputColumnName, string inputColumnName)
  {
- throw new NotImplementedException();
+ var option = new FeaturizeTextOption
+ {
+ InputColumnName = inputColumnName,
+ OutputColumnName = outputColumnName,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateFeaturizeText(option) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing numeric columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] NumericFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] NumericFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(inputColumnNames, nameof(inputColumnNames));
+ Contracts.CheckValue(outputColumnNames, nameof(outputColumnNames));
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+ var replaceMissingValueOption = new ReplaceMissingValueOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateReplaceMissingValues(replaceMissingValueOption) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing catalog columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] CatalogFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] CatalogFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+
+ var option = new OneHotOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new SweepableEstimator[] { SweepableEstimatorFactory.CreateOneHotEncoding(option), SweepableEstimatorFactory.CreateOneHotHashEncoding(option) };
  }
 
  /// <summary>
  /// Create a single featurize pipeline according to <paramref name="data"/>. This function will collect all columns in <paramref name="data"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
  /// <param name="data">input data.</param>
  /// <param name="catalogColumns">columns that should be treated as catalog. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="numericColumns">columns that should be treated as numeric. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="textColumns">columns that should be treated as text. If not specified, it will automatically infer if a column is catalog or not.</param>
  /// <param name="outputColumnName">output feature column.</param>
  /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
-internal MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] excludeColumns = null)
+public MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] numericColumns = null, string[] textColumns = null, string[] excludeColumns = null)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+
+ // validate if there's overlapping among catalogColumns, numericColumns, textColumns and excludeColumns
+ var overallColumns = new string[][] { catalogColumns, numericColumns, textColumns, excludeColumns }
+ .Where(c => c != null)
+ .SelectMany(c => c);
+
+ if (overallColumns != null)
+ {
+ Contracts.Assert(overallColumns.Count() == overallColumns.Distinct().Count(), "detect overlapping among catalogColumns, numericColumns, textColumns and excludedColumns");
+ }
+
+ var columnInfo = new ColumnInformation();
+
+ if (excludeColumns != null)
+ {
+ foreach (var ignoreColumn in excludeColumns)
+ {
+ columnInfo.IgnoredColumnNames.Add(ignoreColumn);
+ }
+ }
+
+ if (catalogColumns != null)
+ {
+ foreach (var catalogColumn in catalogColumns)
+ {
+ columnInfo.CategoricalColumnNames.Add(catalogColumn);
+ }
+ }
+
+ if (numericColumns != null)
+ {
+ foreach (var column in numericColumns)
+ {
+ columnInfo.NumericColumnNames.Add(column);
+ }
+ }
+
+ if (textColumns != null)
+ {
+ foreach (var column in textColumns)
+ {
+ columnInfo.TextColumnNames.Add(column);
+ }
+ }
+
+ return this.Featurizer(data, columnInfo, outputColumnName);
  }
 
  /// <summary>
- /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/>,
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
+ /// <param name="data">input data.</param>
  /// <param name="columnInformation">column information.</param>
  /// <param name="outputColumnName">output feature column.</param>
- /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
- /// <returns></returns>
- internal MultiModelPipeline Featurizer(ColumnInformation columnInformation, string outputColumnName = "Features", string[] excludeColumns = null)
+ /// <returns>A <see cref="MultiModelPipeline"/> for featurization.</returns>
+ public MultiModelPipeline Featurizer(IDataView data, ColumnInformation columnInformation, string outputColumnName = "Features")
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+ Contracts.CheckValue(columnInformation, nameof(columnInformation));
+
+ var columnPurposes = PurposeInference.InferPurposes(this._context, data, columnInformation);
+ var textFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.TextFeature);
+ var numericFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.NumericFeature);
+ var catalogFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.CategoricalFeature);
+ var textFeatureColumnNames = textFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var numericFeatureColumnNames = numericFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var catalogFeatureColumnNames = catalogFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+
+ var pipeline = new MultiModelPipeline();
+ if (numericFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.NumericFeaturizer(numericFeatureColumnNames, numericFeatureColumnNames));
+ }
+
+ if (catalogFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.CatalogFeaturizer(catalogFeatureColumnNames, catalogFeatureColumnNames));
+ }
+
+ foreach (var textColumn in textFeatureColumnNames)
+ {
+ pipeline = pipeline.Append(this.TextFeaturizer(textColumn, textColumn));
+ }
+
+ var option = new ConcatOption
+ {
+ InputColumnNames = textFeatureColumnNames.Concat(numericFeatureColumnNames).Concat(catalogFeatureColumnNames).ToArray(),
+ OutputColumnName = outputColumnName,
+ };
+
+ if (option.InputColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(SweepableEstimatorFactory.CreateConcatenate(option));
+ }
+
+ return pipeline;
  }
  }
 }

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

@@ -182,7 +182,7 @@ internal HashingTransformer(IHostEnvironment env, params HashingEstimator.Column
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
 
  if (column.Combine && column.UseOrderedHashing)
  throw Host.ExceptParam(nameof(HashingEstimator.ColumnOptions.Combine), "When the 'Combine' option is specified, ordered hashing is not supported.");

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

@@ -184,7 +184,7 @@ internal NgramHashingTransformer(IHostEnvironment env, params NgramHashingEstima
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
  }
  }
 

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_iris_test.approved.txt

@@ -0,0 +1,34 @@
+{
+ "schema": "e0 * e1",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "OutputColumnName": "Features"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_uci_adult_test.approved.txt

@@ -0,0 +1,83 @@
+{
+ "schema": "e0 * (e1 \u002B e2) * e3",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "Features"
+ ],
+ "InputColumnNames": [
+ "Features"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "OneHotEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e2": {
+ "estimatorType": "OneHotHashEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e3": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "Features",
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "OutputColumnName": "OutputFeature"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/AutoFeaturizerTests.cs

@@ -0,0 +1,65 @@
+// 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;
+using System.Collections.Generic;
+using System.Text;
+using System.Text.Json;
+using Microsoft.ML.TestFramework;
+using Xunit;
+using Xunit.Abstractions;
+using ApprovalTests;
+using ApprovalTests.Namers;
+using ApprovalTests.Reporters;
+using System.Text.Json.Serialization;
+
+namespace Microsoft.ML.AutoML.Test
+{
+ public class AutoFeaturizerTests : BaseTestClass
+ {
+ private readonly JsonSerializerOptions _jsonSerializerOptions;
+
+ public AutoFeaturizerTests(ITestOutputHelper output)
+ : base(output)
+ {
+ _jsonSerializerOptions = new JsonSerializerOptions()
+ {
+ WriteIndented = true,
+ Converters =
+ {
+ new JsonStringEnumConverter(), new DoubleToDecimalConverter(), new FloatToDecimalConverter(),
+ },
+ };
+
+ if (Environment.GetEnvironmentVariable("HELIX_CORRELATION_ID") != null)
+ {
+ Approvals.UseAssemblyLocationForApprovedFiles();
+ }
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_uci_adult_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetUciAdultDataView();
+ var pipeline = context.Auto().Featurizer(dataset, outputColumnName: "OutputFeature", excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_iris_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetIrisDataView();
+ var pipeline = context.Auto().Featurizer(dataset, excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+ }
+}

src/Microsoft.ML.AutoML/API/AutoCatalog.cs

@@ -4,8 +4,11 @@
 
 using System;
 using System.Collections.Generic;
+using System.Diagnostics.Contracts;
+using System.Linq;
 using Microsoft.ML.AutoML.CodeGen;
 using Microsoft.ML.Data;
+using Microsoft.ML.Runtime;
 using Microsoft.ML.SearchSpace;
 using Microsoft.ML.Trainers.FastTree;
 
@@ -538,55 +541,164 @@ public SweepableEstimator[] Regression(string labelColumnName = DefaultColumnNam
  /// <param name="inputColumnName">input column name.</param>
  internal SweepableEstimator[] TextFeaturizer(string outputColumnName, string inputColumnName)
  {
- throw new NotImplementedException();
+ var option = new FeaturizeTextOption
+ {
+ InputColumnName = inputColumnName,
+ OutputColumnName = outputColumnName,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateFeaturizeText(option) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing numeric columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] NumericFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] NumericFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(inputColumnNames, nameof(inputColumnNames));
+ Contracts.CheckValue(outputColumnNames, nameof(outputColumnNames));
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+ var replaceMissingValueOption = new ReplaceMissingValueOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateReplaceMissingValues(replaceMissingValueOption) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing catalog columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] CatalogFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] CatalogFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+
+ var option = new OneHotOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new SweepableEstimator[] { SweepableEstimatorFactory.CreateOneHotEncoding(option), SweepableEstimatorFactory.CreateOneHotHashEncoding(option) };
  }
 
  /// <summary>
  /// Create a single featurize pipeline according to <paramref name="data"/>. This function will collect all columns in <paramref name="data"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
  /// <param name="data">input data.</param>
  /// <param name="catalogColumns">columns that should be treated as catalog. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="numericColumns">columns that should be treated as numeric. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="textColumns">columns that should be treated as text. If not specified, it will automatically infer if a column is catalog or not.</param>
  /// <param name="outputColumnName">output feature column.</param>
  /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
-internal MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] excludeColumns = null)
+public MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] numericColumns = null, string[] textColumns = null, string[] excludeColumns = null)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+
+ // validate if there's overlapping among catalogColumns, numericColumns, textColumns and excludeColumns
+ var overallColumns = new string[][] { catalogColumns, numericColumns, textColumns, excludeColumns }
+ .Where(c => c != null)
+ .SelectMany(c => c);
+
+ if (overallColumns != null)
+ {
+ Contracts.Assert(overallColumns.Count() == overallColumns.Distinct().Count(), "detect overlapping among catalogColumns, numericColumns, textColumns and excludedColumns");
+ }
+
+ var columnInfo = new ColumnInformation();
+
+ if (excludeColumns != null)
+ {
+ foreach (var ignoreColumn in excludeColumns)
+ {
+ columnInfo.IgnoredColumnNames.Add(ignoreColumn);
+ }
+ }
+
+ if (catalogColumns != null)
+ {
+ foreach (var catalogColumn in catalogColumns)
+ {
+ columnInfo.CategoricalColumnNames.Add(catalogColumn);
+ }
+ }
+
+ if (numericColumns != null)
+ {
+ foreach (var column in numericColumns)
+ {
+ columnInfo.NumericColumnNames.Add(column);
+ }
+ }
+
+ if (textColumns != null)
+ {
+ foreach (var column in textColumns)
+ {
+ columnInfo.TextColumnNames.Add(column);
+ }
+ }
+
+ return this.Featurizer(data, columnInfo, outputColumnName);
  }
 
  /// <summary>
- /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/>,
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
+ /// <param name="data">input data.</param>
  /// <param name="columnInformation">column information.</param>
  /// <param name="outputColumnName">output feature column.</param>
- /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
- /// <returns></returns>
- internal MultiModelPipeline Featurizer(ColumnInformation columnInformation, string outputColumnName = "Features", string[] excludeColumns = null)
+ /// <returns>A <see cref="MultiModelPipeline"/> for featurization.</returns>
+ public MultiModelPipeline Featurizer(IDataView data, ColumnInformation columnInformation, string outputColumnName = "Features")
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+ Contracts.CheckValue(columnInformation, nameof(columnInformation));
+
+ var columnPurposes = PurposeInference.InferPurposes(this._context, data, columnInformation);
+ var textFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.TextFeature);
+ var numericFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.NumericFeature);
+ var catalogFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.CategoricalFeature);
+ var textFeatureColumnNames = textFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var numericFeatureColumnNames = numericFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var catalogFeatureColumnNames = catalogFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+
+ var pipeline = new MultiModelPipeline();
+ if (numericFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.NumericFeaturizer(numericFeatureColumnNames, numericFeatureColumnNames));
+ }
+
+ if (catalogFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.CatalogFeaturizer(catalogFeatureColumnNames, catalogFeatureColumnNames));
+ }
+
+ foreach (var textColumn in textFeatureColumnNames)
+ {
+ pipeline = pipeline.Append(this.TextFeaturizer(textColumn, textColumn));
+ }
+
+ var option = new ConcatOption
+ {
+ InputColumnNames = textFeatureColumnNames.Concat(numericFeatureColumnNames).Concat(catalogFeatureColumnNames).ToArray(),
+ OutputColumnName = outputColumnName,
+ };
+
+ if (option.InputColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(SweepableEstimatorFactory.CreateConcatenate(option));
+ }
+
+ return pipeline;
  }
  }
 }

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

@@ -182,7 +182,7 @@ internal HashingTransformer(IHostEnvironment env, params HashingEstimator.Column
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
 
  if (column.Combine && column.UseOrderedHashing)
  throw Host.ExceptParam(nameof(HashingEstimator.ColumnOptions.Combine), "When the 'Combine' option is specified, ordered hashing is not supported.");

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

@@ -184,7 +184,7 @@ internal NgramHashingTransformer(IHostEnvironment env, params NgramHashingEstima
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
  }
  }
 

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_iris_test.approved.txt

@@ -0,0 +1,34 @@
+{
+ "schema": "e0 * e1",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "OutputColumnName": "Features"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_uci_adult_test.approved.txt

@@ -0,0 +1,83 @@
+{
+ "schema": "e0 * (e1 \u002B e2) * e3",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "Features"
+ ],
+ "InputColumnNames": [
+ "Features"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "OneHotEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e2": {
+ "estimatorType": "OneHotHashEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e3": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "Features",
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "OutputColumnName": "OutputFeature"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/AutoFeaturizerTests.cs

@@ -0,0 +1,65 @@
+// 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;
+using System.Collections.Generic;
+using System.Text;
+using System.Text.Json;
+using Microsoft.ML.TestFramework;
+using Xunit;
+using Xunit.Abstractions;
+using ApprovalTests;
+using ApprovalTests.Namers;
+using ApprovalTests.Reporters;
+using System.Text.Json.Serialization;
+
+namespace Microsoft.ML.AutoML.Test
+{
+ public class AutoFeaturizerTests : BaseTestClass
+ {
+ private readonly JsonSerializerOptions _jsonSerializerOptions;
+
+ public AutoFeaturizerTests(ITestOutputHelper output)
+ : base(output)
+ {
+ _jsonSerializerOptions = new JsonSerializerOptions()
+ {
+ WriteIndented = true,
+ Converters =
+ {
+ new JsonStringEnumConverter(), new DoubleToDecimalConverter(), new FloatToDecimalConverter(),
+ },
+ };
+
+ if (Environment.GetEnvironmentVariable("HELIX_CORRELATION_ID") != null)
+ {
+ Approvals.UseAssemblyLocationForApprovedFiles();
+ }
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_uci_adult_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetUciAdultDataView();
+ var pipeline = context.Auto().Featurizer(dataset, outputColumnName: "OutputFeature", excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_iris_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetIrisDataView();
+ var pipeline = context.Auto().Featurizer(dataset, excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+ }
+}

src/Microsoft.ML.AutoML/API/AutoCatalog.cs

@@ -4,8 +4,11 @@
 
 using System;
 using System.Collections.Generic;
+using System.Diagnostics.Contracts;
+using System.Linq;
 using Microsoft.ML.AutoML.CodeGen;
 using Microsoft.ML.Data;
+using Microsoft.ML.Runtime;
 using Microsoft.ML.SearchSpace;
 using Microsoft.ML.Trainers.FastTree;
 
@@ -538,55 +541,164 @@ public SweepableEstimator[] Regression(string labelColumnName = DefaultColumnNam
  /// <param name="inputColumnName">input column name.</param>
  internal SweepableEstimator[] TextFeaturizer(string outputColumnName, string inputColumnName)
  {
- throw new NotImplementedException();
+ var option = new FeaturizeTextOption
+ {
+ InputColumnName = inputColumnName,
+ OutputColumnName = outputColumnName,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateFeaturizeText(option) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing numeric columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] NumericFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] NumericFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(inputColumnNames, nameof(inputColumnNames));
+ Contracts.CheckValue(outputColumnNames, nameof(outputColumnNames));
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+ var replaceMissingValueOption = new ReplaceMissingValueOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateReplaceMissingValues(replaceMissingValueOption) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing catalog columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] CatalogFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] CatalogFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+
+ var option = new OneHotOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new SweepableEstimator[] { SweepableEstimatorFactory.CreateOneHotEncoding(option), SweepableEstimatorFactory.CreateOneHotHashEncoding(option) };
  }
 
  /// <summary>
  /// Create a single featurize pipeline according to <paramref name="data"/>. This function will collect all columns in <paramref name="data"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
  /// <param name="data">input data.</param>
  /// <param name="catalogColumns">columns that should be treated as catalog. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="numericColumns">columns that should be treated as numeric. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="textColumns">columns that should be treated as text. If not specified, it will automatically infer if a column is catalog or not.</param>
  /// <param name="outputColumnName">output feature column.</param>
  /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
-internal MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] excludeColumns = null)
+public MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] numericColumns = null, string[] textColumns = null, string[] excludeColumns = null)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+
+ // validate if there's overlapping among catalogColumns, numericColumns, textColumns and excludeColumns
+ var overallColumns = new string[][] { catalogColumns, numericColumns, textColumns, excludeColumns }
+ .Where(c => c != null)
+ .SelectMany(c => c);
+
+ if (overallColumns != null)
+ {
+ Contracts.Assert(overallColumns.Count() == overallColumns.Distinct().Count(), "detect overlapping among catalogColumns, numericColumns, textColumns and excludedColumns");
+ }
+
+ var columnInfo = new ColumnInformation();
+
+ if (excludeColumns != null)
+ {
+ foreach (var ignoreColumn in excludeColumns)
+ {
+ columnInfo.IgnoredColumnNames.Add(ignoreColumn);
+ }
+ }
+
+ if (catalogColumns != null)
+ {
+ foreach (var catalogColumn in catalogColumns)
+ {
+ columnInfo.CategoricalColumnNames.Add(catalogColumn);
+ }
+ }
+
+ if (numericColumns != null)
+ {
+ foreach (var column in numericColumns)
+ {
+ columnInfo.NumericColumnNames.Add(column);
+ }
+ }
+
+ if (textColumns != null)
+ {
+ foreach (var column in textColumns)
+ {
+ columnInfo.TextColumnNames.Add(column);
+ }
+ }
+
+ return this.Featurizer(data, columnInfo, outputColumnName);
  }
 
  /// <summary>
- /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/>,
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
+ /// <param name="data">input data.</param>
  /// <param name="columnInformation">column information.</param>
  /// <param name="outputColumnName">output feature column.</param>
- /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
- /// <returns></returns>
- internal MultiModelPipeline Featurizer(ColumnInformation columnInformation, string outputColumnName = "Features", string[] excludeColumns = null)
+ /// <returns>A <see cref="MultiModelPipeline"/> for featurization.</returns>
+ public MultiModelPipeline Featurizer(IDataView data, ColumnInformation columnInformation, string outputColumnName = "Features")
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+ Contracts.CheckValue(columnInformation, nameof(columnInformation));
+
+ var columnPurposes = PurposeInference.InferPurposes(this._context, data, columnInformation);
+ var textFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.TextFeature);
+ var numericFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.NumericFeature);
+ var catalogFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.CategoricalFeature);
+ var textFeatureColumnNames = textFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var numericFeatureColumnNames = numericFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var catalogFeatureColumnNames = catalogFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+
+ var pipeline = new MultiModelPipeline();
+ if (numericFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.NumericFeaturizer(numericFeatureColumnNames, numericFeatureColumnNames));
+ }
+
+ if (catalogFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.CatalogFeaturizer(catalogFeatureColumnNames, catalogFeatureColumnNames));
+ }
+
+ foreach (var textColumn in textFeatureColumnNames)
+ {
+ pipeline = pipeline.Append(this.TextFeaturizer(textColumn, textColumn));
+ }
+
+ var option = new ConcatOption
+ {
+ InputColumnNames = textFeatureColumnNames.Concat(numericFeatureColumnNames).Concat(catalogFeatureColumnNames).ToArray(),
+ OutputColumnName = outputColumnName,
+ };
+
+ if (option.InputColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(SweepableEstimatorFactory.CreateConcatenate(option));
+ }
+
+ return pipeline;
  }
  }
 }

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

@@ -182,7 +182,7 @@ internal HashingTransformer(IHostEnvironment env, params HashingEstimator.Column
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
 
  if (column.Combine && column.UseOrderedHashing)
  throw Host.ExceptParam(nameof(HashingEstimator.ColumnOptions.Combine), "When the 'Combine' option is specified, ordered hashing is not supported.");

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

@@ -184,7 +184,7 @@ internal NgramHashingTransformer(IHostEnvironment env, params NgramHashingEstima
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
  }
  }
 

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_iris_test.approved.txt

@@ -0,0 +1,34 @@
+{
+ "schema": "e0 * e1",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "OutputColumnName": "Features"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_uci_adult_test.approved.txt

@@ -0,0 +1,83 @@
+{
+ "schema": "e0 * (e1 \u002B e2) * e3",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "Features"
+ ],
+ "InputColumnNames": [
+ "Features"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "OneHotEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e2": {
+ "estimatorType": "OneHotHashEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e3": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "Features",
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "OutputColumnName": "OutputFeature"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/AutoFeaturizerTests.cs

@@ -0,0 +1,65 @@
+// 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;
+using System.Collections.Generic;
+using System.Text;
+using System.Text.Json;
+using Microsoft.ML.TestFramework;
+using Xunit;
+using Xunit.Abstractions;
+using ApprovalTests;
+using ApprovalTests.Namers;
+using ApprovalTests.Reporters;
+using System.Text.Json.Serialization;
+
+namespace Microsoft.ML.AutoML.Test
+{
+ public class AutoFeaturizerTests : BaseTestClass
+ {
+ private readonly JsonSerializerOptions _jsonSerializerOptions;
+
+ public AutoFeaturizerTests(ITestOutputHelper output)
+ : base(output)
+ {
+ _jsonSerializerOptions = new JsonSerializerOptions()
+ {
+ WriteIndented = true,
+ Converters =
+ {
+ new JsonStringEnumConverter(), new DoubleToDecimalConverter(), new FloatToDecimalConverter(),
+ },
+ };
+
+ if (Environment.GetEnvironmentVariable("HELIX_CORRELATION_ID") != null)
+ {
+ Approvals.UseAssemblyLocationForApprovedFiles();
+ }
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_uci_adult_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetUciAdultDataView();
+ var pipeline = context.Auto().Featurizer(dataset, outputColumnName: "OutputFeature", excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_iris_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetIrisDataView();
+ var pipeline = context.Auto().Featurizer(dataset, excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+ }
+}

src/Microsoft.ML.AutoML/API/AutoCatalog.cs

@@ -4,8 +4,11 @@
 
 using System;
 using System.Collections.Generic;
+using System.Diagnostics.Contracts;
+using System.Linq;
 using Microsoft.ML.AutoML.CodeGen;
 using Microsoft.ML.Data;
+using Microsoft.ML.Runtime;
 using Microsoft.ML.SearchSpace;
 using Microsoft.ML.Trainers.FastTree;
 
@@ -538,55 +541,164 @@ public SweepableEstimator[] Regression(string labelColumnName = DefaultColumnNam
  /// <param name="inputColumnName">input column name.</param>
  internal SweepableEstimator[] TextFeaturizer(string outputColumnName, string inputColumnName)
  {
- throw new NotImplementedException();
+ var option = new FeaturizeTextOption
+ {
+ InputColumnName = inputColumnName,
+ OutputColumnName = outputColumnName,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateFeaturizeText(option) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing numeric columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] NumericFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] NumericFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(inputColumnNames, nameof(inputColumnNames));
+ Contracts.CheckValue(outputColumnNames, nameof(outputColumnNames));
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+ var replaceMissingValueOption = new ReplaceMissingValueOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateReplaceMissingValues(replaceMissingValueOption) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing catalog columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] CatalogFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] CatalogFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+
+ var option = new OneHotOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new SweepableEstimator[] { SweepableEstimatorFactory.CreateOneHotEncoding(option), SweepableEstimatorFactory.CreateOneHotHashEncoding(option) };
  }
 
  /// <summary>
  /// Create a single featurize pipeline according to <paramref name="data"/>. This function will collect all columns in <paramref name="data"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
  /// <param name="data">input data.</param>
  /// <param name="catalogColumns">columns that should be treated as catalog. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="numericColumns">columns that should be treated as numeric. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="textColumns">columns that should be treated as text. If not specified, it will automatically infer if a column is catalog or not.</param>
  /// <param name="outputColumnName">output feature column.</param>
  /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
-internal MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] excludeColumns = null)
+public MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] numericColumns = null, string[] textColumns = null, string[] excludeColumns = null)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+
+ // validate if there's overlapping among catalogColumns, numericColumns, textColumns and excludeColumns
+ var overallColumns = new string[][] { catalogColumns, numericColumns, textColumns, excludeColumns }
+ .Where(c => c != null)
+ .SelectMany(c => c);
+
+ if (overallColumns != null)
+ {
+ Contracts.Assert(overallColumns.Count() == overallColumns.Distinct().Count(), "detect overlapping among catalogColumns, numericColumns, textColumns and excludedColumns");
+ }
+
+ var columnInfo = new ColumnInformation();
+
+ if (excludeColumns != null)
+ {
+ foreach (var ignoreColumn in excludeColumns)
+ {
+ columnInfo.IgnoredColumnNames.Add(ignoreColumn);
+ }
+ }
+
+ if (catalogColumns != null)
+ {
+ foreach (var catalogColumn in catalogColumns)
+ {
+ columnInfo.CategoricalColumnNames.Add(catalogColumn);
+ }
+ }
+
+ if (numericColumns != null)
+ {
+ foreach (var column in numericColumns)
+ {
+ columnInfo.NumericColumnNames.Add(column);
+ }
+ }
+
+ if (textColumns != null)
+ {
+ foreach (var column in textColumns)
+ {
+ columnInfo.TextColumnNames.Add(column);
+ }
+ }
+
+ return this.Featurizer(data, columnInfo, outputColumnName);
  }
 
  /// <summary>
- /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/>,
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
+ /// <param name="data">input data.</param>
  /// <param name="columnInformation">column information.</param>
  /// <param name="outputColumnName">output feature column.</param>
- /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
- /// <returns></returns>
- internal MultiModelPipeline Featurizer(ColumnInformation columnInformation, string outputColumnName = "Features", string[] excludeColumns = null)
+ /// <returns>A <see cref="MultiModelPipeline"/> for featurization.</returns>
+ public MultiModelPipeline Featurizer(IDataView data, ColumnInformation columnInformation, string outputColumnName = "Features")
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+ Contracts.CheckValue(columnInformation, nameof(columnInformation));
+
+ var columnPurposes = PurposeInference.InferPurposes(this._context, data, columnInformation);
+ var textFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.TextFeature);
+ var numericFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.NumericFeature);
+ var catalogFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.CategoricalFeature);
+ var textFeatureColumnNames = textFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var numericFeatureColumnNames = numericFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var catalogFeatureColumnNames = catalogFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+
+ var pipeline = new MultiModelPipeline();
+ if (numericFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.NumericFeaturizer(numericFeatureColumnNames, numericFeatureColumnNames));
+ }
+
+ if (catalogFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.CatalogFeaturizer(catalogFeatureColumnNames, catalogFeatureColumnNames));
+ }
+
+ foreach (var textColumn in textFeatureColumnNames)
+ {
+ pipeline = pipeline.Append(this.TextFeaturizer(textColumn, textColumn));
+ }
+
+ var option = new ConcatOption
+ {
+ InputColumnNames = textFeatureColumnNames.Concat(numericFeatureColumnNames).Concat(catalogFeatureColumnNames).ToArray(),
+ OutputColumnName = outputColumnName,
+ };
+
+ if (option.InputColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(SweepableEstimatorFactory.CreateConcatenate(option));
+ }
+
+ return pipeline;
  }
  }
 }

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

@@ -182,7 +182,7 @@ internal HashingTransformer(IHostEnvironment env, params HashingEstimator.Column
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
 
  if (column.Combine && column.UseOrderedHashing)
  throw Host.ExceptParam(nameof(HashingEstimator.ColumnOptions.Combine), "When the 'Combine' option is specified, ordered hashing is not supported.");

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

@@ -184,7 +184,7 @@ internal NgramHashingTransformer(IHostEnvironment env, params NgramHashingEstima
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
  }
  }
 

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_iris_test.approved.txt

@@ -0,0 +1,34 @@
+{
+ "schema": "e0 * e1",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "OutputColumnName": "Features"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_uci_adult_test.approved.txt

@@ -0,0 +1,83 @@
+{
+ "schema": "e0 * (e1 \u002B e2) * e3",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "Features"
+ ],
+ "InputColumnNames": [
+ "Features"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "OneHotEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e2": {
+ "estimatorType": "OneHotHashEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e3": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "Features",
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "OutputColumnName": "OutputFeature"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/AutoFeaturizerTests.cs

@@ -0,0 +1,65 @@
+// 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;
+using System.Collections.Generic;
+using System.Text;
+using System.Text.Json;
+using Microsoft.ML.TestFramework;
+using Xunit;
+using Xunit.Abstractions;
+using ApprovalTests;
+using ApprovalTests.Namers;
+using ApprovalTests.Reporters;
+using System.Text.Json.Serialization;
+
+namespace Microsoft.ML.AutoML.Test
+{
+ public class AutoFeaturizerTests : BaseTestClass
+ {
+ private readonly JsonSerializerOptions _jsonSerializerOptions;
+
+ public AutoFeaturizerTests(ITestOutputHelper output)
+ : base(output)
+ {
+ _jsonSerializerOptions = new JsonSerializerOptions()
+ {
+ WriteIndented = true,
+ Converters =
+ {
+ new JsonStringEnumConverter(), new DoubleToDecimalConverter(), new FloatToDecimalConverter(),
+ },
+ };
+
+ if (Environment.GetEnvironmentVariable("HELIX_CORRELATION_ID") != null)
+ {
+ Approvals.UseAssemblyLocationForApprovedFiles();
+ }
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_uci_adult_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetUciAdultDataView();
+ var pipeline = context.Auto().Featurizer(dataset, outputColumnName: "OutputFeature", excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_iris_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetIrisDataView();
+ var pipeline = context.Auto().Featurizer(dataset, excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+ }
+}

src/Microsoft.ML.AutoML/API/AutoCatalog.cs

@@ -4,8 +4,11 @@
 
 using System;
 using System.Collections.Generic;
+using System.Diagnostics.Contracts;
+using System.Linq;
 using Microsoft.ML.AutoML.CodeGen;
 using Microsoft.ML.Data;
+using Microsoft.ML.Runtime;
 using Microsoft.ML.SearchSpace;
 using Microsoft.ML.Trainers.FastTree;
 
@@ -538,55 +541,164 @@ public SweepableEstimator[] Regression(string labelColumnName = DefaultColumnNam
  /// <param name="inputColumnName">input column name.</param>
  internal SweepableEstimator[] TextFeaturizer(string outputColumnName, string inputColumnName)
  {
- throw new NotImplementedException();
+ var option = new FeaturizeTextOption
+ {
+ InputColumnName = inputColumnName,
+ OutputColumnName = outputColumnName,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateFeaturizeText(option) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing numeric columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] NumericFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] NumericFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(inputColumnNames, nameof(inputColumnNames));
+ Contracts.CheckValue(outputColumnNames, nameof(outputColumnNames));
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+ var replaceMissingValueOption = new ReplaceMissingValueOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateReplaceMissingValues(replaceMissingValueOption) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing catalog columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] CatalogFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] CatalogFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+
+ var option = new OneHotOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new SweepableEstimator[] { SweepableEstimatorFactory.CreateOneHotEncoding(option), SweepableEstimatorFactory.CreateOneHotHashEncoding(option) };
  }
 
  /// <summary>
  /// Create a single featurize pipeline according to <paramref name="data"/>. This function will collect all columns in <paramref name="data"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
  /// <param name="data">input data.</param>
  /// <param name="catalogColumns">columns that should be treated as catalog. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="numericColumns">columns that should be treated as numeric. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="textColumns">columns that should be treated as text. If not specified, it will automatically infer if a column is catalog or not.</param>
  /// <param name="outputColumnName">output feature column.</param>
  /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
-internal MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] excludeColumns = null)
+public MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] numericColumns = null, string[] textColumns = null, string[] excludeColumns = null)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+
+ // validate if there's overlapping among catalogColumns, numericColumns, textColumns and excludeColumns
+ var overallColumns = new string[][] { catalogColumns, numericColumns, textColumns, excludeColumns }
+ .Where(c => c != null)
+ .SelectMany(c => c);
+
+ if (overallColumns != null)
+ {
+ Contracts.Assert(overallColumns.Count() == overallColumns.Distinct().Count(), "detect overlapping among catalogColumns, numericColumns, textColumns and excludedColumns");
+ }
+
+ var columnInfo = new ColumnInformation();
+
+ if (excludeColumns != null)
+ {
+ foreach (var ignoreColumn in excludeColumns)
+ {
+ columnInfo.IgnoredColumnNames.Add(ignoreColumn);
+ }
+ }
+
+ if (catalogColumns != null)
+ {
+ foreach (var catalogColumn in catalogColumns)
+ {
+ columnInfo.CategoricalColumnNames.Add(catalogColumn);
+ }
+ }
+
+ if (numericColumns != null)
+ {
+ foreach (var column in numericColumns)
+ {
+ columnInfo.NumericColumnNames.Add(column);
+ }
+ }
+
+ if (textColumns != null)
+ {
+ foreach (var column in textColumns)
+ {
+ columnInfo.TextColumnNames.Add(column);
+ }
+ }
+
+ return this.Featurizer(data, columnInfo, outputColumnName);
  }
 
  /// <summary>
- /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/>,
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
+ /// <param name="data">input data.</param>
  /// <param name="columnInformation">column information.</param>
  /// <param name="outputColumnName">output feature column.</param>
- /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
- /// <returns></returns>
- internal MultiModelPipeline Featurizer(ColumnInformation columnInformation, string outputColumnName = "Features", string[] excludeColumns = null)
+ /// <returns>A <see cref="MultiModelPipeline"/> for featurization.</returns>
+ public MultiModelPipeline Featurizer(IDataView data, ColumnInformation columnInformation, string outputColumnName = "Features")
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+ Contracts.CheckValue(columnInformation, nameof(columnInformation));
+
+ var columnPurposes = PurposeInference.InferPurposes(this._context, data, columnInformation);
+ var textFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.TextFeature);
+ var numericFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.NumericFeature);
+ var catalogFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.CategoricalFeature);
+ var textFeatureColumnNames = textFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var numericFeatureColumnNames = numericFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var catalogFeatureColumnNames = catalogFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+
+ var pipeline = new MultiModelPipeline();
+ if (numericFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.NumericFeaturizer(numericFeatureColumnNames, numericFeatureColumnNames));
+ }
+
+ if (catalogFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.CatalogFeaturizer(catalogFeatureColumnNames, catalogFeatureColumnNames));
+ }
+
+ foreach (var textColumn in textFeatureColumnNames)
+ {
+ pipeline = pipeline.Append(this.TextFeaturizer(textColumn, textColumn));
+ }
+
+ var option = new ConcatOption
+ {
+ InputColumnNames = textFeatureColumnNames.Concat(numericFeatureColumnNames).Concat(catalogFeatureColumnNames).ToArray(),
+ OutputColumnName = outputColumnName,
+ };
+
+ if (option.InputColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(SweepableEstimatorFactory.CreateConcatenate(option));
+ }
+
+ return pipeline;
  }
  }
 }

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

@@ -182,7 +182,7 @@ internal HashingTransformer(IHostEnvironment env, params HashingEstimator.Column
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
 
  if (column.Combine && column.UseOrderedHashing)
  throw Host.ExceptParam(nameof(HashingEstimator.ColumnOptions.Combine), "When the 'Combine' option is specified, ordered hashing is not supported.");

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

@@ -184,7 +184,7 @@ internal NgramHashingTransformer(IHostEnvironment env, params NgramHashingEstima
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
  }
  }
 

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_iris_test.approved.txt

@@ -0,0 +1,34 @@
+{
+ "schema": "e0 * e1",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "OutputColumnName": "Features"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_uci_adult_test.approved.txt

@@ -0,0 +1,83 @@
+{
+ "schema": "e0 * (e1 \u002B e2) * e3",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "Features"
+ ],
+ "InputColumnNames": [
+ "Features"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "OneHotEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e2": {
+ "estimatorType": "OneHotHashEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e3": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "Features",
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "OutputColumnName": "OutputFeature"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/AutoFeaturizerTests.cs

@@ -0,0 +1,65 @@
+// 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;
+using System.Collections.Generic;
+using System.Text;
+using System.Text.Json;
+using Microsoft.ML.TestFramework;
+using Xunit;
+using Xunit.Abstractions;
+using ApprovalTests;
+using ApprovalTests.Namers;
+using ApprovalTests.Reporters;
+using System.Text.Json.Serialization;
+
+namespace Microsoft.ML.AutoML.Test
+{
+ public class AutoFeaturizerTests : BaseTestClass
+ {
+ private readonly JsonSerializerOptions _jsonSerializerOptions;
+
+ public AutoFeaturizerTests(ITestOutputHelper output)
+ : base(output)
+ {
+ _jsonSerializerOptions = new JsonSerializerOptions()
+ {
+ WriteIndented = true,
+ Converters =
+ {
+ new JsonStringEnumConverter(), new DoubleToDecimalConverter(), new FloatToDecimalConverter(),
+ },
+ };
+
+ if (Environment.GetEnvironmentVariable("HELIX_CORRELATION_ID") != null)
+ {
+ Approvals.UseAssemblyLocationForApprovedFiles();
+ }
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_uci_adult_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetUciAdultDataView();
+ var pipeline = context.Auto().Featurizer(dataset, outputColumnName: "OutputFeature", excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_iris_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetIrisDataView();
+ var pipeline = context.Auto().Featurizer(dataset, excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+ }
+}

src/Microsoft.ML.AutoML/API/AutoCatalog.cs

@@ -4,8 +4,11 @@
 
 using System;
 using System.Collections.Generic;
+using System.Diagnostics.Contracts;
+using System.Linq;
 using Microsoft.ML.AutoML.CodeGen;
 using Microsoft.ML.Data;
+using Microsoft.ML.Runtime;
 using Microsoft.ML.SearchSpace;
 using Microsoft.ML.Trainers.FastTree;
 
@@ -538,55 +541,164 @@ public SweepableEstimator[] Regression(string labelColumnName = DefaultColumnNam
  /// <param name="inputColumnName">input column name.</param>
  internal SweepableEstimator[] TextFeaturizer(string outputColumnName, string inputColumnName)
  {
- throw new NotImplementedException();
+ var option = new FeaturizeTextOption
+ {
+ InputColumnName = inputColumnName,
+ OutputColumnName = outputColumnName,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateFeaturizeText(option) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing numeric columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] NumericFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] NumericFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(inputColumnNames, nameof(inputColumnNames));
+ Contracts.CheckValue(outputColumnNames, nameof(outputColumnNames));
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+ var replaceMissingValueOption = new ReplaceMissingValueOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateReplaceMissingValues(replaceMissingValueOption) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing catalog columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] CatalogFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] CatalogFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+
+ var option = new OneHotOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new SweepableEstimator[] { SweepableEstimatorFactory.CreateOneHotEncoding(option), SweepableEstimatorFactory.CreateOneHotHashEncoding(option) };
  }
 
  /// <summary>
  /// Create a single featurize pipeline according to <paramref name="data"/>. This function will collect all columns in <paramref name="data"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
  /// <param name="data">input data.</param>
  /// <param name="catalogColumns">columns that should be treated as catalog. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="numericColumns">columns that should be treated as numeric. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="textColumns">columns that should be treated as text. If not specified, it will automatically infer if a column is catalog or not.</param>
  /// <param name="outputColumnName">output feature column.</param>
  /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
-internal MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] excludeColumns = null)
+public MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] numericColumns = null, string[] textColumns = null, string[] excludeColumns = null)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+
+ // validate if there's overlapping among catalogColumns, numericColumns, textColumns and excludeColumns
+ var overallColumns = new string[][] { catalogColumns, numericColumns, textColumns, excludeColumns }
+ .Where(c => c != null)
+ .SelectMany(c => c);
+
+ if (overallColumns != null)
+ {
+ Contracts.Assert(overallColumns.Count() == overallColumns.Distinct().Count(), "detect overlapping among catalogColumns, numericColumns, textColumns and excludedColumns");
+ }
+
+ var columnInfo = new ColumnInformation();
+
+ if (excludeColumns != null)
+ {
+ foreach (var ignoreColumn in excludeColumns)
+ {
+ columnInfo.IgnoredColumnNames.Add(ignoreColumn);
+ }
+ }
+
+ if (catalogColumns != null)
+ {
+ foreach (var catalogColumn in catalogColumns)
+ {
+ columnInfo.CategoricalColumnNames.Add(catalogColumn);
+ }
+ }
+
+ if (numericColumns != null)
+ {
+ foreach (var column in numericColumns)
+ {
+ columnInfo.NumericColumnNames.Add(column);
+ }
+ }
+
+ if (textColumns != null)
+ {
+ foreach (var column in textColumns)
+ {
+ columnInfo.TextColumnNames.Add(column);
+ }
+ }
+
+ return this.Featurizer(data, columnInfo, outputColumnName);
  }
 
  /// <summary>
- /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/>,
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
+ /// <param name="data">input data.</param>
  /// <param name="columnInformation">column information.</param>
  /// <param name="outputColumnName">output feature column.</param>
- /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
- /// <returns></returns>
- internal MultiModelPipeline Featurizer(ColumnInformation columnInformation, string outputColumnName = "Features", string[] excludeColumns = null)
+ /// <returns>A <see cref="MultiModelPipeline"/> for featurization.</returns>
+ public MultiModelPipeline Featurizer(IDataView data, ColumnInformation columnInformation, string outputColumnName = "Features")
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+ Contracts.CheckValue(columnInformation, nameof(columnInformation));
+
+ var columnPurposes = PurposeInference.InferPurposes(this._context, data, columnInformation);
+ var textFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.TextFeature);
+ var numericFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.NumericFeature);
+ var catalogFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.CategoricalFeature);
+ var textFeatureColumnNames = textFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var numericFeatureColumnNames = numericFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var catalogFeatureColumnNames = catalogFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+
+ var pipeline = new MultiModelPipeline();
+ if (numericFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.NumericFeaturizer(numericFeatureColumnNames, numericFeatureColumnNames));
+ }
+
+ if (catalogFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.CatalogFeaturizer(catalogFeatureColumnNames, catalogFeatureColumnNames));
+ }
+
+ foreach (var textColumn in textFeatureColumnNames)
+ {
+ pipeline = pipeline.Append(this.TextFeaturizer(textColumn, textColumn));
+ }
+
+ var option = new ConcatOption
+ {
+ InputColumnNames = textFeatureColumnNames.Concat(numericFeatureColumnNames).Concat(catalogFeatureColumnNames).ToArray(),
+ OutputColumnName = outputColumnName,
+ };
+
+ if (option.InputColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(SweepableEstimatorFactory.CreateConcatenate(option));
+ }
+
+ return pipeline;
  }
  }
 }

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

@@ -182,7 +182,7 @@ internal HashingTransformer(IHostEnvironment env, params HashingEstimator.Column
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
 
  if (column.Combine && column.UseOrderedHashing)
  throw Host.ExceptParam(nameof(HashingEstimator.ColumnOptions.Combine), "When the 'Combine' option is specified, ordered hashing is not supported.");

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

@@ -184,7 +184,7 @@ internal NgramHashingTransformer(IHostEnvironment env, params NgramHashingEstima
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
  }
  }
 

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_iris_test.approved.txt

@@ -0,0 +1,34 @@
+{
+ "schema": "e0 * e1",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "OutputColumnName": "Features"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_uci_adult_test.approved.txt

@@ -0,0 +1,83 @@
+{
+ "schema": "e0 * (e1 \u002B e2) * e3",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "Features"
+ ],
+ "InputColumnNames": [
+ "Features"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "OneHotEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e2": {
+ "estimatorType": "OneHotHashEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e3": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "Features",
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "OutputColumnName": "OutputFeature"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/AutoFeaturizerTests.cs

@@ -0,0 +1,65 @@
+// 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;
+using System.Collections.Generic;
+using System.Text;
+using System.Text.Json;
+using Microsoft.ML.TestFramework;
+using Xunit;
+using Xunit.Abstractions;
+using ApprovalTests;
+using ApprovalTests.Namers;
+using ApprovalTests.Reporters;
+using System.Text.Json.Serialization;
+
+namespace Microsoft.ML.AutoML.Test
+{
+ public class AutoFeaturizerTests : BaseTestClass
+ {
+ private readonly JsonSerializerOptions _jsonSerializerOptions;
+
+ public AutoFeaturizerTests(ITestOutputHelper output)
+ : base(output)
+ {
+ _jsonSerializerOptions = new JsonSerializerOptions()
+ {
+ WriteIndented = true,
+ Converters =
+ {
+ new JsonStringEnumConverter(), new DoubleToDecimalConverter(), new FloatToDecimalConverter(),
+ },
+ };
+
+ if (Environment.GetEnvironmentVariable("HELIX_CORRELATION_ID") != null)
+ {
+ Approvals.UseAssemblyLocationForApprovedFiles();
+ }
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_uci_adult_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetUciAdultDataView();
+ var pipeline = context.Auto().Featurizer(dataset, outputColumnName: "OutputFeature", excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_iris_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetIrisDataView();
+ var pipeline = context.Auto().Featurizer(dataset, excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+ }
+}

src/Microsoft.ML.AutoML/API/AutoCatalog.cs

@@ -4,8 +4,11 @@
 
 using System;
 using System.Collections.Generic;
+using System.Diagnostics.Contracts;
+using System.Linq;
 using Microsoft.ML.AutoML.CodeGen;
 using Microsoft.ML.Data;
+using Microsoft.ML.Runtime;
 using Microsoft.ML.SearchSpace;
 using Microsoft.ML.Trainers.FastTree;
 
@@ -538,55 +541,164 @@ public SweepableEstimator[] Regression(string labelColumnName = DefaultColumnNam
  /// <param name="inputColumnName">input column name.</param>
  internal SweepableEstimator[] TextFeaturizer(string outputColumnName, string inputColumnName)
  {
- throw new NotImplementedException();
+ var option = new FeaturizeTextOption
+ {
+ InputColumnName = inputColumnName,
+ OutputColumnName = outputColumnName,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateFeaturizeText(option) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing numeric columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] NumericFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] NumericFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(inputColumnNames, nameof(inputColumnNames));
+ Contracts.CheckValue(outputColumnNames, nameof(outputColumnNames));
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+ var replaceMissingValueOption = new ReplaceMissingValueOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new[] { SweepableEstimatorFactory.CreateReplaceMissingValues(replaceMissingValueOption) };
  }
 
  /// <summary>
  /// Create a list of <see cref="SweepableEstimator"/> for featurizing catalog columns.
  /// </summary>
- /// <param name="outputColumnName">output column name.</param>
- /// <param name="inputColumnName">input column name.</param>
- internal SweepableEstimator[] CatalogFeaturizer(string outputColumnName, string inputColumnName)
+ /// <param name="outputColumnNames">output column names.</param>
+ /// <param name="inputColumnNames">input column names.</param>
+ internal SweepableEstimator[] CatalogFeaturizer(string[] outputColumnNames, string[] inputColumnNames)
  {
- throw new NotImplementedException();
+ Contracts.Check(outputColumnNames.Count() == inputColumnNames.Count() && outputColumnNames.Count() > 0, "outputColumnNames and inputColumnNames must have the same length and greater than 0");
+
+ var option = new OneHotOption
+ {
+ InputColumnNames = inputColumnNames,
+ OutputColumnNames = outputColumnNames,
+ };
+
+ return new SweepableEstimator[] { SweepableEstimatorFactory.CreateOneHotEncoding(option), SweepableEstimatorFactory.CreateOneHotHashEncoding(option) };
  }
 
  /// <summary>
  /// Create a single featurize pipeline according to <paramref name="data"/>. This function will collect all columns in <paramref name="data"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
  /// <param name="data">input data.</param>
  /// <param name="catalogColumns">columns that should be treated as catalog. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="numericColumns">columns that should be treated as numeric. If not specified, it will automatically infer if a column is catalog or not.</param>
+ /// <param name="textColumns">columns that should be treated as text. If not specified, it will automatically infer if a column is catalog or not.</param>
  /// <param name="outputColumnName">output feature column.</param>
  /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
-internal MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] excludeColumns = null)
+public MultiModelPipeline Featurizer(IDataView data, string outputColumnName = "Features", string[] catalogColumns = null, string[] numericColumns = null, string[] textColumns = null, string[] excludeColumns = null)
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+
+ // validate if there's overlapping among catalogColumns, numericColumns, textColumns and excludeColumns
+ var overallColumns = new string[][] { catalogColumns, numericColumns, textColumns, excludeColumns }
+ .Where(c => c != null)
+ .SelectMany(c => c);
+
+ if (overallColumns != null)
+ {
+ Contracts.Assert(overallColumns.Count() == overallColumns.Distinct().Count(), "detect overlapping among catalogColumns, numericColumns, textColumns and excludedColumns");
+ }
+
+ var columnInfo = new ColumnInformation();
+
+ if (excludeColumns != null)
+ {
+ foreach (var ignoreColumn in excludeColumns)
+ {
+ columnInfo.IgnoredColumnNames.Add(ignoreColumn);
+ }
+ }
+
+ if (catalogColumns != null)
+ {
+ foreach (var catalogColumn in catalogColumns)
+ {
+ columnInfo.CategoricalColumnNames.Add(catalogColumn);
+ }
+ }
+
+ if (numericColumns != null)
+ {
+ foreach (var column in numericColumns)
+ {
+ columnInfo.NumericColumnNames.Add(column);
+ }
+ }
+
+ if (textColumns != null)
+ {
+ foreach (var column in textColumns)
+ {
+ columnInfo.TextColumnNames.Add(column);
+ }
+ }
+
+ return this.Featurizer(data, columnInfo, outputColumnName);
  }
 
  /// <summary>
- /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/> and not in <paramref name="excludeColumns"/>,
- /// featurizing them using <see cref="CatalogFeaturizer(string, string)"/>, <see cref="NumericFeaturizer(string, string)"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
+ /// Create a single featurize pipeline according to <paramref name="columnInformation"/>. This function will collect all columns in <paramref name="columnInformation"/>,
+ /// featurizing them using <see cref="CatalogFeaturizer(string[], string[])"/>, <see cref="NumericFeaturizer(string[], string[])"/> or <see cref="TextFeaturizer(string, string)"/>. And combine
  /// them into a single feature column as output.
  /// </summary>
+ /// <param name="data">input data.</param>
  /// <param name="columnInformation">column information.</param>
  /// <param name="outputColumnName">output feature column.</param>
- /// <param name="excludeColumns">columns that won't be included when featurizing, like label</param>
- /// <returns></returns>
- internal MultiModelPipeline Featurizer(ColumnInformation columnInformation, string outputColumnName = "Features", string[] excludeColumns = null)
+ /// <returns>A <see cref="MultiModelPipeline"/> for featurization.</returns>
+ public MultiModelPipeline Featurizer(IDataView data, ColumnInformation columnInformation, string outputColumnName = "Features")
  {
- throw new NotImplementedException();
+ Contracts.CheckValue(data, nameof(data));
+ Contracts.CheckValue(columnInformation, nameof(columnInformation));
+
+ var columnPurposes = PurposeInference.InferPurposes(this._context, data, columnInformation);
+ var textFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.TextFeature);
+ var numericFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.NumericFeature);
+ var catalogFeatures = columnPurposes.Where(c => c.Purpose == ColumnPurpose.CategoricalFeature);
+ var textFeatureColumnNames = textFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var numericFeatureColumnNames = numericFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+ var catalogFeatureColumnNames = catalogFeatures.Select(c => data.Schema[c.ColumnIndex].Name).ToArray();
+
+ var pipeline = new MultiModelPipeline();
+ if (numericFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.NumericFeaturizer(numericFeatureColumnNames, numericFeatureColumnNames));
+ }
+
+ if (catalogFeatureColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(this.CatalogFeaturizer(catalogFeatureColumnNames, catalogFeatureColumnNames));
+ }
+
+ foreach (var textColumn in textFeatureColumnNames)
+ {
+ pipeline = pipeline.Append(this.TextFeaturizer(textColumn, textColumn));
+ }
+
+ var option = new ConcatOption
+ {
+ InputColumnNames = textFeatureColumnNames.Concat(numericFeatureColumnNames).Concat(catalogFeatureColumnNames).ToArray(),
+ OutputColumnName = outputColumnName,
+ };
+
+ if (option.InputColumnNames.Length > 0)
+ {
+ pipeline = pipeline.Append(SweepableEstimatorFactory.CreateConcatenate(option));
+ }
+
+ return pipeline;
  }
  }
 }

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

@@ -182,7 +182,7 @@ internal HashingTransformer(IHostEnvironment env, params HashingEstimator.Column
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found column with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(HashingEstimator)} instead");
 
  if (column.Combine && column.UseOrderedHashing)
  throw Host.ExceptParam(nameof(HashingEstimator.ColumnOptions.Combine), "When the 'Combine' option is specified, ordered hashing is not supported.");

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

@@ -184,7 +184,7 @@ internal NgramHashingTransformer(IHostEnvironment env, params NgramHashingEstima
  foreach (var column in _columns)
  {
  if (column.MaximumNumberOfInverts != 0)
- throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
+ throw Host.ExceptParam(nameof(columns), $"Found colunm with {nameof(column.MaximumNumberOfInverts)} set to non zero value, please use {nameof(NgramHashingEstimator)} instead");
  }
  }
 

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_iris_test.approved.txt

@@ -0,0 +1,34 @@
+{
+ "schema": "e0 * e1",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "col1",
+ "col2",
+ "col3",
+ "col4"
+ ],
+ "OutputColumnName": "Features"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/ApprovalTests/AutoFeaturizerTests.AutoFeaturizer_uci_adult_test.approved.txt

@@ -0,0 +1,83 @@
+{
+ "schema": "e0 * (e1 \u002B e2) * e3",
+ "estimators": {
+ "e0": {
+ "estimatorType": "ReplaceMissingValues",
+ "parameter": {
+ "OutputColumnNames": [
+ "Features"
+ ],
+ "InputColumnNames": [
+ "Features"
+ ]
+ }
+ },
+ "e1": {
+ "estimatorType": "OneHotEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e2": {
+ "estimatorType": "OneHotHashEncoding",
+ "parameter": {
+ "OutputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "InputColumnNames": [
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ]
+ }
+ },
+ "e3": {
+ "estimatorType": "Concatenate",
+ "parameter": {
+ "InputColumnNames": [
+ "Features",
+ "Workclass",
+ "education",
+ "marital-status",
+ "occupation",
+ "relationship",
+ "ethnicity",
+ "sex",
+ "native-country-region"
+ ],
+ "OutputColumnName": "OutputFeature"
+ }
+ }
+ }
+}

test/Microsoft.ML.AutoML.Tests/AutoFeaturizerTests.cs

@@ -0,0 +1,65 @@
+// 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;
+using System.Collections.Generic;
+using System.Text;
+using System.Text.Json;
+using Microsoft.ML.TestFramework;
+using Xunit;
+using Xunit.Abstractions;
+using ApprovalTests;
+using ApprovalTests.Namers;
+using ApprovalTests.Reporters;
+using System.Text.Json.Serialization;
+
+namespace Microsoft.ML.AutoML.Test
+{
+ public class AutoFeaturizerTests : BaseTestClass
+ {
+ private readonly JsonSerializerOptions _jsonSerializerOptions;
+
+ public AutoFeaturizerTests(ITestOutputHelper output)
+ : base(output)
+ {
+ _jsonSerializerOptions = new JsonSerializerOptions()
+ {
+ WriteIndented = true,
+ Converters =
+ {
+ new JsonStringEnumConverter(), new DoubleToDecimalConverter(), new FloatToDecimalConverter(),
+ },
+ };
+
+ if (Environment.GetEnvironmentVariable("HELIX_CORRELATION_ID") != null)
+ {
+ Approvals.UseAssemblyLocationForApprovedFiles();
+ }
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_uci_adult_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetUciAdultDataView();
+ var pipeline = context.Auto().Featurizer(dataset, outputColumnName: "OutputFeature", excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+
+ [Fact]
+ [UseReporter(typeof(DiffReporter))]
+ [UseApprovalSubdirectory("ApprovalTests")]
+ public void AutoFeaturizer_iris_test()
+ {
+ var context = new MLContext(1);
+ var dataset = DatasetUtil.GetIrisDataView();
+ var pipeline = context.Auto().Featurizer(dataset, excludeColumns: new[] { "Label" });
+
+ Approvals.Verify(JsonSerializer.Serialize(pipeline, _jsonSerializerOptions));
+ }
+ }
+}