Skip to content

Commit 745167b

Browse files
sfilipisfilipi
authored
Renaming transforms to transformers Part 1 (dotnet#1588)
* renaming transforms to transformers Part 1
1 parent e803520 commit 745167b

File tree

25 files changed

+200
-200
lines changed

25 files changed

+200
-200
lines changed

‎src/Microsoft.ML.Ensemble/Selector/SubsetSelector/BootstrapSelector.cs

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -46,7 +46,7 @@ public override IEnumerable<Subset> GetSubsets(Batch batch, IRandom rand)
4646
for (int i = 0; i < Size; i++)
4747
{
4848
// REVIEW: Consider ways to reintroduce "balanced" samples.
49-
var viewTrain = new BootstrapSampleTransform(Host, new BootstrapSampleTransform.Arguments(), Data.Data);
49+
var viewTrain = new BootstrapSampleTransformer(Host, new BootstrapSampleTransformer.Arguments(), Data.Data);
5050
var dataTrain = new RoleMappedData(viewTrain, Data.Schema.GetColumnRoleNames());
5151
yield return FeatureSelector.SelectFeatures(dataTrain, rand);
5252
}

‎src/Microsoft.ML.Legacy/AssemblyRegistration.cs

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -46,7 +46,7 @@ private static bool LoadStandardAssemblies()
4646
_ = typeof(Maml).Assembly; // ML.Maml
4747
_ = typeof(PcaPredictor).Assembly; // ML.PCA
4848
_ = typeof(SweepCommand).Assembly; // ML.Sweeper
49-
_ = typeof(CategoricalTransform).Assembly; // ML.Transforms
49+
_ = typeof(OneHotEncodingTransformer).Assembly; // ML.Transforms
5050

5151
// The following assemblies reference this assembly, so we can't directly reference them
5252

‎src/Microsoft.ML.Legacy/CSharpApi.cs

Lines changed: 23 additions & 23 deletions
Original file line numberDiff line numberDiff line change
@@ -11061,7 +11061,7 @@ public BinNormalizerPipelineStep(Output output)
1106111061

1106211062
namespace Legacy.Transforms
1106311063
{
11064-
public enum CategoricalTransformOutputKind : byte
11064+
public enum OneHotEncodingTransformerOutputKind : byte
1106511065
{
1106611066
Bag = 1,
1106711067
Ind = 2,
@@ -11070,7 +11070,7 @@ public enum CategoricalTransformOutputKind : byte
1107011070
}
1107111071

1107211072

11073-
public sealed partial class CategoricalHashTransformColumn : OneToOneColumn<CategoricalHashTransformColumn>, IOneToOneColumn
11073+
public sealed partial class OneHotHashEncodingTransformerColumn : OneToOneColumn<OneHotHashEncodingTransformerColumn>, IOneToOneColumn
1107411074
{
1107511075
/// <summary>
1107611076
/// The number of bits to hash into. Must be between 1 and 30, inclusive.
@@ -11095,7 +11095,7 @@ public sealed partial class CategoricalHashTransformColumn : OneToOneColumn<Cate
1109511095
/// <summary>
1109611096
/// Output kind: Bag (multi-set vector), Ind (indicator vector), or Key (index)
1109711097
/// </summary>
11098-
public CategoricalTransformOutputKind? OutputKind { get; set; }
11098+
public OneHotEncodingTransformerOutputKind? OutputKind { get; set; }
1109911099

1110011100
/// <summary>
1110111101
/// Name of the new column
@@ -11142,23 +11142,23 @@ public CategoricalHashOneHotVectorizer(params (string inputColumn, string output
1114211142

1114311143
public void AddColumn(string inputColumn)
1114411144
{
11145-
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.CategoricalHashTransformColumn>() : new List<Microsoft.ML.Legacy.Transforms.CategoricalHashTransformColumn>(Column);
11146-
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.CategoricalHashTransformColumn>.Create(inputColumn));
11145+
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.OneHotHashEncodingTransformerColumn>() : new List<Microsoft.ML.Legacy.Transforms.OneHotHashEncodingTransformerColumn>(Column);
11146+
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.OneHotHashEncodingTransformerColumn>.Create(inputColumn));
1114711147
Column = list.ToArray();
1114811148
}
1114911149

1115011150
public void AddColumn(string outputColumn, string inputColumn)
1115111151
{
11152-
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.CategoricalHashTransformColumn>() : new List<Microsoft.ML.Legacy.Transforms.CategoricalHashTransformColumn>(Column);
11153-
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.CategoricalHashTransformColumn>.Create(outputColumn, inputColumn));
11152+
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.OneHotHashEncodingTransformerColumn>() : new List<Microsoft.ML.Legacy.Transforms.OneHotHashEncodingTransformerColumn>(Column);
11153+
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.OneHotHashEncodingTransformerColumn>.Create(outputColumn, inputColumn));
1115411154
Column = list.ToArray();
1115511155
}
1115611156

1115711157

1115811158
/// <summary>
1115911159
/// New column definition(s) (optional form: name:hashBits:src)
1116011160
/// </summary>
11161-
public CategoricalHashTransformColumn[] Column { get; set; }
11161+
public OneHotHashEncodingTransformerColumn[] Column { get; set; }
1116211162

1116311163
/// <summary>
1116411164
/// Number of bits to hash into. Must be between 1 and 30, inclusive.
@@ -11183,7 +11183,7 @@ public void AddColumn(string outputColumn, string inputColumn)
1118311183
/// <summary>
1118411184
/// Output kind: Bag (multi-set vector), Ind (indicator vector), or Key (index)
1118511185
/// </summary>
11186-
public CategoricalTransformOutputKind OutputKind { get; set; } = CategoricalTransformOutputKind.Bag;
11186+
public OneHotEncodingTransformerOutputKind OutputKind { get; set; } = OneHotEncodingTransformerOutputKind.Bag;
1118711187

1118811188
/// <summary>
1118911189
/// Input dataset
@@ -11244,12 +11244,12 @@ public enum TermTransformSortOrder : byte
1124411244
}
1124511245

1124611246

11247-
public sealed partial class CategoricalTransformColumn : OneToOneColumn<CategoricalTransformColumn>, IOneToOneColumn
11247+
public sealed partial class OneHotEncodingTransformerColumn : OneToOneColumn<OneHotEncodingTransformerColumn>, IOneToOneColumn
1124811248
{
1124911249
/// <summary>
1125011250
/// Output kind: Bag (multi-set vector), Ind (indicator vector), Key (index), or Binary encoded indicator vector
1125111251
/// </summary>
11252-
public CategoricalTransformOutputKind? OutputKind { get; set; }
11252+
public OneHotEncodingTransformerOutputKind? OutputKind { get; set; }
1125311253

1125411254
/// <summary>
1125511255
/// Maximum number of terms to keep when auto-training
@@ -11316,28 +11316,28 @@ public CategoricalOneHotVectorizer(params (string inputColumn, string outputColu
1131611316

1131711317
public void AddColumn(string inputColumn)
1131811318
{
11319-
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.CategoricalTransformColumn>() : new List<Microsoft.ML.Legacy.Transforms.CategoricalTransformColumn>(Column);
11320-
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.CategoricalTransformColumn>.Create(inputColumn));
11319+
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.OneHotEncodingTransformerColumn>() : new List<Microsoft.ML.Legacy.Transforms.OneHotEncodingTransformerColumn>(Column);
11320+
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.OneHotEncodingTransformerColumn>.Create(inputColumn));
1132111321
Column = list.ToArray();
1132211322
}
1132311323

1132411324
public void AddColumn(string outputColumn, string inputColumn)
1132511325
{
11326-
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.CategoricalTransformColumn>() : new List<Microsoft.ML.Legacy.Transforms.CategoricalTransformColumn>(Column);
11327-
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.CategoricalTransformColumn>.Create(outputColumn, inputColumn));
11326+
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.OneHotEncodingTransformerColumn>() : new List<Microsoft.ML.Legacy.Transforms.OneHotEncodingTransformerColumn>(Column);
11327+
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.OneHotEncodingTransformerColumn>.Create(outputColumn, inputColumn));
1132811328
Column = list.ToArray();
1132911329
}
1133011330

1133111331

1133211332
/// <summary>
1133311333
/// New column definition(s) (optional form: name:src)
1133411334
/// </summary>
11335-
public CategoricalTransformColumn[] Column { get; set; }
11335+
public OneHotEncodingTransformerColumn[] Column { get; set; }
1133611336

1133711337
/// <summary>
1133811338
/// Output kind: Bag (multi-set vector), Ind (indicator vector), or Key (index)
1133911339
/// </summary>
11340-
public CategoricalTransformOutputKind OutputKind { get; set; } = CategoricalTransformOutputKind.Ind;
11340+
public OneHotEncodingTransformerOutputKind OutputKind { get; set; } = OneHotEncodingTransformerOutputKind.Ind;
1134111341

1134211342
/// <summary>
1134311343
/// Maximum number of terms to keep per column when auto-training
@@ -14938,7 +14938,7 @@ public MissingValueHandlerPipelineStep(Output output)
1493814938
namespace Legacy.Transforms
1493914939
{
1494014940

14941-
public sealed partial class NAIndicatorTransformColumn : OneToOneColumn<NAIndicatorTransformColumn>, IOneToOneColumn
14941+
public sealed partial class MissingValueIndicatorTransformerColumn : OneToOneColumn<MissingValueIndicatorTransformerColumn>, IOneToOneColumn
1494214942
{
1494314943
/// <summary>
1494414944
/// Name of the new column
@@ -14985,23 +14985,23 @@ public MissingValueIndicator(params (string inputColumn, string outputColumn)[]
1498514985

1498614986
public void AddColumn(string inputColumn)
1498714987
{
14988-
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.NAIndicatorTransformColumn>() : new List<Microsoft.ML.Legacy.Transforms.NAIndicatorTransformColumn>(Column);
14989-
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.NAIndicatorTransformColumn>.Create(inputColumn));
14988+
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.MissingValueIndicatorTransformerColumn>() : new List<Microsoft.ML.Legacy.Transforms.MissingValueIndicatorTransformerColumn>(Column);
14989+
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.MissingValueIndicatorTransformerColumn>.Create(inputColumn));
1499014990
Column = list.ToArray();
1499114991
}
1499214992

1499314993
public void AddColumn(string outputColumn, string inputColumn)
1499414994
{
14995-
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.NAIndicatorTransformColumn>() : new List<Microsoft.ML.Legacy.Transforms.NAIndicatorTransformColumn>(Column);
14996-
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.NAIndicatorTransformColumn>.Create(outputColumn, inputColumn));
14995+
var list = Column == null ? new List<Microsoft.ML.Legacy.Transforms.MissingValueIndicatorTransformerColumn>() : new List<Microsoft.ML.Legacy.Transforms.MissingValueIndicatorTransformerColumn>(Column);
14996+
list.Add(OneToOneColumn<Microsoft.ML.Legacy.Transforms.MissingValueIndicatorTransformerColumn>.Create(outputColumn, inputColumn));
1499714997
Column = list.ToArray();
1499814998
}
1499914999

1500015000

1500115001
/// <summary>
1500215002
/// New column definition(s) (optional form: name:src)
1500315003
/// </summary>
15004-
public NAIndicatorTransformColumn[] Column { get; set; }
15004+
public MissingValueIndicatorTransformerColumn[] Column { get; set; }
1500515005

1500615006
/// <summary>
1500715007
/// Input dataset

‎src/Microsoft.ML.PipelineInference/TransformInference.cs

Lines changed: 5 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -477,7 +477,7 @@ public override IEnumerable<SuggestedTransform> Apply(IntermediateColumn[] colum
477477
{
478478
Column = new[]
479479
{
480-
new ML.Legacy.Transforms.CategoricalHashTransformColumn
480+
new ML.Legacy.Transforms.OneHotHashEncodingTransformerColumn
481481
{
482482
Name = dest,
483483
Source = source
@@ -590,8 +590,8 @@ public override IEnumerable<SuggestedTransform> Apply(IntermediateColumn[] colum
590590
bool foundCatHash = false;
591591
var colSpecCat = new StringBuilder();
592592
var colSpecCatHash = new StringBuilder();
593-
var catColumns = new List<ML.Legacy.Transforms.CategoricalTransformColumn>();
594-
var catHashColumns = new List<ML.Legacy.Transforms.CategoricalHashTransformColumn>();
593+
var catColumns = new List<ML.Legacy.Transforms.OneHotEncodingTransformerColumn>();
594+
var catHashColumns = new List<ML.Legacy.Transforms.OneHotHashEncodingTransformerColumn>();
595595
var featureCols = new List<string>();
596596

597597
foreach (var column in columns)
@@ -622,7 +622,7 @@ public override IEnumerable<SuggestedTransform> Apply(IntermediateColumn[] colum
622622
{
623623
foundCat = true;
624624
colSpecCat.Append(columnArgument);
625-
catColumns.Add(new ML.Legacy.Transforms.CategoricalTransformColumn
625+
catColumns.Add(new ML.Legacy.Transforms.OneHotEncodingTransformerColumn
626626
{
627627
Name = columnNameQuoted.ToString(),
628628
Source = columnNameQuoted.ToString()
@@ -633,7 +633,7 @@ public override IEnumerable<SuggestedTransform> Apply(IntermediateColumn[] colum
633633
ch.Info("Categorical column '{0}' has extremely high cardinality. Suggested hash-based category encoding.", column.ColumnName);
634634
foundCatHash = true;
635635
colSpecCatHash.Append(columnArgument);
636-
catHashColumns.Add(new ML.Legacy.Transforms.CategoricalHashTransformColumn
636+
catHashColumns.Add(new ML.Legacy.Transforms.OneHotHashEncodingTransformerColumn
637637
{
638638
Name = columnNameQuoted.ToString(),
639639
Source = columnNameQuoted.ToString()

‎src/Microsoft.ML.Transforms/BootstrapSampleTransform.cs renamed to ‎src/Microsoft.ML.Transforms/BootstrapSampleTransformer.cs

Lines changed: 16 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -11,11 +11,11 @@
1111
using Microsoft.ML.Transforms;
1212
using System;
1313

14-
[assembly: LoadableClass(BootstrapSampleTransform.Summary, typeof(BootstrapSampleTransform), typeof(BootstrapSampleTransform.Arguments), typeof(SignatureDataTransform),
15-
BootstrapSampleTransform.UserName, "BootstrapSampleTransform", "BootstrapSample")]
14+
[assembly: LoadableClass(BootstrapSampleTransformer.Summary, typeof(BootstrapSampleTransformer), typeof(BootstrapSampleTransformer.Arguments), typeof(SignatureDataTransform),
15+
BootstrapSampleTransformer.UserName, "BootstrapSampleTransform", "BootstrapSample")]
1616

17-
[assembly: LoadableClass(BootstrapSampleTransform.Summary, typeof(BootstrapSampleTransform), null, typeof(SignatureLoadDataTransform),
18-
BootstrapSampleTransform.UserName, BootstrapSampleTransform.LoaderSignature)]
17+
[assembly: LoadableClass(BootstrapSampleTransformer.Summary, typeof(BootstrapSampleTransformer), null, typeof(SignatureLoadDataTransform),
18+
BootstrapSampleTransformer.UserName, BootstrapSampleTransformer.LoaderSignature)]
1919

2020
[assembly: EntryPointModule(typeof(BootstrapSample))]
2121

@@ -24,7 +24,7 @@ namespace Microsoft.ML.Transforms
2424
/// <summary>
2525
/// This class approximates bootstrap sampling of a dataview.
2626
/// </summary>
27-
public sealed class BootstrapSampleTransform : FilterBase
27+
public sealed class BootstrapSampleTransformer : FilterBase
2828
{
2929
private static class Defaults
3030
{
@@ -61,7 +61,7 @@ private static VersionInfo GetVersionInfo()
6161
verReadableCur: 0x00010001,
6262
verWeCanReadBack: 0x00010001,
6363
loaderSignature: LoaderSignature,
64-
loaderAssemblyName: typeof(BootstrapSampleTransform).Assembly.FullName);
64+
loaderAssemblyName: typeof(BootstrapSampleTransformer).Assembly.FullName);
6565
}
6666

6767
internal const string RegistrationName = "BootstrapSample";
@@ -73,7 +73,7 @@ private static VersionInfo GetVersionInfo()
7373
private readonly bool _shuffleInput;
7474
private readonly int _poolSize;
7575

76-
public BootstrapSampleTransform(IHostEnvironment env, Arguments args, IDataView input)
76+
public BootstrapSampleTransformer(IHostEnvironment env, Arguments args, IDataView input)
7777
: base(env, RegistrationName, input)
7878
{
7979
Host.CheckValue(args, nameof(args));
@@ -94,7 +94,7 @@ public BootstrapSampleTransform(IHostEnvironment env, Arguments args, IDataView
9494
/// <param name="seed">The random seed. If unspecified random state will be instead derived from the environment.</param>
9595
/// <param name="shuffleInput">Whether we should attempt to shuffle the source data. By default on, but can be turned off for efficiency.</param>
9696
/// <param name="poolSize">When shuffling the output, the number of output rows to keep in that pool. Note that shuffling of output is completely distinct from shuffling of input.</param>
97-
public BootstrapSampleTransform(IHostEnvironment env,
97+
public BootstrapSampleTransformer(IHostEnvironment env,
9898
IDataView input,
9999
bool complement = Defaults.Complement,
100100
uint? seed = null,
@@ -104,7 +104,7 @@ public BootstrapSampleTransform(IHostEnvironment env,
104104
{
105105
}
106106

107-
private BootstrapSampleTransform(IHost host, ModelLoadContext ctx, IDataView input)
107+
private BootstrapSampleTransformer(IHost host, ModelLoadContext ctx, IDataView input)
108108
: base(host, input)
109109
{
110110
host.AssertValue(ctx);
@@ -148,14 +148,14 @@ public override void Save(ModelSaveContext ctx)
148148
ctx.Writer.Write(_poolSize);
149149
}
150150

151-
public static BootstrapSampleTransform Create(IHostEnvironment env, ModelLoadContext ctx, IDataView input)
151+
public static BootstrapSampleTransformer Create(IHostEnvironment env, ModelLoadContext ctx, IDataView input)
152152
{
153153
Contracts.CheckValue(env, nameof(env));
154154
var h = env.Register(RegistrationName);
155155
h.CheckValue(ctx, nameof(ctx));
156156
h.CheckValue(input, nameof(input));
157157
ctx.CheckAtModel(GetVersionInfo());
158-
return h.Apply("Loading Model", ch => new BootstrapSampleTransform(h, ctx, input));
158+
return h.Apply("Loading Model", ch => new BootstrapSampleTransformer(h, ctx, input));
159159
}
160160

161161
protected override bool? ShouldUseParallelCursors(Func<int, bool> predicate)
@@ -184,14 +184,14 @@ public override IRowCursor[] GetRowCursorSet(out IRowCursorConsolidator consolid
184184
private sealed class RowCursor : LinkedRootCursorBase<IRowCursor>, IRowCursor
185185
{
186186
private int _remaining;
187-
private readonly BootstrapSampleTransform _parent;
187+
private readonly BootstrapSampleTransformer _parent;
188188
private readonly IRandom _rgen;
189189

190190
public override long Batch { get { return 0; } }
191191

192192
public Schema Schema { get { return Input.Schema; } }
193193

194-
public RowCursor(BootstrapSampleTransform parent, IRowCursor input, IRandom rgen)
194+
public RowCursor(BootstrapSampleTransformer parent, IRowCursor input, IRandom rgen)
195195
: base(parent.Host, input)
196196
{
197197
Ch.AssertValue(rgen);
@@ -237,14 +237,14 @@ protected override bool MoveNextCore()
237237

238238
public static class BootstrapSample
239239
{
240-
[TlcModule.EntryPoint(Name = "Transforms.ApproximateBootstrapSampler", Desc = BootstrapSampleTransform.Summary, UserName = BootstrapSampleTransform.UserName, ShortName = BootstrapSampleTransform.RegistrationName)]
241-
public static CommonOutputs.TransformOutput GetSample(IHostEnvironment env, BootstrapSampleTransform.Arguments input)
240+
[TlcModule.EntryPoint(Name = "Transforms.ApproximateBootstrapSampler", Desc = BootstrapSampleTransformer.Summary, UserName = BootstrapSampleTransformer.UserName, ShortName = BootstrapSampleTransformer.RegistrationName)]
241+
public static CommonOutputs.TransformOutput GetSample(IHostEnvironment env, BootstrapSampleTransformer.Arguments input)
242242
{
243243
Contracts.CheckValue(env, nameof(env));
244244
env.CheckValue(input, nameof(input));
245245

246246
var h = EntryPointUtils.CheckArgsAndCreateHost(env, "BootstrapSample", input);
247-
var view = new BootstrapSampleTransform(h, input, input.Data);
247+
var view = new BootstrapSampleTransformer(h, input, input.Data);
248248
return new CommonOutputs.TransformOutput()
249249
{
250250
Model = new TransformModel(h, view, input.Data),

0 commit comments

Comments
 (0)