Skip to content

Metadata fixes for the ValueMappingEstimator #2098

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 6 commits into from
Jan 15, 2019
Merged

Metadata fixes for the ValueMappingEstimator #2098

Show file tree
Hide file tree
Changes from 4 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
7aeb5f7
The ValueMappingEstimator had a couple of issues when using setting the
singlis Jan 9, 2019
904975a
-Fixing runs for release build. Code that needed to run was wrapped i…
singlis Jan 10, 2019
3b3f6da
Merge remote-tracking branch 'origin/master' into singlis/valuemap-fix
singlis Jan 11, 2019
c4729ff
- Updating changes based upon feedback.
singlis Jan 11, 2019
5759218
Merge remote-tracking branch 'origin/master' into singlis/valuemap-fix
singlis Jan 14, 2019
fa27d84
- Updating from feedback
singlis Jan 15, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
15 changes: 6 additions & 9 deletions src/Microsoft.ML.Data/Transforms/ValueMappingTransformer.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -100,13 +100,14 @@ public override SchemaShape GetOutputSchema(SchemaShape inputSchema)
var isKey = Transformer.ValueColumnType.IsKey;
var columnType = (isKey) ? PrimitiveType.FromKind(DataKind.U4) :
Transformer.ValueColumnType;
var metadata = SchemaShape.Create(Transformer.ValueColumnMetadata.Schema);
Copy link
Member

@wschin wschin Jan 11, 2019 •
edited by singlis
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
var metadata = SchemaShape.Create(Transformer.ValueColumnMetadata.Schema);
var metadataShape = SchemaShape.Create(Transformer.ValueColumnMetadata.Schema);

Maybe? #Resolved

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure, I will update.

In reply to: 247282604 [](ancestors = 247282604)

foreach (var (Input, Output) in _columns)
{
if (!inputSchema.TryFindColumn(Input, out var originalColumn))
throw Host.ExceptSchemaMismatch(nameof(inputSchema), "input", Input);

// Get the type from TOutputType
Copy link
Member

@wschin wschin Jan 11, 2019 •
edited by singlis
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What is TOutputType here? Is it not Output? #Resolved

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That is old, I used to have the generic arguments be TInputType, TOutputType, but then changed to TKeyType and TValueType. I will update the comment.

In reply to: 247282739 [](ancestors = 247282739)

var col = new SchemaShape.Column(Output, vectorKind, columnType, isKey, originalColumn.Metadata);
var col = new SchemaShape.Column(Output, vectorKind, columnType, isKey, metadata);
Copy link
Member

@wschin wschin Jan 11, 2019 •
edited by singlis
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
var col = new SchemaShape.Column(Output, vectorKind, columnType, isKey, metadata);
var col = new SchemaShape.Column(OutputColumnName, vectorKind, columnType, isKey, metadata);
``` #WontFix

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Although Name is more specific to what the object actually is, I think it goes against thinking of these as columns since you are talking about a input/output (or source/name). I would prefer to keep it as Output/Input. BTW - this will also go through another revision with #2064

In reply to: 247282892 [](ancestors = 247282892)

resultDic[Output] = col;
}
return new SchemaShape(resultDic.Values);
Expand Down Expand Up @@ -191,18 +192,14 @@ internal static IDataView CreateDataView<TKey, TValue>(IHostEnvironment env,
// set of values. This is used for generating the metadata of
// the column.
HashSet<TValue> valueSet = new HashSet<TValue>();
HashSet<TKey> keySet = new HashSet<TKey>();
for (int i = 0; i < values.Count(); ++i)
foreach(var v in values)
{
var v = values.ElementAt(i);
if (valueSet.Contains(v))
continue;
valueSet.Add(v);

var k = keys.ElementAt(i);
keySet.Add(k);
}
var metaKeys = keySet.ToArray();

var metaKeys = valueSet.ToArray();

// Key Values are treated in one of two ways:
// If the values are of type uint or ulong, these values are used directly as the keys types and no new keys are created.
Expand Down Expand Up @@ -387,7 +384,7 @@ protected ValueMappingTransformer(IHostEnvironment env, IDataView lookupMap,
Host.CheckNonEmpty(valueColumn, nameof(valueColumn), "A value column must be specified when passing in an IDataView for the value mapping");
_valueMap = CreateValueMapFromDataView(lookupMap, keyColumn, valueColumn);
int valueColumnIdx = 0;
Host.Assert(lookupMap.Schema.TryGetColumnIndex(valueColumn, out valueColumnIdx));
Host.Check(lookupMap.Schema.TryGetColumnIndex(valueColumn, out valueColumnIdx));
_valueMetadata = lookupMap.Schema[valueColumnIdx].Metadata;

// Create the byte array of the original IDataView, this is used for saving out the data.
Expand Down
98 changes: 62 additions & 36 deletions test/Microsoft.ML.Tests/Transformers/ValueMappingTests.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,17 +52,17 @@ public void ValueMapOneValueTest()
var data = new[] { new TestClass() { A = "bar", B = "test", C = "foo" } };
var dataView = ComponentCreation.CreateDataView(Env, data);

IEnumerable<ReadOnlyMemory<char>> keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
IEnumerable<int> values = new List<int>() { 1, 2, 3, 4 };
var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
var values = new List<int>() { 1, 2, 3, 4 };

var estimator = new ValueMappingEstimator<ReadOnlyMemory<char>, int>(Env, keys, values, new[] { ("A", "D"), ("B", "E"), ("C", "F") });
var t = estimator.Fit(dataView);

var result = t.Transform(dataView);
var cursor = result.GetRowCursor((col) => true);
var getterD = cursor.GetGetter<int>(3);
var getterE = cursor.GetGetter<int>(4);
var getterF = cursor.GetGetter<int>(5);
var getterD = cursor.GetGetter<int>(result.Schema["D"].Index);
var getterE = cursor.GetGetter<int>(result.Schema["E"].Index);
var getterF = cursor.GetGetter<int>(result.Schema["F"].Index);
cursor.MoveNext();

int dValue = 0;
Expand Down Expand Up @@ -93,9 +93,9 @@ public void ValueMapVectorValueTest()

var result = t.Transform(dataView);
var cursor = result.GetRowCursor((col) => true);
var getterD = cursor.GetGetter<VBuffer<int>>(3);
var getterE = cursor.GetGetter<VBuffer<int>>(4);
var getterF = cursor.GetGetter<VBuffer<int>>(5);
var getterD = cursor.GetGetter<VBuffer<int>>(result.Schema["D"].Index);
var getterE = cursor.GetGetter<VBuffer<int>>(result.Schema["E"].Index);
var getterF = cursor.GetGetter<VBuffer<int>>(result.Schema["F"].Index);
cursor.MoveNext();

var valuesArray = values.ToArray();
Expand All @@ -116,17 +116,17 @@ public void ValueMappingMissingKey()
var data = new[] { new TestClass() { A = "barTest", B = "test", C = "foo" } };
var dataView = ComponentCreation.CreateDataView(Env, data);

IEnumerable<ReadOnlyMemory<char>> keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
IEnumerable<int> values = new List<int>() { 1, 2, 3, 4 };
var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
var values = new List<int>() { 1, 2, 3, 4 };

var estimator = new ValueMappingEstimator<ReadOnlyMemory<char>, int>(Env, keys, values, new[] { ("A", "D"), ("B", "E"), ("C", "F") });
var t = estimator.Fit(dataView);

var result = t.Transform(dataView);
var cursor = result.GetRowCursor((col) => true);
var getterD = cursor.GetGetter<int>(3);
var getterE = cursor.GetGetter<int>(4);
var getterF = cursor.GetGetter<int>(5);
var getterD = cursor.GetGetter<int>(result.Schema["D"].Index);
var getterE = cursor.GetGetter<int>(result.Schema["E"].Index);
var getterF = cursor.GetGetter<int>(result.Schema["F"].Index);
cursor.MoveNext();

int dValue = 1;
Expand All @@ -146,8 +146,8 @@ void TestDuplicateKeys()
var data = new[] { new TestClass() { A = "barTest", B = "test", C = "foo" } };
var dataView = ComponentCreation.CreateDataView(Env, data);

IEnumerable<ReadOnlyMemory<char>> keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "foo".AsMemory() };
IEnumerable<int> values = new List<int>() { 1, 2 };
var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "foo".AsMemory() };
var values = new List<int>() { 1, 2 };

Assert.Throws<InvalidOperationException>(() => new ValueMappingEstimator<ReadOnlyMemory<char>, int>(Env, keys, values, new[] { ("A", "D"), ("B", "E"), ("C", "F") }));
}
Expand All @@ -158,8 +158,8 @@ public void ValueMappingOutputSchema()
var data = new[] { new TestClass() { A = "barTest", B = "test", C = "foo" } };
var dataView = ComponentCreation.CreateDataView(Env, data);

IEnumerable<ReadOnlyMemory<char>> keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
IEnumerable<int> values = new List<int>() { 1, 2, 3, 4 };
var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
var values = new List<int>() { 1, 2, 3, 4 };

var estimator = new ValueMappingEstimator<ReadOnlyMemory<char>, int>(Env, keys, values, new[] { ("A", "D"), ("B", "E"), ("C", "F") });
var outputSchema = estimator.GetOutputSchema(SchemaShape.Create(dataView.Schema));
Expand All @@ -184,8 +184,8 @@ public void ValueMappingWithValuesAsKeyTypesOutputSchema()
var data = new[] { new TestClass() { A = "bar", B = "test", C = "foo" } };
var dataView = ComponentCreation.CreateDataView(Env, data);

IEnumerable<ReadOnlyMemory<char>> keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
IEnumerable<ReadOnlyMemory<char>> values = new List<ReadOnlyMemory<char>>() { "t".AsMemory(), "s".AsMemory(), "u".AsMemory(), "v".AsMemory() };
var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
var values = new List<ReadOnlyMemory<char>>() { "t".AsMemory(), "s".AsMemory(), "u".AsMemory(), "v".AsMemory() };

var estimator = new ValueMappingEstimator<ReadOnlyMemory<char>, ReadOnlyMemory<char>>(Env, keys, values, true, new[] { ("A", "D"), ("B", "E"), ("C", "F") });
var outputSchema = estimator.GetOutputSchema(SchemaShape.Create(dataView.Schema));
Expand All @@ -212,20 +212,20 @@ public void ValueMappingValuesAsUintKeyTypes()
var data = new[] { new TestClass() { A = "bar", B = "test2", C = "wahoo" } };
var dataView = ComponentCreation.CreateDataView(Env, data);

IEnumerable<ReadOnlyMemory<char>> keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };

// These are the expected key type values
IEnumerable<uint> values = new List<uint>() { 51, 25, 42, 61 };
var values = new List<uint>() { 51, 25, 42, 61 };

var estimator = new ValueMappingEstimator<ReadOnlyMemory<char>, uint>(Env, keys, values, true, new[] { ("A", "D"), ("B", "E"), ("C", "F") });

var t = estimator.Fit(dataView);

var result = t.Transform(dataView);
var cursor = result.GetRowCursor((col) => true);
var getterD = cursor.GetGetter<uint>(3);
var getterE = cursor.GetGetter<uint>(4);
var getterF = cursor.GetGetter<uint>(5);
var getterD = cursor.GetGetter<uint>(result.Schema["D"].Index);
var getterE = cursor.GetGetter<uint>(result.Schema["E"].Index);
var getterF = cursor.GetGetter<uint>(result.Schema["F"].Index);
cursor.MoveNext();

// The expected values will contain the actual uints and are not generated.
Expand All @@ -251,20 +251,20 @@ public void ValueMappingValuesAsUlongKeyTypes()
var data = new[] { new TestClass() { A = "bar", B = "test2", C = "wahoo" } };
var dataView = ComponentCreation.CreateDataView(Env, data);

IEnumerable<ReadOnlyMemory<char>> keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };

// These are the expected key type values
IEnumerable<ulong> values = new List<ulong>() { 51, Int32.MaxValue, 42, 61 };
var values = new List<ulong>() { 51, Int32.MaxValue, 42, 61 };

var estimator = new ValueMappingEstimator<ReadOnlyMemory<char>, ulong>(Env, keys, values, true, new[] { ("A", "D"), ("B", "E"), ("C", "F") });

var t = estimator.Fit(dataView);

var result = t.Transform(dataView);
var cursor = result.GetRowCursor((col) => true);
var getterD = cursor.GetGetter<ulong>(3);
var getterE = cursor.GetGetter<ulong>(4);
var getterF = cursor.GetGetter<ulong>(5);
var getterD = cursor.GetGetter<ulong>(result.Schema["D"].Index);
var getterE = cursor.GetGetter<ulong>(result.Schema["E"].Index);
var getterF = cursor.GetGetter<ulong>(result.Schema["F"].Index);
cursor.MoveNext();

// The expected values will contain the actual uints and are not generated.
Expand All @@ -289,19 +289,19 @@ public void ValueMappingValuesAsStringKeyTypes()
var data = new[] { new TestClass() { A = "bar", B = "test", C = "notfound" } };
var dataView = ComponentCreation.CreateDataView(Env, data);

IEnumerable<ReadOnlyMemory<char>> keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };

// Generating the list of strings for the key type values, note that foo1 is duplicated as intended to test that the same index value is returned
IEnumerable<ReadOnlyMemory<char>> values = new List<ReadOnlyMemory<char>>() { "foo1".AsMemory(), "foo2".AsMemory(), "foo1".AsMemory(), "foo3".AsMemory() };
var values = new List<ReadOnlyMemory<char>>() { "foo1".AsMemory(), "foo2".AsMemory(), "foo1".AsMemory(), "foo3".AsMemory() };

var estimator = new ValueMappingEstimator<ReadOnlyMemory<char>, ReadOnlyMemory<char>>(Env, keys, values, true, new[] { ("A", "D"), ("B", "E"), ("C", "F") });
var t = estimator.Fit(dataView);

var result = t.Transform(dataView);
var cursor = result.GetRowCursor((col) => true);
var getterD = cursor.GetGetter<uint>(3);
var getterE = cursor.GetGetter<uint>(4);
var getterF = cursor.GetGetter<uint>(5);
var getterD = cursor.GetGetter<uint>(result.Schema["D"].Index);
var getterE = cursor.GetGetter<uint>(result.Schema["E"].Index);
var getterF = cursor.GetGetter<uint>(result.Schema["F"].Index);
cursor.MoveNext();

// The expected values will contain the generated key type values starting from 1.
Expand All @@ -320,6 +320,32 @@ public void ValueMappingValuesAsStringKeyTypes()
Assert.Equal<uint>(0, fValue);
}

[Fact]
public void ValueMappingValuesAsKeyTypesReverseLookup()
{
var data = new[] { new TestClass() { A = "bar", B = "test", C = "notfound" } };
var dataView = ComponentCreation.CreateDataView(Env, data);

var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };

// Generating the list of strings for the key type values, note that foo1 is duplicated as intended to test that the same index value is returned
var values = new List<ReadOnlyMemory<char>>() { "foo1".AsMemory(), "foo2".AsMemory(), "foo1".AsMemory(), "foo3".AsMemory() };

var estimator = new ValueMappingEstimator<ReadOnlyMemory<char>, ReadOnlyMemory<char>>(Env, keys, values, true, new[] { ("A", "D"), ("B", "E"), ("C", "F") })
Copy link
Member

@wschin wschin Jan 11, 2019 •
edited by singlis
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why do you need three columns? Only "D" is used below. #Resolved

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

True - that is a result of copying code from another test. I have removed the other two columns.

In reply to: 247285144 [](ancestors = 247285144)

.Append(new KeyToValueMappingEstimator(Env, ("D","DOutput")));
Copy link
Member

@wschin wschin Jan 11, 2019 •
edited by singlis
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
.Append(new KeyToValueMappingEstimator(Env, ("D","DOutput")));
.Append(new KeyToValueMappingEstimator(Env, ("D", "DOutput")));

Please format all files touched. #Resolved

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, they are both formatted now.

In reply to: 247284278 [](ancestors = 247284278)

var t = estimator.Fit(dataView);

var result = t.Transform(dataView);
var cursor = result.GetRowCursor((col) => true);
var getterD = cursor.GetGetter<ReadOnlyMemory<char>>(result.Schema["DOutput"].Index);
cursor.MoveNext();

// The expected values will contain the generated key type values starting from 1.
ReadOnlyMemory<char> dValue = default;
getterD(ref dValue);
Copy link
Contributor

@Ivanidzo4ka Ivanidzo4ka Jan 11, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

dValue [](start = 24, length = 6)

can we validate it's a "bar"? #Closed

Assert.Equal("foo2".AsMemory(), dValue);
}

[Fact]
public void ValueMappingWorkout()
{
Expand All @@ -328,8 +354,8 @@ public void ValueMappingWorkout()
var badData = new[] { new TestWrong() { A = "bar", B = 1.2f } };
var badDataView = ComponentCreation.CreateDataView(Env, badData);

IEnumerable<ReadOnlyMemory<char>> keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
IEnumerable<int> values = new List<int>() { 1, 2, 3, 4 };
var keys = new List<ReadOnlyMemory<char>>() { "foo".AsMemory(), "bar".AsMemory(), "test".AsMemory(), "wahoo".AsMemory() };
var values = new List<int>() { 1, 2, 3, 4 };

// Workout on value mapping
var est = ML.Transforms.Conversion.ValueMap(keys, values, new[] { ("A", "D"), ("B", "E"), ("C", "F") });
Expand Down