private int[][] CompileSlotMap(string slotMapString, int srcSlotCount)
{
var parts = ReadOnlyMemoryUtils.Split(slotMapString.AsMemory(), new[] { ';' }).ToArray();
var slotMap = new int[parts.Length][];
for (int i = 0; i < slotMap.Length; i++)
{
var slotIndices = ReadOnlyMemoryUtils.Split(parts[i], new[] { ',' }).ToArray();
var slots = new int[slotIndices.Length];
slotMap[i] = slots;
for (int j = 0; j < slots.Length; j++)
{
int index;
if (!int.TryParse(slotIndices[j].ToString(), out index) || index < 0 || index >= srcSlotCount)
{
throw Host.Except("Unexpected slot index '{1}' in group {0}. Expected 0 to {2}", i, slotIndices[j], srcSlotCount - 1);
}
slots[j] = index;
}
if (slots.Distinct().Count() < slots.Length)
{
throw Host.Except("Group '{0}' has duplicate slot indices", parts[i]);
}
}
return(slotMap);
}
public static CommonOutputs.TransformOutput RenameBinaryPredictionScoreColumns(IHostEnvironment env,
RenameBinaryPredictionScoreColumnsInput input)
{
Contracts.CheckValue(env, nameof(env));
var host = env.Register("ScoreModel");
host.CheckValue(input, nameof(input));
EntryPointUtils.CheckInputArgs(host, input);
if (input.PredictorModel.Predictor.PredictionKind == PredictionKind.BinaryClassification)
{
DataViewType labelType;
var labelNames = input.PredictorModel.GetLabelInfo(host, out labelType);
if (labelNames != null && labelNames.Length == 2)
{
var positiveClass = labelNames[1];
// Rename all the score columns.
int colMax;
var maxScoreId = input.Data.Schema.GetMaxAnnotationKind(out colMax, AnnotationUtils.Kinds.ScoreColumnSetId);
var copyCols = new List <(string name, string source)>();
for (int i = 0; i < input.Data.Schema.Count; i++)
{
if (input.Data.Schema[i].IsHidden)
{
continue;
}
if (!ShouldAddColumn(input.Data.Schema, i, null, maxScoreId))
{
continue;
}
// Do not rename the PredictedLabel column.
ReadOnlyMemory <char> tmp = default;
if (input.Data.Schema.TryGetAnnotation(TextDataViewType.Instance, AnnotationUtils.Kinds.ScoreValueKind, i,
ref tmp) &&
ReadOnlyMemoryUtils.EqualsStr(AnnotationUtils.Const.ScoreValueKind.PredictedLabel, tmp))
{
continue;
}
var source = input.Data.Schema[i].Name;
var name = source + "." + positiveClass;
copyCols.Add((name, source));
}
var copyColumn = new ColumnCopyingTransformer(env, copyCols.ToArray()).Transform(input.Data);
var dropColumn = ColumnSelectingTransformer.CreateDrop(env, copyColumn, copyCols.Select(c => c.source).ToArray());
return(new CommonOutputs.TransformOutput {
Model = new TransformModelImpl(env, dropColumn, input.Data), OutputData = dropColumn
});
}
}
var newView = NopTransform.CreateIfNeeded(env, input.Data);
return(new CommonOutputs.TransformOutput {
Model = new TransformModelImpl(env, newView, input.Data), OutputData = newView
});
}
private void Evaluate(IndentedTextWriter wrt, Delegate del, DataViewType typeRes, DataViewType[] types,
string text, int ichMin, int ichLim)
{
Contracts.AssertValue(del);
Contracts.AssertNonEmpty(types);
var args = new object[types.Length];
var getters = new Func <ReadOnlyMemory <char>, bool> [types.Length];
for (int i = 0; i < getters.Length; i++)
{
getters[i] = GetGetter(i, types[i], args);
}
StringBuilder sb = new StringBuilder();
Action <object> printer = GetPrinter(typeRes, sb);
ReadOnlyMemory <char> chars = text.AsMemory().Slice(ichMin, ichLim - ichMin);
for (bool more = true; more;)
{
ReadOnlyMemory <char> line;
if (RuntimeInformation.IsOSPlatform(OSPlatform.Windows))
{
more = ReadOnlyMemoryUtils.SplitOne(chars, '\x0D', out line, out chars);
}
else
{
more = ReadOnlyMemoryUtils.SplitOne(chars, '\x0A', out line, out chars);
}
line = ReadOnlyMemoryUtils.TrimWhiteSpace(line);
if (line.IsEmpty)
{
continue;
}
// Note this "hack" to map _ to empty. It's easier than fully handling quoting and is sufficient
// for these tests.
var vals = ReadOnlyMemoryUtils.Split(line, new char[] { ',' })
.Select(x => ReadOnlyMemoryUtils.TrimWhiteSpace(x))
.Select(x => ReadOnlyMemoryUtils.EqualsStr("_", x) ? ReadOnlyMemory <char> .Empty : x)
.ToArray();
Contracts.Assert(vals.Length == getters.Length);
for (int i = 0; i < getters.Length; i++)
{
if (!getters[i](vals[i]))
{
wrt.Write("*** Parsing {0} Failed *** ", vals[i]);
}
}
var res = del.DynamicInvoke(args);
printer(res);
wrt.WriteLine(sb);
}
}
public static TX Lower(TX a)
{
if (a.IsEmpty)
{
return(a);
}
var sb = new StringBuilder();
ReadOnlyMemoryUtils.AddLowerCaseToStringBuilder(a.Span, sb);
return(sb.ToString().AsMemory());
}
protected override Delegate GetGetterCore(IChannel ch, IRow input, int iinfo, out Action disposer)
{
Host.AssertValueOrNull(ch);
Host.AssertValue(input);
Host.Assert(0 <= iinfo && iinfo < Infos.Length);
Host.Assert(Infos[iinfo].TypeSrc.IsVector & Infos[iinfo].TypeSrc.ItemType.IsText);
disposer = null;
var ex = _exes[iinfo];
Language stopWordslang = ex.Lang;
var lang = default(ReadOnlyMemory <char>);
var getLang = ex.LangsColIndex >= 0 ? input.GetGetter <ReadOnlyMemory <char> >(ex.LangsColIndex) : null;
var getSrc = GetSrcGetter <VBuffer <ReadOnlyMemory <char> > >(input, iinfo);
var src = default(VBuffer <ReadOnlyMemory <char> >);
var buffer = new StringBuilder();
var list = new List <ReadOnlyMemory <char> >();
ValueGetter <VBuffer <ReadOnlyMemory <char> > > del =
(ref VBuffer <ReadOnlyMemory <char> > dst) =>
{
var langToUse = stopWordslang;
UpdateLanguage(ref langToUse, getLang, ref lang);
getSrc(ref src);
list.Clear();
var srcValues = src.GetValues();
for (int i = 0; i < srcValues.Length; i++)
{
if (srcValues[i].IsEmpty)
{
continue;
}
buffer.Clear();
ReadOnlyMemoryUtils.AddLowerCaseToStringBuilder(srcValues[i].Span, buffer);
// REVIEW nihejazi: Consider using a trie for string matching (Aho-Corasick, etc.)
if (StopWords[(int)langToUse].Get(buffer) == null)
{
list.Add(srcValues[i]);
}
}
VBufferUtils.Copy(list, ref dst, list.Count);
};
return(del);
}
public void TestCrossValidationMacroWithNonDefaultNames()
{
string dataPath = GetDataPath(@"adult.tiny.with-schema.txt");
var env = new MLContext(42);
var subGraph = env.CreateExperiment();
var textToKey = new Legacy.Transforms.TextToKeyConverter();
textToKey.Column = new[] { new Legacy.Transforms.ValueToKeyMappingTransformerColumn()
{
Name = "Label1", Source = "Label"
} };
var textToKeyOutput = subGraph.Add(textToKey);
var hash = new Legacy.Transforms.HashConverter();
hash.Column = new[] { new Legacy.Transforms.HashJoiningTransformColumn()
{
Name = "GroupId1", Source = "Workclass"
} };
hash.Data = textToKeyOutput.OutputData;
var hashOutput = subGraph.Add(hash);
var learnerInput = new Legacy.Trainers.FastTreeRanker
{
TrainingData = hashOutput.OutputData,
NumThreads = 1,
LabelColumn = "Label1",
GroupIdColumn = "GroupId1"
};
var learnerOutput = subGraph.Add(learnerInput);
var modelCombine = new Legacy.Transforms.ManyHeterogeneousModelCombiner
{
TransformModels = new ArrayVar <TransformModel>(textToKeyOutput.Model, hashOutput.Model),
PredictorModel = learnerOutput.PredictorModel
};
var modelCombineOutput = subGraph.Add(modelCombine);
var experiment = env.CreateExperiment();
var importInput = new Legacy.Data.TextLoader(dataPath);
importInput.Arguments.HasHeader = true;
importInput.Arguments.Column = new TextLoaderColumn[]
{
new TextLoaderColumn {
Name = "Label", Source = new[] { new TextLoaderRange(0) }
},
new TextLoaderColumn {
Name = "Workclass", Source = new[] { new TextLoaderRange(1) }, Type = Legacy.Data.DataKind.Text
},
new TextLoaderColumn {
Name = "Features", Source = new[] { new TextLoaderRange(9, 14) }
}
};
var importOutput = experiment.Add(importInput);
var crossValidate = new Legacy.Models.CrossValidator
{
Data = importOutput.Data,
Nodes = subGraph,
TransformModel = null,
LabelColumn = "Label1",
GroupColumn = "GroupId1",
NameColumn = "Workclass",
Kind = Legacy.Models.MacroUtilsTrainerKinds.SignatureRankerTrainer
};
crossValidate.Inputs.Data = textToKey.Data;
crossValidate.Outputs.PredictorModel = modelCombineOutput.PredictorModel;
var crossValidateOutput = experiment.Add(crossValidate);
experiment.Compile();
experiment.SetInput(importInput.InputFile, new SimpleFileHandle(env, dataPath, false, false));
experiment.Run();
var data = experiment.GetOutput(crossValidateOutput.OverallMetrics);
var schema = data.Schema;
var b = schema.TryGetColumnIndex("NDCG", out int metricCol);
Assert.True(b);
b = schema.TryGetColumnIndex("Fold Index", out int foldCol);
Assert.True(b);
using (var cursor = data.GetRowCursor(col => col == metricCol || col == foldCol))
{
var getter = cursor.GetGetter <VBuffer <double> >(metricCol);
var foldGetter = cursor.GetGetter <ReadOnlyMemory <char> >(foldCol);
ReadOnlyMemory <char> fold = default;
// Get the verage.
b = cursor.MoveNext();
Assert.True(b);
var avg = default(VBuffer <double>);
getter(ref avg);
foldGetter(ref fold);
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Average", fold));
// Get the standard deviation.
b = cursor.MoveNext();
Assert.True(b);
var stdev = default(VBuffer <double>);
getter(ref stdev);
foldGetter(ref fold);
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Standard Deviation", fold));
var stdevValues = stdev.GetValues();
Assert.Equal(2.462, stdevValues[0], 3);
Assert.Equal(2.763, stdevValues[1], 3);
Assert.Equal(3.273, stdevValues[2], 3);
var sumBldr = new BufferBuilder <double>(R8Adder.Instance);
sumBldr.Reset(avg.Length, true);
var val = default(VBuffer <double>);
for (int f = 0; f < 2; f++)
{
b = cursor.MoveNext();
Assert.True(b);
getter(ref val);
foldGetter(ref fold);
sumBldr.AddFeatures(0, in val);
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Fold " + f, fold));
}
var sum = default(VBuffer <double>);
sumBldr.GetResult(ref sum);
var avgValues = avg.GetValues();
var sumValues = sum.GetValues();
for (int i = 0; i < avgValues.Length; i++)
{
Assert.Equal(avgValues[i], sumValues[i] / 2);
}
b = cursor.MoveNext();
Assert.False(b);
}
data = experiment.GetOutput(crossValidateOutput.PerInstanceMetrics);
Assert.True(data.Schema.TryGetColumnIndex("Instance", out int nameCol));
using (var cursor = data.GetRowCursor(col => col == nameCol))
{
var getter = cursor.GetGetter <ReadOnlyMemory <char> >(nameCol);
while (cursor.MoveNext())
{
ReadOnlyMemory <char> name = default;
getter(ref name);
Assert.Subset(new HashSet <string>()
{
"Private", "?", "Federal-gov"
}, new HashSet <string>()
{
name.ToString()
});
if (cursor.Position > 4)
{
break;
}
}
}
}
public void TestCrossValidationMacroWithStratification()
{
var dataPath = GetDataPath(@"breast-cancer.txt");
var env = new MLContext(42);
var subGraph = env.CreateExperiment();
var nop = new Legacy.Transforms.NoOperation();
var nopOutput = subGraph.Add(nop);
var learnerInput = new Legacy.Trainers.StochasticDualCoordinateAscentBinaryClassifier
{
TrainingData = nopOutput.OutputData,
NumThreads = 1
};
var learnerOutput = subGraph.Add(learnerInput);
var modelCombine = new Legacy.Transforms.ManyHeterogeneousModelCombiner
{
TransformModels = new ArrayVar <TransformModel>(nopOutput.Model),
PredictorModel = learnerOutput.PredictorModel
};
var modelCombineOutput = subGraph.Add(modelCombine);
var experiment = env.CreateExperiment();
var importInput = new Legacy.Data.TextLoader(dataPath);
importInput.Arguments.Column = new Legacy.Data.TextLoaderColumn[]
{
new Legacy.Data.TextLoaderColumn {
Name = "Label", Source = new[] { new Legacy.Data.TextLoaderRange(0) }
},
new Legacy.Data.TextLoaderColumn {
Name = "Strat", Source = new[] { new Legacy.Data.TextLoaderRange(1) }
},
new Legacy.Data.TextLoaderColumn {
Name = "Features", Source = new[] { new Legacy.Data.TextLoaderRange(2, 9) }
}
};
var importOutput = experiment.Add(importInput);
var crossValidate = new Legacy.Models.CrossValidator
{
Data = importOutput.Data,
Nodes = subGraph,
TransformModel = null,
StratificationColumn = "Strat"
};
crossValidate.Inputs.Data = nop.Data;
crossValidate.Outputs.PredictorModel = modelCombineOutput.PredictorModel;
var crossValidateOutput = experiment.Add(crossValidate);
experiment.Compile();
experiment.SetInput(importInput.InputFile, new SimpleFileHandle(env, dataPath, false, false));
experiment.Run();
var data = experiment.GetOutput(crossValidateOutput.OverallMetrics);
var schema = data.Schema;
var b = schema.TryGetColumnIndex("AUC", out int metricCol);
Assert.True(b);
b = schema.TryGetColumnIndex("Fold Index", out int foldCol);
Assert.True(b);
using (var cursor = data.GetRowCursor(col => col == metricCol || col == foldCol))
{
var getter = cursor.GetGetter <double>(metricCol);
var foldGetter = cursor.GetGetter <ReadOnlyMemory <char> >(foldCol);
ReadOnlyMemory <char> fold = default;
// Get the verage.
b = cursor.MoveNext();
Assert.True(b);
double avg = 0;
getter(ref avg);
foldGetter(ref fold);
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Average", fold));
// Get the standard deviation.
b = cursor.MoveNext();
Assert.True(b);
double stdev = 0;
getter(ref stdev);
foldGetter(ref fold);
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Standard Deviation", fold));
Assert.Equal(0.00488, stdev, 5);
double sum = 0;
double val = 0;
for (int f = 0; f < 2; f++)
{
b = cursor.MoveNext();
Assert.True(b);
getter(ref val);
foldGetter(ref fold);
sum += val;
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Fold " + f, fold));
}
Assert.Equal(avg, sum / 2);
b = cursor.MoveNext();
Assert.False(b);
}
}
public void TestCrossValidationMacroWithMultiClass()
{
var dataPath = GetDataPath(@"Train-Tiny-28x28.txt");
var env = new MLContext(42);
var subGraph = env.CreateExperiment();
var nop = new Legacy.Transforms.NoOperation();
var nopOutput = subGraph.Add(nop);
var learnerInput = new Legacy.Trainers.StochasticDualCoordinateAscentClassifier
{
TrainingData = nopOutput.OutputData,
NumThreads = 1
};
var learnerOutput = subGraph.Add(learnerInput);
var modelCombine = new Legacy.Transforms.ManyHeterogeneousModelCombiner
{
TransformModels = new ArrayVar <TransformModel>(nopOutput.Model),
PredictorModel = learnerOutput.PredictorModel
};
var modelCombineOutput = subGraph.Add(modelCombine);
var experiment = env.CreateExperiment();
var importInput = new Legacy.Data.TextLoader(dataPath);
var importOutput = experiment.Add(importInput);
var crossValidate = new Legacy.Models.CrossValidator
{
Data = importOutput.Data,
Nodes = subGraph,
Kind = Legacy.Models.MacroUtilsTrainerKinds.SignatureMultiClassClassifierTrainer,
TransformModel = null
};
crossValidate.Inputs.Data = nop.Data;
crossValidate.Outputs.PredictorModel = modelCombineOutput.PredictorModel;
var crossValidateOutput = experiment.Add(crossValidate);
experiment.Compile();
importInput.SetInput(env, experiment);
experiment.Run();
var data = experiment.GetOutput(crossValidateOutput.OverallMetrics);
var schema = data.Schema;
var b = schema.TryGetColumnIndex("Accuracy(micro-avg)", out int metricCol);
Assert.True(b);
b = schema.TryGetColumnIndex("Fold Index", out int foldCol);
Assert.True(b);
using (var cursor = data.GetRowCursor(col => col == metricCol || col == foldCol))
{
var getter = cursor.GetGetter <double>(metricCol);
var foldGetter = cursor.GetGetter <ReadOnlyMemory <char> >(foldCol);
ReadOnlyMemory <char> fold = default;
// Get the average.
b = cursor.MoveNext();
Assert.True(b);
double avg = 0;
getter(ref avg);
foldGetter(ref fold);
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Average", fold));
// Get the standard deviation.
b = cursor.MoveNext();
Assert.True(b);
double stdev = 0;
getter(ref stdev);
foldGetter(ref fold);
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Standard Deviation", fold));
Assert.Equal(0.015, stdev, 3);
double sum = 0;
double val = 0;
for (int f = 0; f < 2; f++)
{
b = cursor.MoveNext();
Assert.True(b);
getter(ref val);
foldGetter(ref fold);
sum += val;
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Fold " + f, fold));
}
Assert.Equal(avg, sum / 2);
b = cursor.MoveNext();
Assert.False(b);
}
var confusion = experiment.GetOutput(crossValidateOutput.ConfusionMatrix);
schema = confusion.Schema;
b = schema.TryGetColumnIndex("Count", out int countCol);
Assert.True(b);
b = schema.TryGetColumnIndex("Fold Index", out foldCol);
Assert.True(b);
var type = schema[countCol].Metadata.Schema[MetadataUtils.Kinds.SlotNames].Type;
Assert.True(type is VectorType vecType && vecType.ItemType is TextType && vecType.Size == 10);
var slotNames = default(VBuffer <ReadOnlyMemory <char> >);
schema[countCol].GetSlotNames(ref slotNames);
var slotNameValues = slotNames.GetValues();
for (int i = 0; i < slotNameValues.Length; i++)
{
Assert.True(ReadOnlyMemoryUtils.EqualsStr(i.ToString(), slotNameValues[i]));
}
using (var curs = confusion.GetRowCursor(col => true))
{
var countGetter = curs.GetGetter <VBuffer <double> >(countCol);
var foldGetter = curs.GetGetter <ReadOnlyMemory <char> >(foldCol);
var confCount = default(VBuffer <double>);
var foldIndex = default(ReadOnlyMemory <char>);
int rowCount = 0;
var foldCur = "Fold 0";
while (curs.MoveNext())
{
countGetter(ref confCount);
foldGetter(ref foldIndex);
rowCount++;
Assert.True(ReadOnlyMemoryUtils.EqualsStr(foldCur, foldIndex));
if (rowCount == 10)
{
rowCount = 0;
foldCur = "Fold 1";
}
}
Assert.Equal(0, rowCount);
}
var warnings = experiment.GetOutput(crossValidateOutput.Warnings);
using (var cursor = warnings.GetRowCursor(col => true))
Assert.False(cursor.MoveNext());
}
[ConditionalFact(typeof(BaseTestBaseline), nameof(BaseTestBaseline.LessThanNetCore30OrNotNetCore))] // netcore3.0 output differs from Baseline
public void TestCrossValidationMacro()
{
var dataPath = GetDataPath(TestDatasets.generatedRegressionDatasetmacro.trainFilename);
var env = new MLContext(42);
var subGraph = env.CreateExperiment();
var nop = new Legacy.Transforms.NoOperation();
var nopOutput = subGraph.Add(nop);
var generate = new Legacy.Transforms.RandomNumberGenerator();
generate.Column = new[] { new Legacy.Transforms.GenerateNumberTransformColumn()
{
Name = "Weight1"
} };
generate.Data = nopOutput.OutputData;
var generateOutput = subGraph.Add(generate);
var learnerInput = new Legacy.Trainers.PoissonRegressor
{
TrainingData = generateOutput.OutputData,
NumThreads = 1,
WeightColumn = "Weight1"
};
var learnerOutput = subGraph.Add(learnerInput);
var modelCombine = new Legacy.Transforms.ManyHeterogeneousModelCombiner
{
TransformModels = new ArrayVar <TransformModel>(nopOutput.Model, generateOutput.Model),
PredictorModel = learnerOutput.PredictorModel
};
var modelCombineOutput = subGraph.Add(modelCombine);
var experiment = env.CreateExperiment();
var importInput = new Legacy.Data.TextLoader(dataPath)
{
Arguments = new Legacy.Data.TextLoaderArguments
{
Separator = new[] { ';' },
HasHeader = true,
Column = new[]
{
new TextLoaderColumn()
{
Name = "Label",
Source = new [] { new TextLoaderRange(11) },
Type = Legacy.Data.DataKind.Num
},
new TextLoaderColumn()
{
Name = "Features",
Source = new [] { new TextLoaderRange(0, 10) },
Type = Legacy.Data.DataKind.Num
}
}
}
};
var importOutput = experiment.Add(importInput);
var crossValidate = new Legacy.Models.CrossValidator
{
Data = importOutput.Data,
Nodes = subGraph,
Kind = Legacy.Models.MacroUtilsTrainerKinds.SignatureRegressorTrainer,
TransformModel = null,
WeightColumn = "Weight1"
};
crossValidate.Inputs.Data = nop.Data;
crossValidate.Outputs.PredictorModel = modelCombineOutput.PredictorModel;
var crossValidateOutput = experiment.Add(crossValidate);
experiment.Compile();
importInput.SetInput(env, experiment);
experiment.Run();
var data = experiment.GetOutput(crossValidateOutput.OverallMetrics);
var schema = data.Schema;
var b = schema.TryGetColumnIndex("L1(avg)", out int metricCol);
Assert.True(b);
b = schema.TryGetColumnIndex("Fold Index", out int foldCol);
Assert.True(b);
b = schema.TryGetColumnIndex("IsWeighted", out int isWeightedCol);
using (var cursor = data.GetRowCursor(col => col == metricCol || col == foldCol || col == isWeightedCol))
{
var getter = cursor.GetGetter <double>(metricCol);
var foldGetter = cursor.GetGetter <ReadOnlyMemory <char> >(foldCol);
ReadOnlyMemory <char> fold = default;
var isWeightedGetter = cursor.GetGetter <bool>(isWeightedCol);
bool isWeighted = default;
double avg = 0;
double weightedAvg = 0;
for (int w = 0; w < 2; w++)
{
// Get the average.
b = cursor.MoveNext();
Assert.True(b);
if (w == 1)
{
getter(ref weightedAvg);
}
else
{
getter(ref avg);
}
foldGetter(ref fold);
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Average", fold));
isWeightedGetter(ref isWeighted);
Assert.True(isWeighted == (w == 1));
// Get the standard deviation.
b = cursor.MoveNext();
Assert.True(b);
double stdev = 0;
getter(ref stdev);
foldGetter(ref fold);
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Standard Deviation", fold));
if (w == 1)
{
Assert.Equal(1.585, stdev, 3);
}
else
{
Assert.Equal(1.39, stdev, 2);
}
isWeightedGetter(ref isWeighted);
Assert.True(isWeighted == (w == 1));
}
double sum = 0;
double weightedSum = 0;
for (int f = 0; f < 2; f++)
{
for (int w = 0; w < 2; w++)
{
b = cursor.MoveNext();
Assert.True(b);
double val = 0;
getter(ref val);
foldGetter(ref fold);
if (w == 1)
{
weightedSum += val;
}
else
{
sum += val;
}
Assert.True(ReadOnlyMemoryUtils.EqualsStr("Fold " + f, fold));
isWeightedGetter(ref isWeighted);
Assert.True(isWeighted == (w == 1));
}
}
Assert.Equal(weightedAvg, weightedSum / 2);
Assert.Equal(avg, sum / 2);
b = cursor.MoveNext();
Assert.False(b);
}
}
private void LoadStopWords(IHostEnvironment env, IChannel ch, ArgumentsBase loaderArgs, out NormStr.Pool stopWordsMap)
{
Contracts.AssertValue(env);
env.AssertValue(ch);
ch.AssertValue(loaderArgs);
if ((!string.IsNullOrEmpty(loaderArgs.Stopwords) || Utils.Size(loaderArgs.Stopword) > 0) &&
(!string.IsNullOrWhiteSpace(loaderArgs.DataFile) || loaderArgs.Loader != null ||
!string.IsNullOrWhiteSpace(loaderArgs.StopwordsColumn)))
{
ch.Warning("Explicit stopwords list specified. Data file arguments will be ignored");
}
var src = default(ReadOnlyMemory <char>);
stopWordsMap = new NormStr.Pool();
var buffer = new StringBuilder();
var stopwords = loaderArgs.Stopwords.AsMemory();
stopwords = ReadOnlyMemoryUtils.TrimSpaces(stopwords);
if (!stopwords.IsEmpty)
{
bool warnEmpty = true;
for (bool more = true; more;)
{
ReadOnlyMemory <char> stopword;
more = ReadOnlyMemoryUtils.SplitOne(stopwords, ',', out stopword, out stopwords);
stopword = ReadOnlyMemoryUtils.TrimSpaces(stopword);
if (!stopword.IsEmpty)
{
buffer.Clear();
ReadOnlyMemoryUtils.AddLowerCaseToStringBuilder(stopword.Span, buffer);
stopWordsMap.Add(buffer);
}
else if (warnEmpty)
{
ch.Warning("Empty strings ignored in 'stopwords' specification");
warnEmpty = false;
}
}
ch.CheckUserArg(stopWordsMap.Count > 0, nameof(Arguments.Stopwords), "stopwords is empty");
}
else if (Utils.Size(loaderArgs.Stopword) > 0)
{
bool warnEmpty = true;
foreach (string word in loaderArgs.Stopword)
{
var stopword = word.AsSpan();
stopword = stopword.Trim(' ');
if (!stopword.IsEmpty)
{
buffer.Clear();
ReadOnlyMemoryUtils.AddLowerCaseToStringBuilder(stopword, buffer);
stopWordsMap.Add(buffer);
}
else if (warnEmpty)
{
ch.Warning("Empty strings ignored in 'stopword' specification");
warnEmpty = false;
}
}
}
else
{
string srcCol = loaderArgs.StopwordsColumn;
var loader = LoadStopwords(env, ch, loaderArgs.DataFile, loaderArgs.Loader, ref srcCol);
int colSrc;
if (!loader.Schema.TryGetColumnIndex(srcCol, out colSrc))
{
throw ch.ExceptUserArg(nameof(Arguments.StopwordsColumn), "Unknown column '{0}'", srcCol);
}
var typeSrc = loader.Schema[colSrc].Type;
ch.CheckUserArg(typeSrc.IsText, nameof(Arguments.StopwordsColumn), "Must be a scalar text column");
// Accumulate the stopwords.
using (var cursor = loader.GetRowCursor(col => col == colSrc))
{
bool warnEmpty = true;
var getter = cursor.GetGetter <ReadOnlyMemory <char> >(colSrc);
while (cursor.MoveNext())
{
getter(ref src);
if (!src.IsEmpty)
{
buffer.Clear();
ReadOnlyMemoryUtils.AddLowerCaseToStringBuilder(src.Span, buffer);
stopWordsMap.Add(buffer);
}
else if (warnEmpty)
{
ch.Warning("Empty rows ignored in data file");
warnEmpty = false;
}
}
}
ch.CheckUserArg(stopWordsMap.Count > 0, nameof(Arguments.DataFile), "dataFile is empty");
}
}
