private IDataView WrapPerInstance(RoleMappedData perInst)
{
var idv = perInst.Data;
// Make a list of column names that Maml outputs as part of the per-instance data view, and then wrap
// the per-instance data computed by the evaluator in a SelectColumnsTransform.
var cols = new List <(string name, string source)>();
var colsToKeep = new List <string>();
// If perInst is the result of cross-validation and contains a fold Id column, include it.
int foldCol;
if (perInst.Schema.Schema.TryGetColumnIndex(MetricKinds.ColumnNames.FoldIndex, out foldCol))
{
colsToKeep.Add(MetricKinds.ColumnNames.FoldIndex);
}
// Maml always outputs a name column, if it doesn't exist add a GenerateNumberTransform.
if (perInst.Schema.Name?.Name is string nameName)
{
cols.Add(("Instance", nameName));
colsToKeep.Add("Instance");
}
else
{
var args = new GenerateNumberTransform.Options();
args.Columns = new[] { new GenerateNumberTransform.Column()
{
Name = "Instance"
} };
args.UseCounter = true;
idv = new GenerateNumberTransform(Host, args, idv);
colsToKeep.Add("Instance");
}
// Maml outputs the weight column if it exists.
if (perInst.Schema.Weight?.Name is string weightName)
{
colsToKeep.Add(weightName);
}
// Get the other columns from the evaluator.
foreach (var col in GetPerInstanceColumnsToSave(perInst.Schema))
{
colsToKeep.Add(col);
}
idv = new ColumnCopyingTransformer(Host, cols.ToArray()).Transform(idv);
idv = ColumnSelectingTransformer.CreateKeep(Host, idv, colsToKeep.ToArray());
return(GetPerInstanceMetricsCore(idv, perInst.Schema));
}
public IEnumerable <Batch> GetBatches(Random rand)
{
Host.Assert(Data != null, "Must call Initialize first!");
Host.AssertValue(rand);
using (var ch = Host.Start("Getting batches"))
{
RoleMappedData dataTest;
RoleMappedData dataTrain;
// Split the data, if needed.
if (!(ValidationDatasetProportion > 0))
{
dataTest = dataTrain = Data;
}
else
{
// Split the data into train and test sets.
string name = Data.Data.Schema.GetTempColumnName();
var args = new GenerateNumberTransform.Options();
args.Columns = new[] { new GenerateNumberTransform.Column()
{
Name = name
} };
args.Seed = (uint)rand.Next();
var view = new GenerateNumberTransform(Host, args, Data.Data);
var viewTest = new RangeFilter(Host, new RangeFilter.Options()
{
Column = name, Max = ValidationDatasetProportion
}, view);
var viewTrain = new RangeFilter(Host, new RangeFilter.Options()
{
Column = name, Max = ValidationDatasetProportion, Complement = true
}, view);
dataTest = new RoleMappedData(viewTest, Data.Schema.GetColumnRoleNames());
dataTrain = new RoleMappedData(viewTrain, Data.Schema.GetColumnRoleNames());
}
if (BatchSize > 0)
{
// REVIEW: How should we carve the data into batches?
ch.Warning("Batch support is temporarily disabled");
}
yield return(new Batch(dataTrain, dataTest));
}
}
public override IEnumerable <Subset> GetSubsets(Batch batch, Random rand)
{
string name = Data.Data.Schema.GetTempColumnName();
var args = new GenerateNumberTransform.Options();
args.Columns = new[] { new GenerateNumberTransform.Column()
{
Name = name
} };
args.Seed = (uint)rand.Next();
IDataTransform view = new GenerateNumberTransform(Host, args, Data.Data);
// REVIEW: This won't be very efficient when Size is large.
for (int i = 0; i < Size; i++)
{
var viewTrain = new RangeFilter(Host, new RangeFilter.Options()
{
Column = name, Min = (Double)i / Size, Max = (Double)(i + 1) / Size
}, view);
var dataTrain = new RoleMappedData(viewTrain, Data.Schema.GetColumnRoleNames());
yield return(FeatureSelector.SelectFeatures(dataTrain, rand));
}
}
protected override TVectorPredictor TrainPredictor(IChannel ch, TScalarTrainer trainer, RoleMappedData data, int count)
{
var data0 = data;
#region adding group ID
// We insert a group Id.
string groupColumnTemp = DataViewUtils.GetTempColumnName(data.Schema.Schema) + "GR";
var groupArgs = new GenerateNumberTransform.Options
{
Columns = new[] { GenerateNumberTransform.Column.Parse(groupColumnTemp) },
UseCounter = true
};
var withGroup = new GenerateNumberTransform(Host, groupArgs, data.Data);
data = new RoleMappedData(withGroup, data.Schema.GetColumnRoleNames());
#endregion
#region preparing the training dataset
string dstName, labName;
var trans = MapLabelsAndInsertTransform(ch, data, out dstName, out labName, count, true, _args);
var newFeatures = trans.Schema.GetTempColumnName() + "NF";
// We check the label is not boolean.
int indexLab = SchemaHelper.GetColumnIndex(trans.Schema, dstName);
var typeLab = trans.Schema[indexLab].Type;
if (typeLab.RawKind() == DataKind.Boolean)
{
throw Host.Except("Column '{0}' has an unexpected type {1}.", dstName, typeLab.RawKind());
}
var args3 = new DescribeTransform.Arguments {
columns = new string[] { labName, dstName }, oneRowPerColumn = true
};
var desc = new DescribeTransform(Host, args3, trans);
IDataView viewI;
if (_args.singleColumn && data.Schema.Label.Value.Type.RawKind() == DataKind.Single)
{
viewI = desc;
}
else if (_args.singleColumn)
{
var sch = new TypeReplacementSchema(desc.Schema, new[] { labName }, new[] { NumberDataViewType.Single });
viewI = new TypeReplacementDataView(desc, sch);
#region debug
#if (DEBUG)
DebugChecking0(viewI, labName, false);
#endif
#endregion
}
else if (data.Schema.Label.Value.Type.IsKey())
{
ulong nb = data.Schema.Label.Value.Type.AsKey().GetKeyCount();
var sch = new TypeReplacementSchema(desc.Schema, new[] { labName }, new[] { new VectorDataViewType(NumberDataViewType.Single, (int)nb) });
viewI = new TypeReplacementDataView(desc, sch);
#region debug
#if (DEBUG)
int nb_;
MinMaxLabelOverDataSet(trans, labName, out nb_);
int count3;
data.CheckMulticlassLabel(out count3);
if ((ulong)count3 != nb)
{
throw ch.Except("Count mismatch (KeyCount){0} != {1}", nb, count3);
}
DebugChecking0(viewI, labName, true);
DebugChecking0Vfloat(viewI, labName, nb);
#endif
#endregion
}
else
{
int nb;
if (count <= 0)
{
MinMaxLabelOverDataSet(trans, labName, out nb);
}
else
{
nb = count;
}
var sch = new TypeReplacementSchema(desc.Schema, new[] { labName }, new[] { new VectorDataViewType(NumberDataViewType.Single, nb) });
viewI = new TypeReplacementDataView(desc, sch);
#region debug
#if (DEBUG)
DebugChecking0(viewI, labName, true);
#endif
#endregion
}
ch.Info("Merging column label '{0}' with features '{1}'", labName, data.Schema.Feature.Value.Name);
var args = string.Format("Concat{{col={0}:{1},{2}}}", newFeatures, data.Schema.Feature.Value.Name, labName);
var after_concatenation_ = ComponentCreation.CreateTransform(Host, args, viewI);
#endregion
#region converting label and group into keys
// We need to convert the label into a Key.
var convArgs = new MulticlassConvertTransform.Arguments
{
column = new[] { MulticlassConvertTransform.Column.Parse(string.Format("{0}k:{0}", dstName)) },
keyCount = new KeyCount(4),
resultType = DataKind.UInt32
};
IDataView after_concatenation_key_label = new MulticlassConvertTransform(Host, convArgs, after_concatenation_);
// The group must be a key too!
convArgs = new MulticlassConvertTransform.Arguments
{
column = new[] { MulticlassConvertTransform.Column.Parse(string.Format("{0}k:{0}", groupColumnTemp)) },
keyCount = new KeyCount(),
resultType = _args.groupIsU4 ? DataKind.UInt32 : DataKind.UInt64
};
after_concatenation_key_label = new MulticlassConvertTransform(Host, convArgs, after_concatenation_key_label);
#endregion
#region preparing the RoleMapData view
string groupColumn = groupColumnTemp + "k";
dstName += "k";
var roles = data.Schema.GetColumnRoleNames();
var rolesArray = roles.ToArray();
roles = roles
.Where(kvp => kvp.Key.Value != RoleMappedSchema.ColumnRole.Label.Value)
.Where(kvp => kvp.Key.Value != RoleMappedSchema.ColumnRole.Feature.Value)
.Where(kvp => kvp.Key.Value != groupColumn)
.Where(kvp => kvp.Key.Value != groupColumnTemp);
rolesArray = roles.ToArray();
if (rolesArray.Any() && rolesArray[0].Value == groupColumnTemp)
{
throw ch.Except("Duplicated group.");
}
roles = roles
.Prepend(RoleMappedSchema.ColumnRole.Feature.Bind(newFeatures))
.Prepend(RoleMappedSchema.ColumnRole.Label.Bind(dstName))
.Prepend(RoleMappedSchema.ColumnRole.Group.Bind(groupColumn));
var trainer_input = new RoleMappedData(after_concatenation_key_label, roles);
#endregion
ch.Info("New Features: {0}:{1}", trainer_input.Schema.Feature.Value.Name, trainer_input.Schema.Feature.Value.Type);
ch.Info("New Label: {0}:{1}", trainer_input.Schema.Label.Value.Name, trainer_input.Schema.Label.Value.Type);
// We train the unique binary classifier.
var trainedPredictor = trainer.Train(trainer_input);
var predictors = new TScalarPredictor[] { trainedPredictor };
// We train the reclassification classifier.
if (_args.reclassicationPredictor != null)
{
var pred = CreateFinalPredictor(ch, data, trans, count, _args, predictors, null);
TrainReclassificationPredictor(data0, pred, ScikitSubComponent <ITrainer, SignatureTrainer> .AsSubComponent(_args.reclassicationPredictor));
}
return(CreateFinalPredictor(ch, data, trans, count, _args, predictors, _reclassPredictor));
}
private string GetSplitColumn(IChannel ch, IDataView input, ref IDataView output)
{
// The stratification column and/or group column, if they exist at all, must be present at this point.
var schema = input.Schema;
output = input;
// If no stratification column was specified, but we have a group column of type Single, Double or
// Key (contiguous) use it.
string stratificationColumn = null;
if (!string.IsNullOrWhiteSpace(Args.StratificationColumn))
{
stratificationColumn = Args.StratificationColumn;
}
else
{
string group = TrainUtils.MatchNameOrDefaultOrNull(ch, schema, nameof(Args.GroupColumn), Args.GroupColumn, DefaultColumnNames.GroupId);
int index;
if (group != null && schema.TryGetColumnIndex(group, out index))
{
// Check if group column key type with known cardinality.
var type = schema[index].Type;
if (type.GetKeyCount() > 0)
{
stratificationColumn = group;
}
}
}
if (string.IsNullOrEmpty(stratificationColumn))
{
stratificationColumn = "StratificationColumn";
int tmp;
int inc = 0;
while (input.Schema.TryGetColumnIndex(stratificationColumn, out tmp))
{
stratificationColumn = string.Format("StratificationColumn_{0:000}", ++inc);
}
var keyGenArgs = new GenerateNumberTransform.Options();
var col = new GenerateNumberTransform.Column();
col.Name = stratificationColumn;
keyGenArgs.Columns = new[] { col };
output = new GenerateNumberTransform(Host, keyGenArgs, input);
}
else
{
int col;
if (!input.Schema.TryGetColumnIndex(stratificationColumn, out col))
{
throw ch.ExceptUserArg(nameof(Arguments.StratificationColumn), "Column '{0}' does not exist", stratificationColumn);
}
var type = input.Schema[col].Type;
if (!RangeFilter.IsValidRangeFilterColumnType(ch, type))
{
ch.Info("Hashing the stratification column");
var origStratCol = stratificationColumn;
int tmp;
int inc = 0;
while (input.Schema.TryGetColumnIndex(stratificationColumn, out tmp))
{
stratificationColumn = string.Format("{0}_{1:000}", origStratCol, ++inc);
}
output = new HashingEstimator(Host, origStratCol, stratificationColumn, 30).Fit(input).Transform(input);
}
}
return(stratificationColumn);
}
private void RunCore(IChannel ch, string cmd)
{
Host.AssertValue(ch);
IPredictor inputPredictor = null;
if (Args.ContinueTrain && !TrainUtils.TryLoadPredictor(ch, Host, Args.InputModelFile, out inputPredictor))
{
ch.Warning("No input model file specified or model file did not contain a predictor. The model state cannot be initialized.");
}
ch.Trace("Constructing data pipeline");
IDataLoader loader = CreateRawLoader();
// If the per-instance results are requested and there is no name column, add a GenerateNumberTransform.
var preXf = Args.PreTransforms;
if (!string.IsNullOrEmpty(Args.OutputDataFile))
{
string name = TrainUtils.MatchNameOrDefaultOrNull(ch, loader.Schema, nameof(Args.NameColumn), Args.NameColumn, DefaultColumnNames.Name);
if (name == null)
{
preXf = preXf.Concat(
new[]
{
new KeyValuePair <string, IComponentFactory <IDataView, IDataTransform> >(
"", ComponentFactoryUtils.CreateFromFunction <IDataView, IDataTransform>(
(env, input) =>
{
var args = new GenerateNumberTransform.Options();
args.Columns = new[] { new GenerateNumberTransform.Column()
{
Name = DefaultColumnNames.Name
}, };
args.UseCounter = true;
return(new GenerateNumberTransform(env, args, input));
}))
}).ToArray();
}
}
loader = CompositeDataLoader.Create(Host, loader, preXf);
ch.Trace("Binding label and features columns");
IDataView pipe = loader;
var stratificationColumn = GetSplitColumn(ch, loader, ref pipe);
var scorer = Args.Scorer;
var evaluator = Args.Evaluator;
Func <IDataView> validDataCreator = null;
if (Args.ValidationFile != null)
{
validDataCreator =
() =>
{
// Fork the command.
var impl = new CrossValidationCommand(this);
return(impl.CreateRawLoader(dataFile: Args.ValidationFile));
};
}
FoldHelper fold = new FoldHelper(Host, RegistrationName, pipe, stratificationColumn,
Args, CreateRoleMappedData, ApplyAllTransformsToData, scorer, evaluator,
validDataCreator, ApplyAllTransformsToData, inputPredictor, cmd, loader, !string.IsNullOrEmpty(Args.OutputDataFile));
var tasks = fold.GetCrossValidationTasks();
var eval = evaluator?.CreateComponent(Host) ??
EvaluateUtils.GetEvaluator(Host, tasks[0].Result.ScoreSchema);
// Print confusion matrix and fold results for each fold.
for (int i = 0; i < tasks.Length; i++)
{
var dict = tasks[i].Result.Metrics;
MetricWriter.PrintWarnings(ch, dict);
eval.PrintFoldResults(ch, dict);
}
// Print the overall results.
if (!TryGetOverallMetrics(tasks.Select(t => t.Result.Metrics).ToArray(), out var overallList))
{
throw ch.Except("No overall metrics found");
}
var overall = eval.GetOverallResults(overallList.ToArray());
MetricWriter.PrintOverallMetrics(Host, ch, Args.SummaryFilename, overall, Args.NumFolds);
eval.PrintAdditionalMetrics(ch, tasks.Select(t => t.Result.Metrics).ToArray());
Dictionary <string, IDataView>[] metricValues = tasks.Select(t => t.Result.Metrics).ToArray();
SendTelemetryMetric(metricValues);
// Save the per-instance results.
if (!string.IsNullOrWhiteSpace(Args.OutputDataFile))
{
var perInstance = EvaluateUtils.ConcatenatePerInstanceDataViews(Host, eval, Args.CollateMetrics,
Args.OutputExampleFoldIndex, tasks.Select(t => t.Result.PerInstanceResults).ToArray(), out var variableSizeVectorColumnNames);
if (variableSizeVectorColumnNames.Length > 0)
{
ch.Warning("Detected columns of variable length: {0}. Consider setting collateMetrics- for meaningful per-Folds results.",
string.Join(", ", variableSizeVectorColumnNames));
}
if (Args.CollateMetrics)
{
ch.Assert(perInstance.Length == 1);
MetricWriter.SavePerInstance(Host, ch, Args.OutputDataFile, perInstance[0]);
}
else
{
int i = 0;
foreach (var idv in perInstance)
{
MetricWriter.SavePerInstance(Host, ch, ConstructPerFoldName(Args.OutputDataFile, i), idv);
i++;
}
}
}
}
IDataTransform AppendToPipeline(IDataView input)
{
IDataView current = input;
if (_shuffleInput)
{
var args1 = new RowShufflingTransformer.Options()
{
ForceShuffle = false,
ForceShuffleSeed = _seedShuffle,
PoolRows = _poolRows,
PoolOnly = false,
};
current = new RowShufflingTransformer(Host, args1, current);
}
// We generate a random number.
var columnName = current.Schema.GetTempColumnName();
var args2 = new GenerateNumberTransform.Options()
{
Columns = new GenerateNumberTransform.Column[] { new GenerateNumberTransform.Column()
{
Name = columnName
} },
Seed = _seed ?? 42
};
IDataTransform currentTr = new GenerateNumberTransform(Host, args2, current);
// We convert this random number into a part.
var cRatios = new float[_ratios.Length];
cRatios[0] = 0;
for (int i = 1; i < _ratios.Length; ++i)
{
cRatios[i] = cRatios[i - 1] + _ratios[i - 1];
}
ValueMapper <float, int> mapper = (in float src, ref int dst) =>
{
for (int i = cRatios.Length - 1; i > 0; --i)
{
if (src >= cRatios[i])
{
dst = i;
return;
}
}
dst = 0;
};
// Get location of columnName
int index = SchemaHelper.GetColumnIndex(currentTr.Schema, columnName);
var ct = currentTr.Schema[index].Type;
var view = LambdaColumnMapper.Create(Host, "Key to part mapper", currentTr,
columnName, _newColumn, ct, NumberDataViewType.Int32, mapper);
// We cache the result to avoid the pipeline to change the random number.
var args3 = new ExtendedCacheTransform.Arguments()
{
inDataFrame = string.IsNullOrEmpty(_cacheFile),
numTheads = _numThreads,
cacheFile = _cacheFile,
reuse = _reuse,
};
currentTr = new ExtendedCacheTransform(Host, args3, view);
// Removing the temporary column.
var objtr = ColumnSelectingTransformer.CreateDrop(Host, currentTr, new string[] { columnName });
var finalTr = objtr as IDataTransform;
if (finalTr == null)
{
throw Contracts.ExceptNotSupp("Desgin change.");
}
var taggedViews = new List <Tuple <string, ITaggedDataView> >();
// filenames
if (_filenames != null || _tags != null)
{
int nbf = _filenames == null ? 0 : _filenames.Length;
if (nbf > 0 && nbf != _ratios.Length)
{
throw Host.Except("Differen number of filenames and ratios.");
}
int nbt = _tags == null ? 0 : _tags.Length;
if (nbt > 0 && nbt != _ratios.Length)
{
throw Host.Except("Differen number of filenames and ratios.");
}
int nb = Math.Max(nbf, nbt);
using (var ch = Host.Start("Split the datasets and stores each part."))
{
for (int i = 0; i < nb; ++i)
{
if (_filenames == null || !_filenames.Any())
{
ch.Info("Create part {0}: {1} (tag: {2})", i + 1, _ratios[i], _tags[i]);
}
else
{
ch.Info("Create part {0}: {1} (file: {2})", i + 1, _ratios[i], _filenames[i]);
}
var ar1 = new RangeFilter.Options()
{
Column = _newColumn, Min = i, Max = i, IncludeMax = true
};
int pardId = i;
var filtView = LambdaFilter.Create <int>(Host, string.Format("Select part {0}", i), currentTr,
_newColumn, NumberDataViewType.Int32,
(in int part) => { return(part.Equals(pardId)); });
