Libraries/Microsoft.ML.Api.xml
|
<?xml version="1.0"?>
<doc> <assembly> <name>Microsoft.ML.Api</name> </assembly> <members> <member name="M:Microsoft.ML.Runtime.Api.ApiUtils.GeneratePeek``2(Microsoft.ML.Runtime.Api.InternalSchemaDefinition.Column)"> <summary> Each of the specialized 'peek' methods copies the appropriate field value of an instance of T into the provided buffer. So, the call is 'peek(userObject, ref destination)' and the logic is indentical to 'destination = userObject.##FIELD##', where ##FIELD## is defined per peek method. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ApiUtils.GeneratePoke``2(Microsoft.ML.Runtime.Api.InternalSchemaDefinition.Column)"> <summary> Each of the specialized 'poke' methods sets the appropriate field value of an instance of T to the provided value. So, the call is 'peek(userObject, providedValue)' and the logic is indentical to 'userObject.##FIELD## = providedValue', where ##FIELD## is defined per poke method. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.CodeGenerationUtils"> <summary> Utility methods for code generation. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.CodeGenerationUtils.MultiReplace(System.String,System.Collections.Generic.Dictionary{System.String,System.String})"> <summary> Replace placeholders with provided values. Assert that every placeholder is found. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.CodeGenerationUtils.AppendFieldDeclaration(Microsoft.CSharp.CSharpCodeProvider,System.Text.StringBuilder,System.Int32,System.String,Microsoft.ML.Runtime.Data.ColumnType,System.Boolean,System.Boolean)"> <summary> Append a field declaration to the provided <see cref="T:System.Text.StringBuilder"/>. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.CodeGenerationUtils.GetCSharpString(Microsoft.CSharp.CSharpCodeProvider,System.String)"> <summary> Generates a C# string for a given input (with proper escaping). </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.CodeGenerationUtils.GetBackingTypeName(Microsoft.ML.Runtime.Data.ColumnType,System.Boolean,System.Collections.Generic.List{System.String})"> <summary> Gets the C# strings representing the type name for a variable corresponding to the <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/> column type. If the type is a vector, then <paramref name="useVBuffer"/> controls whether the array field is generated or <see cref="T:Microsoft.ML.Runtime.Data.VBuffer`1"/>. If additional attributes are required, they are appended to the <paramref name="attributes"/> list. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.ComponentCreation"> <summary> This class defines extension methods for an <see cref="T:Microsoft.ML.Runtime.IHostEnvironment"/> to facilitate creating components (loaders, transforms, trainers, scorers, evaluators, savers). </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.Zip(Microsoft.ML.Runtime.IHostEnvironment,System.Collections.Generic.IEnumerable{Microsoft.ML.Runtime.Data.IDataView})"> <summary> Create a new data view which is obtained by appending all columns of all the source data views. If the data views are of different length, the resulting data view will have the length equal to the length of the shortest source. </summary> <param name="env">The host environment to use.</param> <param name="sources">A non-empty collection of data views to zip together.</param> <returns>The resulting data view.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateExamples(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.String,System.String,System.String,System.String,System.Collections.Generic.IEnumerable{System.Collections.Generic.KeyValuePair{Microsoft.ML.Runtime.Data.RoleMappedSchema.ColumnRole,System.String}})"> <summary> Generate training examples for training a predictor or instantiating a scorer. </summary> <param name="env">The host environment to use.</param> <param name="data">The data to use for training or scoring.</param> <param name="features">The name of the features column. Can be null.</param> <param name="label">The name of the label column. Can be null.</param> <param name="group">The name of the group ID column (for ranking). Can be null.</param> <param name="weight">The name of the weight column. Can be null.</param> <param name="custom">Additional column mapping to be passed to the trainer or scorer (specific to the prediction type). Can be null or empty.</param> <returns>The constructed examples.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateDataView``1(Microsoft.ML.Runtime.IHostEnvironment,System.Collections.Generic.IList{``0},Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Create a new <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/> over an in-memory collection of the items of user-defined type. The user maintains ownership of the <paramref name="data"/> and the resulting data view will never alter the contents of the <paramref name="data"/>. Since <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/> is assumed to be immutable, the user is expected to not modify the contents of <paramref name="data"/> while the data view is being actively cursored. One typical usage for in-memory data view could be: create the data view, train a predictor. Once the predictor is fully trained, modify the contents of the underlying collection and train another predictor. </summary> <typeparam name="TRow">The user-defined item type.</typeparam> <param name="env">The host environment to use for data view creation.</param> <param name="data">The data to wrap around.</param> <param name="schemaDefinition">The optional schema definition of the data view to create. If <c>null</c>, the schema definition is inferred from <typeparamref name="TRow"/>.</param> <returns>The constructed <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateStreamingDataView``1(Microsoft.ML.Runtime.IHostEnvironment,System.Collections.Generic.IEnumerable{``0},Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Create a new <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/> over an enumerable of the items of user-defined type. The user maintains ownership of the <paramref name="data"/> and the resulting data view will never alter the contents of the <paramref name="data"/>. Since <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/> is assumed to be immutable, the user is expected to support multiple enumeration of the <paramref name="data"/> that would return the same results, unless the user knows that the data will only be cursored once. One typical usage for streaming data view could be: create the data view that lazily loads data as needed, then apply pre-trained transformations to it and cursor through it for transformation results. This is how <see cref="T:Microsoft.ML.Runtime.Api.BatchPredictionEngine`2"/> is implemented. </summary> <typeparam name="TRow">The user-defined item type.</typeparam> <param name="env">The host environment to use for data view creation.</param> <param name="data">The data to wrap around.</param> <param name="schemaDefinition">The optional schema definition of the data view to create. If <c>null</c>, the schema definition is inferred from <typeparamref name="TRow"/>.</param> <returns>The constructed <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateBatchPredictionEngine``2(Microsoft.ML.Runtime.IHostEnvironment,System.IO.Stream,System.Boolean,Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Create a batch prediction engine. </summary> <param name="env">The host environment to use.</param> <param name="modelStream">The stream to deserialize the pipeline (transforms and predictor) from.</param> <param name="ignoreMissingColumns">Whether to ignore missing columns in the data view.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <param name="outputSchemaDefinition">The optional output schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TDst"/> type.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateBatchPredictionEngine``2(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Boolean,Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Create a batch prediction engine. </summary> <param name="env">The host environment to use.</param> <param name="dataPipe">The transformation pipe that may or may not include a scorer.</param> <param name="ignoreMissingColumns">Whether to ignore missing columns in the data view.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <param name="outputSchemaDefinition">The optional output schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TDst"/> type.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreatePredictionEngine``2(Microsoft.ML.Runtime.IHostEnvironment,System.IO.Stream,System.Boolean,Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Create an on-demand prediction engine. </summary> <param name="env">The host environment to use.</param> <param name="modelStream">The stream to deserialize the pipeline (transforms and predictor) from.</param> <param name="ignoreMissingColumns">Whether to ignore missing columns in the data view.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <param name="outputSchemaDefinition">The optional output schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TDst"/> type.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreatePredictionEngine``2(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Boolean,Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Create an on-demand prediction engine. </summary> <param name="env">The host environment to use.</param> <param name="dataPipe">The transformation pipe that may or may not include a scorer.</param> <param name="ignoreMissingColumns">Whether to ignore missing columns in the data view.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <param name="outputSchemaDefinition">The optional output schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TDst"/> type.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateSimplePredictionEngine(Microsoft.ML.Runtime.IHostEnvironment,System.IO.Stream,System.Int32)"> <summary> Create a prediction engine. This encapsulates the 'classic' prediction problem, where the input is denoted by the float array of features, and the output is a float score. For binary classification predictors that can output probability, there are output fields that report the predicted label and probability. </summary> <param name="env">The host environment to use.</param> <param name="modelStream">The model stream to load pipeline from.</param> <param name="nFeatures">Number of features.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.LoadTransforms(Microsoft.ML.Runtime.IHostEnvironment,System.IO.Stream,Microsoft.ML.Runtime.Data.IDataView)"> <summary> Load the transforms (but not loader) from the model steram and apply them to the specified data. It is acceptable to have no transforms in the model stream: in this case the original <paramref name="data"/> will be returned. </summary> <param name="env">The host environment to use.</param> <param name="modelStream">The model stream to load from.</param> <param name="data">The data to apply transforms to.</param> <returns>The transformed data.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateLoader``1(Microsoft.ML.Runtime.IHostEnvironment,``0,Microsoft.ML.Runtime.Data.IMultiStreamSource)"> <summary> Creates a data loader from the arguments object. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateLoader(Microsoft.ML.Runtime.IHostEnvironment,System.String,Microsoft.ML.Runtime.Data.IMultiStreamSource)"> <summary> Creates a data loader from the 'LoadName{settings}' string. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateSaver``1(Microsoft.ML.Runtime.IHostEnvironment,``0)"> <summary> Creates a data saver from the arguments object. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateSaver(Microsoft.ML.Runtime.IHostEnvironment,System.String)"> <summary> Creates a data saver from the 'LoadName{settings}' string. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateTransform``1(Microsoft.ML.Runtime.IHostEnvironment,``0,Microsoft.ML.Runtime.Data.IDataView)"> <summary> Creates a data transform from the arguments object. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateTransform(Microsoft.ML.Runtime.IHostEnvironment,System.String,Microsoft.ML.Runtime.Data.IDataView)"> <summary> Creates a data transform from the 'LoadName{settings}' string. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateScorer(Microsoft.ML.Runtime.IHostEnvironment,System.String,Microsoft.ML.Runtime.Data.RoleMappedData,Microsoft.ML.Runtime.Api.Predictor,Microsoft.ML.Runtime.Data.RoleMappedSchema)"> <summary> Creates a data scorer from the 'LoadName{settings}' string. </summary> <param name="env">The host environment to use.</param> <param name="settings">The settings string.</param> <param name="data">The data to score.</param> <param name="predictor">The predictor to score.</param> <param name="trainSchema">The training data schema from which the scorer can optionally extract additional information, e.g., label names. If this is <c>null</c>, no information will be extracted.</param> <returns>The scored data.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.CreateDefaultScorer(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.RoleMappedData,Microsoft.ML.Runtime.Api.Predictor,Microsoft.ML.Runtime.Data.RoleMappedSchema)"> <summary> Creates a default data scorer appropriate to the predictor's prediction kind. </summary> <param name="env">The host environment to use.</param> <param name="data">The data to score.</param> <param name="predictor">The predictor to score.</param> <param name="trainSchema">The training data schema from which the scorer can optionally extract additional information, e.g., label names. If this is <c>null</c>, no information will be extracted.</param> <returns>The scored data.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.ComponentCreation.LoadPredictorOrNull(Microsoft.ML.Runtime.IHostEnvironment,System.IO.Stream)"> <summary> Loads a predictor from the model stream. Returns null iff there's no predictor. </summary> <param name="env">The host environment to use.</param> <param name="modelStream">The model stream.</param> </member> <member name="T:Microsoft.ML.Runtime.Api.DataViewConstructionUtils"> <summary> A helper class to create data views based on the user-provided types. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.DataViewConstructionUtils.DataViewBase`1"> <summary> The base class for the data view over items of user-defined type. </summary> <typeparam name="TRow">The user-defined data type.</typeparam> </member> <member name="T:Microsoft.ML.Runtime.Api.DataViewConstructionUtils.ListDataView`1"> <summary> An in-memory data view based on the IList of data. Supports shuffling. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.DataViewConstructionUtils.StreamingDataView`1"> <summary> An in-memory data view based on the IEnumerable of data. Doesn't support shuffling. This class is public because prediction engine wants to call its <see cref="M:Microsoft.ML.Runtime.Api.DataViewConstructionUtils.StreamingDataView`1.SetData(System.Collections.Generic.IEnumerable{`0})"/> for performance reasons. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.DataViewConstructionUtils.StreamingDataView`1.SetData(System.Collections.Generic.IEnumerable{`0})"> <summary> Since all the cursors only depend on an enumerator (rather than the data itself), it's safe to 'swap' the data inside the streaming data view. This doesn't affect the current 'live' cursors, only the ones that will be created later. This is used for efficiency in <see cref="T:Microsoft.ML.Runtime.Api.BatchPredictionEngine`2"/>. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.DataViewConstructionUtils.SingleRowLoopDataView`1"> <summary> This represents the 'infinite data view' over one (mutable) user-defined object. The 'current row' object can be updated at any time, this will affect all the newly created cursors, but not the ones already existing. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.MetadataInfo"> <summary> A single instance of metadata information, associated with a column. </summary> </member> <member name="F:Microsoft.ML.Runtime.Api.MetadataInfo.MetadataType"> <summary> The type of the metadata. </summary> </member> <member name="F:Microsoft.ML.Runtime.Api.MetadataInfo.Kind"> <summary> The string identifier of the metadata. Some identifiers have special meaning, like "SlotNames", but any other identifiers can be used. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.MetadataInfo`1"> <summary> Strongly-typed version of <see cref="T:Microsoft.ML.Runtime.Api.MetadataInfo"/>, that contains the actual value of the metadata. </summary> <typeparam name="T">Type of the metadata value.</typeparam> </member> <member name="M:Microsoft.ML.Runtime.Api.MetadataInfo`1.#ctor(System.String,`0,Microsoft.ML.Runtime.Data.ColumnType)"> <summary> Constructor for metadata of value type T. </summary> <param name="kind">The string identifier of the metadata. Some identifiers have special meaning, like "SlotNames", but any other identifiers can be used.</param> <param name="value">Metadata value.</param> <param name="metadataType">Type of the metadata.</param> </member> <member name="T:Microsoft.ML.Runtime.Api.GenerateCodeCommand"> <summary> Generates the sample prediction code for a given model file, with correct input and output classes. REVIEW: Consider adding support for generating VBuffers instead of arrays, maybe for high dimensionality vectors. </summary> </member> <member name="F:Microsoft.ML.Runtime.Api.GenerateCodeCommand.Arguments.SparseVectorDeclaration"> <summary> Whether to use the <see cref="T:Microsoft.ML.Runtime.Data.VBuffer`1"/> to represent vector columns (supports sparse vectors). </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.InternalSchemaDefinition"> <summary> An internal class that holds the (already validated) mapping between a custom type and an IDataView schema. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.InternalSchemaDefinition.Column.AssertRep"> <summary> Function that checks whether the InternalSchemaDefinition.Column is a valid one. To be valid, the Column must: 1. Have non-empty values for ColumnName and ColumnType 2. Have a non-empty value for FieldInfo iff it is a field column, else ReturnParameterInfo and Generator iff it is a computed column 3. Generator must have the method inputs (TRow rowObject, long position, ref TValue outputValue) in that order. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.InternalSchemaDefinition.GetVectorAndKind(System.Reflection.FieldInfo,System.Boolean@,Microsoft.ML.Runtime.Data.DataKind@)"> <summary> Given a field info on a type, returns whether this appears to be a vector type, and also the associated data kind for this type. If a data kind could not be determined, this will throw. </summary> <param name="fieldInfo">The field info to inspect.</param> <param name="isVector">Whether this appears to be a vector type.</param> <param name="kind">The data kind of the type, or items of this type if vector.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.InternalSchemaDefinition.GetVectorAndKind(System.Reflection.ParameterInfo,System.Boolean@,Microsoft.ML.Runtime.Data.DataKind@)"> <summary> Given a parameter info on a type, returns whether this appears to be a vector type, and also the associated data kind for this type. If a data kind could not be determined, this will throw. </summary> <param name="parameterInfo">The parameter info to inspect.</param> <param name="isVector">Whether this appears to be a vector type.</param> <param name="kind">The data kind of the type, or items of this type if vector.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.InternalSchemaDefinition.GetVectorAndKind(System.Type,System.String,System.Boolean@,Microsoft.ML.Runtime.Data.DataKind@)"> <summary> Given a type and name for a variable, returns whether this appears to be a vector type, and also the associated data kind for this type. If a data kind could not be determined, this will throw. </summary> <param name="rawType">The type of the variable to inspect.</param> <param name="name">The name of the variable to inspect.</param> <param name="isVector">Whether this appears to be a vector type.</param> <param name="kind">The data kind of the type, or items of this type if vector.</param> </member> <member name="T:Microsoft.ML.Runtime.Api.LambdaTransform"> <summary> Utility class for creating transforms easily. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.LambdaTransform.LoadDelegate"> <summary> A delegate type to create a persistent transform, utilized by the creation functions as a callback to reconstitute a transform from binary data. </summary> <param name="reader">The binary stream from which the load is persisted</param> <param name="env">The host environment</param> <param name="input">The dataview this transform should be persisted on</param> <returns>A transform of the input data, as parameterized by the binary input stream</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.LambdaTransform.CreateMap``2(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Action{``0,``1},Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> This creates a transform that generates additional columns to the provided <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>. It does not change the number of rows, and can be seen as a result of application of the user's function to every row of the input data. Similarly to the existing <see cref="T:Microsoft.ML.Runtime.Data.IDataTransform"/>'s, this object can be treated as both the 'transformation' algorithm (which can be then applied to different data by calling <see cref="M:Microsoft.ML.Runtime.Data.ITransformTemplate.ApplyToData(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView)"/>), and the transformed data (which can be enumerated upon by calling <c>GetRowCursor</c> or <c>AsCursorable{TRow}</c>). If <typeparamref name="TSrc"/> or <typeparamref name="TDst"/> implement the <see cref="T:System.IDisposable" /> interface, they will be disposed after use. This is a 'stateless non-savable' version of the transform. </summary> <param name="env">The host environment to use.</param> <param name="source">The input data to apply transformation to.</param> <param name="mapAction">The function that performs the transformation. This must be a 'pure' function which should only depend on the current <typeparamref name="TSrc"/> object and be re-entrant: the system may call the function on some (or all) the input rows, may do this in any order and from multiple threads at once. The function may utilize closures, as long as this is done in re-entrant fashion and without side effects.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <param name="outputSchemaDefinition">The optional output schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TDst"/> type.</param> <typeparam name="TSrc">The type that describes what 'source' columns are consumed from the input <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</typeparam> <typeparam name="TDst">The type that describes what new columns are added by this transform.</typeparam> <returns>A non-savable mapping transform from the input source to the destination</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.LambdaTransform.CreateMap``2(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Action{``0,``1},System.Action{System.IO.BinaryWriter},Microsoft.ML.Runtime.Api.LambdaTransform.LoadDelegate,Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> This creates a transform that generates additional columns to the provided <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>. It does not change the number of rows, and can be seen as a result of application of the user's function to every row of the input data. Similarly to the existing <see cref="T:Microsoft.ML.Runtime.Data.IDataTransform"/>'s, this object can be treated as both the 'transformation' algorithm (which can be then applied to different data by calling <see cref="M:Microsoft.ML.Runtime.Data.ITransformTemplate.ApplyToData(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView)"/>), and the transformed data (which can be enumerated upon by calling <c>GetRowCursor</c> or <c>AsCursorable{TRow}</c>). If <typeparamref name="TSrc"/> or <typeparamref name="TDst"/> implement the <see cref="T:System.IDisposable" /> interface, they will be disposed after use. This is a 'stateless savable' version of the transform: save and load routines must be provided. </summary> <param name="env">The host environment to use.</param> <param name="source">The input data to apply transformation to.</param> <param name="mapAction">The function that performs the transformation. This must be a 'pure' function which should only depend on the current <typeparamref name="TSrc"/> object and be re-entrant: the system may call the function on some (or all) the input rows, may do this in any order and from multiple threads at once. The function may utilize closures, as long as this is done in re-entrant fashion and without side effects.</param> <param name="saveAction">An action that allows us to save state to the serialization stream</param> <param name="loadFunc">A function that given the serialization stream and a data view, returns an <see cref="T:Microsoft.ML.Runtime.Data.ITransformTemplate"/>. The intent is, this returned object should itself be the same as if we had recreated it using this method, but this is impossible to enforce. This transform will do its best to save a description of this method through assembly qualified names of the defining class, method name, and generic type parameters (if any), and then recover this same method on load, so it should be a static non-lambda method that this assembly can legally call.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <param name="outputSchemaDefinition">The optional output schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TDst"/> type.</param> <typeparam name="TSrc">The type that describes what 'source' columns are consumed from the input <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</typeparam> <typeparam name="TDst">The type that describes what new columns are added by this transform.</typeparam> <returns>A savable mapping transform from the input source to the destination</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.LambdaTransform.CreateMap``3(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Action{``0,``1,``2},System.Action{``2},Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> This is a 'stateful non-savable' version of the map transform: the mapping function is guaranteed to be invoked once per every row of the data set, in sequence; one user-defined state object will be allocated per cursor and passed to the map function every time. If <typeparamref name="TSrc"/>, <typeparamref name="TDst"/>, or <typeparamref name="TState"/> implement the <see cref="T:System.IDisposable" /> interface, they will be disposed after use. </summary> <typeparam name="TSrc">The type that describes what 'source' columns are consumed from the input <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</typeparam> <typeparam name="TState">The type of the state object to allocate per cursor.</typeparam> <typeparam name="TDst">The type that describes what new columns are added by this transform.</typeparam> <param name="env">The host environment to use.</param> <param name="source">The input data to apply transformation to.</param> <param name="mapAction">The function that performs the transformation. The function should transform its <typeparamref name="TSrc"/> argument into its <typeparamref name="TDst"/> argument and can utilize the per-cursor <typeparamref name="TState"/> state.</param> <param name="initStateAction">The function that is called once per cursor to initialize state. Can be null.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <param name="outputSchemaDefinition">The optional output schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TDst"/> type.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.LambdaTransform.CreateMap``3(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Action{``0,``1,``2},System.Action{``2},System.Action{System.IO.BinaryWriter},Microsoft.ML.Runtime.Api.LambdaTransform.LoadDelegate,Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> This is a 'stateful savable' version of the map transform: the mapping function is guaranteed to be invoked once per every row of the data set, in sequence (non-parallelizable); one user-defined state object will be allocated per cursor and passed to the map function every time; save and load routines must be provided. If <typeparamref name="TSrc"/>, <typeparamref name="TDst"/>, or <typeparamref name="TState"/> implement the <see cref="T:System.IDisposable" /> interface, they will be disposed after use. </summary> <typeparam name="TSrc">The type that describes what 'source' columns are consumed from the input <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</typeparam> <typeparam name="TState">The type of the state object to allocate per cursor.</typeparam> <typeparam name="TDst">The type that describes what new columns are added by this transform.</typeparam> <param name="env">The host environment to use.</param> <param name="source">The input data to apply transformation to.</param> <param name="mapAction">The function that performs the transformation. The function should transform its <typeparamref name="TSrc"/> <param name="initStateAction">The function that is called once per cursor to initialize state. Can be null.</param> <param name="saveAction">An action that allows us to save state to the serialization stream</param> <param name="loadFunc">A function that given the serialization stream and a data view, returns an <see cref="T:Microsoft.ML.Runtime.Data.ITransformTemplate"/>. The intent is, this returned object should itself be the same as if we had recreated it using this method, but this is impossible to enforce. This transform will do its best to save a description of this method through assembly qualified names of the defining class, method name, and generic type parameters (if any), and then recover this same method on load, so it should be a static non-lambda method that this assembly can legally call.</param> argument into its <typeparamref name="TDst"/> argument and can utilize the per-cursor <typeparamref name="TState"/> state.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <param name="outputSchemaDefinition">The optional output schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TDst"/> type.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.LambdaTransform.CreateFilter``2(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Func{``0,``1,System.Boolean},System.Action{``1},Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> This creates a filter transform that can 'accept' or 'decline' any row of the data based on the contents of the row or state of the cursor. This is a 'stateful non-savable' version of the filter: the filter function is guaranteed to be invoked once per every row of the data set, in sequence (non-parallelizable); one user-defined state object will be allocated per cursor and passed to the filter function every time. If <typeparamref name="TSrc"/> or <typeparamref name="TState"/> implement the <see cref="T:System.IDisposable" /> interface, they will be disposed after use. </summary> <typeparam name="TSrc">The type that describes what 'source' columns are consumed from the input <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</typeparam> <typeparam name="TState">The type of the state object to allocate per cursor.</typeparam> <param name="env">The host environment to use.</param> <param name="source">The input data to apply transformation to.</param> <param name="filterFunc">The user-defined function that determines whether to keep the row or discard it. First parameter is the current row's contents, the second parameter is the cursor-specific state object.</param> <param name="initStateAction">The function that is called once per cursor to initialize state. Can be null.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <returns></returns> </member> <member name="M:Microsoft.ML.Runtime.Api.LambdaTransform.CreateFilter``2(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Func{``0,``1,System.Boolean},System.Action{``1},System.Action{System.IO.BinaryWriter},Microsoft.ML.Runtime.Api.LambdaTransform.LoadDelegate,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> This creates a filter transform that can 'accept' or 'decline' any row of the data based on the contents of the row or state of the cursor. This is a 'stateful savable' version of the filter: the filter function is guaranteed to be invoked once per every row of the data set, in sequence (non-parallelizable); one user-defined state object will be allocated per cursor and passed to the filter function every time; save and load routines must be provided. If <typeparamref name="TSrc"/> or <typeparamref name="TState"/> implement the <see cref="T:System.IDisposable" /> interface, they will be disposed after use. </summary> <typeparam name="TSrc">The type that describes what 'source' columns are consumed from the input <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</typeparam> <typeparam name="TState">The type of the state object to allocate per cursor.</typeparam> <param name="env">The host environment to use.</param> <param name="source">The input data to apply transformation to.</param> <param name="filterFunc">The user-defined function that determines whether to keep the row or discard it. First parameter is the current row's contents, the second parameter is the cursor-specific state object.</param> <param name="initStateAction">The function that is called once per cursor to initialize state. Can be null.</param> <param name="saveAction">An action that allows us to save state to the serialization stream</param> <param name="loadFunc">A function that given the serialization stream and a data view, returns an <see cref="T:Microsoft.ML.Runtime.Data.ITransformTemplate"/>. The intent is, this returned object should itself be the same as if we had recreated it using this method, but this is impossible to enforce. This transform will do its best to save a description of this method through assembly qualified names of the defining class, method name, and generic type parameters (if any), and then recover this same method on load, so it should be a static non-lambda method that this assembly can legally call.</param> <param name="inputSchemaDefinition">The optional input schema. If <c>null</c>, the schema is inferred from the <typeparamref name="TSrc"/> type.</param> <returns></returns> </member> <member name="T:Microsoft.ML.Runtime.Api.LambdaTransformBase"> <summary> Defines common ancestor for various flavors of lambda-based user-defined transforms that may or may not be serializable. In order for the transform to be serializable, the user should specify a save and load delegate. Specifically, for this the user has to provide the following things: * a custom save action that serializes the transform 'state' to the binary writer. * a custom load action that de-serializes the transform from the binary reader. This must be a public static method of a public class. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.LambdaTransformBase.#ctor(Microsoft.ML.Runtime.IHostEnvironment,System.String,Microsoft.ML.Runtime.Api.LambdaTransformBase)"> <summary> The 'reapply' constructor. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.MapTransform`2"> <summary> This transform generates additional columns to the provided <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>. It doesn't change the number of rows, and can be seen as a result of application of the user's function to every row of the input data. Similarly to the existing <see cref="T:Microsoft.ML.Runtime.Data.IDataTransform"/>'s, this object can be treated as both the 'transformation' algorithm (which can be then applied to different data by calling <see cref="M:Microsoft.ML.Runtime.Api.MapTransform`2.ApplyToData(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView)"/>), and the transformed data (which can be enumerated upon by calling <c>GetRowCursor</c> or <c>AsCursorable{TRow}</c>). </summary> <typeparam name="TSrc">The type that describes what 'source' columns are consumed from the input <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</typeparam> <typeparam name="TDst">The type that describes what new columns are added by this transform.</typeparam> </member> <member name="M:Microsoft.ML.Runtime.Api.MapTransform`2.#ctor(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Action{`0,`1},System.Action{System.IO.BinaryWriter},Microsoft.ML.Runtime.Api.LambdaTransform.LoadDelegate,Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Create a a map transform that is savable iff <paramref name="saveAction"/> and <paramref name="loadFunc"/> are not null. </summary> <param name="env">The host environment</param> <param name="source">The dataview upon which we construct the transform</param> <param name="mapAction">The action by which we map source to destination columns</param> <param name="saveAction">An action that allows us to save state to the serialization stream. May be null simultaneously with <paramref name="loadFunc"/>.</param> <param name="loadFunc">A function that given the serialization stream and a data view, returns an <see cref="T:Microsoft.ML.Runtime.Data.ITransformTemplate"/>. The intent is, this returned object should itself be a <see cref="T:Microsoft.ML.Runtime.Api.MapTransform`2"/>, but this is not strictly necessary. This delegate should be a static non-lambda method that this assembly can legally call. May be null simultaneously with <paramref name="saveAction"/>.</param> <param name="inputSchemaDefinition">The schema definition overrides for <typeparamref name="TSrc"/></param> <param name="outputSchemaDefinition">The schema definition overrides for <typeparamref name="TDst"/></param> </member> <member name="M:Microsoft.ML.Runtime.Api.MapTransform`2.#ctor(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Api.MapTransform{`0,`1},Microsoft.ML.Runtime.Data.IDataView)"> <summary> The 'reapply' constructor. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.MergedSchema"> <summary> This is the schema that is obtained by adding columns from <see cref="T:Microsoft.ML.Runtime.Api.InternalSchemaDefinition"/> to the existing schema with auto-hiding. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.BatchPredictionEngine`2"> <summary> A class that runs the previously trained model (and the preceding transform pipeline) on the in-memory data in batch mode. This can also be used with trained pipelines that do not end with a predictor: in this case, the 'prediction' will be just the outcome of all the transformations. </summary> <typeparam name="TSrc">The user-defined type that holds the example.</typeparam> <typeparam name="TDst">The user-defined type that holds the prediction.</typeparam> </member> <member name="M:Microsoft.ML.Runtime.Api.BatchPredictionEngine`2.Predict(System.Collections.Generic.IEnumerable{`0},System.Boolean)"> <summary> Run the prediction pipe. This will enumerate the <paramref name="examples"/> exactly once, cache all the examples (by reference) into its internal representation and then run the transformation pipe. </summary> <param name="examples">The examples to run the prediction on.</param> <param name="reuseRowObjects">If <c>true</c>, the engine will not allocate memory per output, and the returned <typeparamref name="TDst"/> objects will actually always be the same object. The user is expected to clone the values himself if needed.</param> <returns>The <see cref="T:System.Collections.Generic.IEnumerable`1"/> that contains all the pipeline results.</returns> </member> <member name="T:Microsoft.ML.Runtime.Api.PipeEngine`1"> <summary> Utility class to run the pipeline to completion and produce a strongly-typed IEnumerable as a result. Doesn't allocate memory for every row: instead, yields the same row object on every step. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.PredictionEngine`2"> <summary> A class that runs the previously trained model (and the preceding transform pipeline) on the in-memory data, one example at a time. This can also be used with trained pipelines that do not end with a predictor: in this case, the 'prediction' will be just the outcome of all the transformations. This is essentially a wrapper for <see cref="T:Microsoft.ML.Runtime.Api.BatchPredictionEngine`2"/> that throws if more than one result is returned per call to <see cref="M:Microsoft.ML.Runtime.Api.PredictionEngine`2.Predict(`0)"/>. </summary> <typeparam name="TSrc">The user-defined type that holds the example.</typeparam> <typeparam name="TDst">The user-defined type that holds the prediction.</typeparam> </member> <member name="M:Microsoft.ML.Runtime.Api.PredictionEngine`2.Predict(`0)"> <summary> Run prediction pipeline on one example. </summary> <param name="example">The example to run on.</param> <returns>The result of prediction. A new object is created for every call.</returns> </member> <member name="T:Microsoft.ML.Runtime.Api.SimplePredictionEngine"> <summary> This class encapsulates the 'classic' prediction problem, where the input is denoted by the float array of features, and the output is a float score. For binary classification predictors that can output probability, there are output fields that report the predicted label and probability. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.SimplePredictionEngine.Prediction"> <summary> The prediction output. For every field, if there are no column with the matched name in the scoring pipeline, the field will be left intact by the engine (and keep 0 as value unless the user code changes it). </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.SimplePredictionEngine.#ctor(Microsoft.ML.Runtime.IHostEnvironment,System.IO.Stream,System.Int32,System.String)"> <summary> Create a prediction engine. </summary> <param name="env">The host environment to use.</param> <param name="modelStream">The model stream to load pipeline from.</param> <param name="nFeatures">Number of features.</param> <param name="featureColumnName">Name of the features column.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.SimplePredictionEngine.Predict(System.Single[])"> <summary> Score an example. </summary> <param name="features">The feature array of the example.</param> <returns>The prediction object. New object is created on every call.</returns> </member> <member name="T:Microsoft.ML.Runtime.Api.Predictor"> <summary> An opaque 'holder' of the predictor, meant to insulate the user from the internal TLC predictor structure, which is subject to change. </summary> </member> <member name="F:Microsoft.ML.Runtime.Api.Predictor.Pred"> <summary> The actual predictor. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.Predictor.GetPredictorObject"> <summary> A way for the user to extract the predictor object and 'delve into the underworld' of unsupported non-API methods. This is needed, for instance, to inspect the weights of a predictor programmatically. The intention is to expose most methods through the API and make usage of this method increasingly unnecessary. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.KeyTypeAttribute"> <summary> Attach to a member of a class to indicate that the item type should be of class key. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.KeyTypeAttribute.Min"> <summary> The minimum key value. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.KeyTypeAttribute.Count"> <summary> The key count, if it is a known cardinality key. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.KeyTypeAttribute.Contiguous"> <summary> Whether keys should be considered to be contiguous. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.KeyTypeAttribute.#ctor"> <summary> Public KeyTypeAttribute constuctor. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.VectorTypeAttribute"> <summary> Allows a member to be marked as a vector valued field, primarily allowing one to set the dimensionality of the resulting array. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.VectorTypeAttribute.Dims"> <summary> The length of the vectors from this vector valued field. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.ColumnAttribute"> <summary> Describes column information such as name and the source columns indicies that this column encapsulates. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.ColumnAttribute.Name"> <summary> Column name. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.ColumnAttribute.Ordinal"> <summary> Contains positions of indices of source columns in the form of ranges. Examples of range: if we want to include just column with index 1 we can write the range as 1, if we want to include columns 1 to 10 then we can write the range as 1-10 and we want to include all the columns from column with index 1 until end then we can write 1-*. This takes sequence of ranges that are comma seperated, example: 1,2-5,10-* </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.ColumnNameAttribute"> <summary> Allows a member to specify its column name directly, as opposed to the default behavior of using the member name as the column name. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.ColumnNameAttribute.Name"> <summary> Column name. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ColumnNameAttribute.#ctor(System.String)"> <summary> Allows one to specify a name to expose this column as, as opposed to simply the field name. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.NoColumnAttribute"> <summary> Mark this member as not being exposed as a column in the schema. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.CursorChannelAttribute"> <summary> Mark a member that implements exactly IChannel as being permitted to receive channel information from an external channel. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.CursorChannelAttribute.TrySetCursorChannel``1(Microsoft.ML.Runtime.IExceptionContext,``0,Microsoft.ML.Runtime.IChannel)"> <summary> When passed some object, and a channel, it attempts to pass the channel to the object. It passes the channel to the object iff the object has exactly one field marked with the CursorChannelAttribute, and that field implements only the IChannel interface. The function returns the modified object, as well as a boolean indicator of whether it was able to pass the channel to the object. </summary> <param name="obj">The object that attempts to acquire the channel.</param> <param name="channel">The channel to pass to the object.</param> <param name="ectx">The exception context.</param> <returns>1. A boolean indicator of whether the channel was sucessfully passed to the object. 2. The object passed in (only modified by the addition of the channel to the field with the CursorChannelAttribute, if the channel was added sucessfully).</returns> </member> <member name="T:Microsoft.ML.Runtime.Api.SchemaDefinition"> <summary> This class defines a schema of a typed data view. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.SchemaDefinition.Column"> <summary> One column of the data view. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.SchemaDefinition.Column.MemberName"> <summary> The name of the member the column is taken from. The API requires this to not be null, and a valid name of a member of the type for which we are creating a schema. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.SchemaDefinition.Column.ColumnName"> <summary> The name of the column that's created in the data view. If this is null, the API uses the <see cref="P:Microsoft.ML.Runtime.Api.SchemaDefinition.Column.MemberName"/>. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.SchemaDefinition.Column.ColumnType"> <summary> The column type. If this is null, the API attempts to derive a type from the member's type. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.SchemaDefinition.Column.IsComputed"> <summary> Whether the column is a computed type. </summary> </member> <member name="P:Microsoft.ML.Runtime.Api.SchemaDefinition.Column.Generator"> <summary> The generator function. if the column is computed. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.SchemaDefinition.Column.AddMetadata``1(System.String,``0,Microsoft.ML.Runtime.Data.ColumnType)"> <summary> Add metadata to the column. </summary> <typeparam name="T">Type of Metadata being added. Types suported as entries in columns are also supported as entries in Metadata. Multiple metadata may be added to one column. </typeparam> <param name="kind">The string identifier of the metadata.</param> <param name="value">Value of metadata.</param> <param name="metadataType">Type of value.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.SchemaDefinition.Column.RemoveMetadata(System.String)"> <summary> Remove metadata from the column if it exists. </summary> <param name="kind">The string identifier of the metadata. </param> </member> <member name="P:Microsoft.ML.Runtime.Api.SchemaDefinition.Column.GetMetadataTypes"> <summary> Returns metadata kind and type associated with this column. </summary> <returns>A dictionary with the kind of the metadata as the key, and the metadata type as the associated value.</returns> </member> <member name="P:Microsoft.ML.Runtime.Api.SchemaDefinition.Item(System.String)"> <summary> Get or set the column definition by column name. If there's no such column: - get returns null, - set adds a new column. If there's more than one column with the same name: - get returns the first column, - set replaces the first column. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.SchemaDefinition.Create(System.Type)"> <summary> Create a schema definition by enumerating all public fields of the given type. </summary> <param name="userType">The type to base the schema on.</param> <returns>The generated schema definition.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.SerializableLambdaTransform.Create(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Model.ModelLoadContext,Microsoft.ML.Runtime.Data.IDataView)"> <summary> Creates an instance of the transform from a context. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.SerializableLambdaTransform.GetSerializedStaticDelegate(Microsoft.ML.Runtime.Api.LambdaTransform.LoadDelegate)"> <summary> Given a single item function that should be a static method, this builds a serialized version of that method that should be enough to "recover" it, assuming it is a "recoverable" method (recoverable here is a loose definition, meaning that <see cref="M:Microsoft.ML.Runtime.Api.SerializableLambdaTransform.DeserializeStaticDelegateOrNull(Microsoft.ML.Runtime.IExceptionContext,System.Byte[],System.Exception@)"/> is capable of creating it, which includes among other things that it's static, non-lambda, accessible to this assembly, etc.). </summary> <param name="func">The method that should be "recoverable"</param> <returns>A string array describing the input method</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.SerializableLambdaTransform.DeserializeStaticDelegateOrNull(Microsoft.ML.Runtime.IExceptionContext,System.Byte[],System.Exception@)"> <summary> This is essentially the inverse function to <see cref="M:Microsoft.ML.Runtime.Api.SerializableLambdaTransform.GetSerializedStaticDelegate(Microsoft.ML.Runtime.Api.LambdaTransform.LoadDelegate)"/>. If the function is not recoverable for any reason, this will return <c>null</c>, and the error parameter will be set. </summary> <param name="ectx">Exception context.</param> <param name="serialized">The serialized bytes, as returned by <see cref="M:Microsoft.ML.Runtime.Api.SerializableLambdaTransform.GetSerializedStaticDelegate(Microsoft.ML.Runtime.Api.LambdaTransform.LoadDelegate)"/></param> <param name="inner">An exception the caller may raise as an inner exception if the return value is <c>null</c>, else, this itself will be <c>null</c></param> <returns>The recovered function wrapping the recovered method, or <c>null</c> if it could not be created, for some reason</returns> </member> <member name="T:Microsoft.ML.Runtime.Api.StatefulFilterTransform`3"> <summary> This transform is similar to <see cref="T:Microsoft.ML.Runtime.Api.MapTransform`2"/>, but it allows per-cursor state, as well as the ability to 'accept' or 'filter out' some rows of the supplied <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>. The downside is that the provided lambda is eagerly called on every row (not lazily when needed), and parallel cursors are not allowed. </summary> <typeparam name="TSrc">The type that describes what 'source' columns are consumed from the input <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>.</typeparam> <typeparam name="TDst">The type that describes what new columns are added by this transform.</typeparam> <typeparam name="TState">The type that describes per-cursor state.</typeparam> </member> <member name="M:Microsoft.ML.Runtime.Api.StatefulFilterTransform`3.#ctor(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Func{`0,`1,`2,System.Boolean},System.Action{`2},System.Action{System.IO.BinaryWriter},Microsoft.ML.Runtime.Api.LambdaTransform.LoadDelegate,Microsoft.ML.Runtime.Api.SchemaDefinition,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Create a filter transform that is savable iff <paramref name="saveAction"/> and <paramref name="loadFunc"/> are not null. </summary> <param name="env">The host environment</param> <param name="source">The dataview upon which we construct the transform</param> <param name="filterFunc">The function by which we transform source to destination columns and decide whether to keep the row.</param> <param name="initStateAction">The function that is called once per cursor to initialize state. Can be null.</param> <param name="saveAction">An action that allows us to save state to the serialization stream. May be null simultaneously with <paramref name="loadFunc"/>.</param> <param name="loadFunc">A function that given the serialization stream and a data view, returns an <see cref="T:Microsoft.ML.Runtime.Data.ITransformTemplate"/>. The intent is, this returned object should itself be a <see cref="T:Microsoft.ML.Runtime.Api.MapTransform`2"/>, but this is not strictly necessary. This delegate should be a static non-lambda method that this assembly can legally call. May be null simultaneously with <paramref name="saveAction"/>.</param> <param name="inputSchemaDefinition">The schema definition overrides for <typeparamref name="TSrc"/></param> <param name="outputSchemaDefinition">The schema definition overrides for <typeparamref name="TDst"/></param> </member> <member name="M:Microsoft.ML.Runtime.Api.StatefulFilterTransform`3.#ctor(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Api.StatefulFilterTransform{`0,`1,`2},Microsoft.ML.Runtime.Data.IDataView)"> <summary> The 'reapply' constructor. </summary> </member> <member name="T:Microsoft.ML.Runtime.Api.IRow`1"> <summary> This interface is an <see cref="T:Microsoft.ML.Runtime.Data.IRow"/> with 'strongly typed' binding. It can populate the user-supplied object's fields with the values of the current row. </summary> <typeparam name="TRow">The user-defined type that is being populated while cursoring.</typeparam> </member> <member name="M:Microsoft.ML.Runtime.Api.IRow`1.FillValues(`0)"> <summary> Populates the fields of the user-supplied <paramref name="row"/> object with the values of the current row. </summary> <param name="row">The row object. Cannot be null.</param> </member> <member name="T:Microsoft.ML.Runtime.Api.IRowCursor`1"> <summary> This interface provides cursoring through a <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/> via a 'strongly typed' binding. It can populate the user-supplied object's fields with the values of the current row. </summary> <typeparam name="TRow">The user-defined type that is being populated while cursoring.</typeparam> </member> <member name="T:Microsoft.ML.Runtime.Api.ICursorable`1"> <summary> This interface allows to create strongly typed cursors over a <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>. </summary> <typeparam name="TRow">The user-defined type that is being populated while cursoring.</typeparam> </member> <member name="M:Microsoft.ML.Runtime.Api.ICursorable`1.GetCursor"> <summary> Get a new cursor. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.ICursorable`1.GetRandomizedCursor(System.Int32)"> <summary> Get a new randomized cursor. </summary> <param name="randomSeed">The random seed to use.</param> </member> <member name="T:Microsoft.ML.Runtime.Api.TypedCursorable`1"> <summary> Implementation of the strongly typed Cursorable. Similarly to the 'DataView{T}, this class uses IL generation to create the 'poke' methods that write directly into the fields of the user-defined type. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.TypedCursorable`1.IsCompatibleType(Microsoft.ML.Runtime.Data.ColumnType,System.Reflection.FieldInfo)"> <summary> Returns whether the column type <paramref name="colType"/> can be bound to field <paramref name="fieldInfo"/>. They must both be vectors or scalars, and the raw data kind should match. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.TypedCursorable`1.GetCursor"> <summary> Create and return a new cursor. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.TypedCursorable`1.GetRandomizedCursor(System.Int32)"> <summary> Create and return a new randomized cursor. </summary> <param name="randomSeed">The random seed to use.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.TypedCursorable`1.GetCursor(System.Func{System.Int32,System.Boolean},System.Nullable{System.Int32})"> <summary> Create a new cursor with additional active columns. </summary> <param name="additionalColumnsPredicate">Predicate that denotes which additional columns to include in the cursor, in addition to the columns that are needed for populating the <typeparamref name="TRow"/> object.</param> <param name="randomSeed">The random seed to use. If <c>null</c>, the cursor will be non-randomized.</param> </member> <member name="M:Microsoft.ML.Runtime.Api.TypedCursorable`1.GetCursorSet(Microsoft.ML.Runtime.Data.IRowCursorConsolidator@,System.Func{System.Int32,System.Boolean},System.Int32,Microsoft.ML.Runtime.IRandom)"> <summary> Create a set of cursors with additional active columns. </summary> <param name="consolidator">The consolidator for the original row cursors</param> <param name="additionalColumnsPredicate">Predicate that denotes which additional columns to include in the cursor, in addition to the columns that are needed for populating the <typeparamref name="TRow"/> object.</param> <param name="n">Number of cursors to create</param> <param name="rand">Random generator to use</param> </member> <member name="M:Microsoft.ML.Runtime.Api.TypedCursorable`1.Create(Microsoft.ML.Runtime.IHostEnvironment,Microsoft.ML.Runtime.Data.IDataView,System.Boolean,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Create a Cursorable object on a given data view. </summary> <param name="env">Host enviroment.</param> <param name="data">The underlying data view.</param> <param name="ignoreMissingColumns">Whether to ignore missing columns in the data view.</param> <param name="schemaDefinition">The optional user-provided schema.</param> <returns>The constructed Cursorable.</returns> </member> <member name="T:Microsoft.ML.Runtime.Api.CursoringUtils"> <summary> Utility methods that facilitate strongly-typed cursoring. </summary> </member> <member name="M:Microsoft.ML.Runtime.Api.CursoringUtils.AsCursorable``1(Microsoft.ML.Runtime.Data.IDataView,Microsoft.ML.Runtime.IHostEnvironment,System.Boolean,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Generate a strongly-typed cursorable wrapper of the <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>. </summary> <typeparam name="TRow">The user-defined row type.</typeparam> <param name="data">The underlying data view.</param> <param name="env">The environment.</param> <param name="ignoreMissingColumns">Whether to ignore the case when a requested column is not present in the data view.</param> <param name="schemaDefinition">Optional user-provided schema definition. If it is not present, the schema is inferred from the definition of T.</param> <returns>The cursorable wrapper of <paramref name="data"/>.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.CursoringUtils.AsCursorable``1(Microsoft.ML.Runtime.Data.IDataView,System.Boolean,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Generate a strongly-typed cursorable wrapper of the <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/>. </summary> <typeparam name="TRow">The user-defined row type.</typeparam> <param name="data">The underlying data view.</param> <param name="ignoreMissingColumns">Whether to ignore the case when a requested column is not present in the data view.</param> <param name="schemaDefinition">Optional user-provided schema definition. If it is not present, the schema is inferred from the definition of T.</param> <returns>The cursorable wrapper of <paramref name="data"/>.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.CursoringUtils.AsEnumerable``1(Microsoft.ML.Runtime.Data.IDataView,Microsoft.ML.Runtime.IHostEnvironment,System.Boolean,System.Boolean,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Convert an <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/> into a strongly-typed <see cref="T:System.Collections.Generic.IEnumerable`1"/>. </summary> <typeparam name="TRow">The user-defined row type.</typeparam> <param name="data">The underlying data view.</param> <param name="env">The environment.</param> <param name="reuseRowObject">Whether to return the same object on every row, or allocate a new one per row.</param> <param name="ignoreMissingColumns">Whether to ignore the case when a requested column is not present in the data view.</param> <param name="schemaDefinition">Optional user-provided schema definition. If it is not present, the schema is inferred from the definition of T.</param> <returns>The <see cref="T:System.Collections.Generic.IEnumerable`1"/> that holds the data in <paramref name="data"/>. It can be enumerated multiple times.</returns> </member> <member name="M:Microsoft.ML.Runtime.Api.CursoringUtils.AsEnumerable``1(Microsoft.ML.Runtime.Data.IDataView,System.Boolean,System.Boolean,Microsoft.ML.Runtime.Api.SchemaDefinition)"> <summary> Convert an <see cref="T:Microsoft.ML.Runtime.Data.IDataView"/> into a strongly-typed <see cref="T:System.Collections.Generic.IEnumerable`1"/>. </summary> <typeparam name="TRow">The user-defined row type.</typeparam> <param name="data">The underlying data view.</param> <param name="reuseRowObject">Whether to return the same object on every row, or allocate a new one per row.</param> <param name="ignoreMissingColumns">Whether to ignore the case when a requested column is not present in the data view.</param> <param name="schemaDefinition">Optional user-provided schema definition. If it is not present, the schema is inferred from the definition of T.</param> <returns>The <see cref="T:System.Collections.Generic.IEnumerable`1"/> that holds the data in <paramref name="data"/>. It can be enumerated multiple times.</returns> </member> </members> </doc> |