AiModelDataPipeline<T, TInput, TOutput>
Models & Types · AiDotNet.Configuration
Default implementation of IAiModelDataPipeline. Mirrors the
preprocessing / postprocessing / data-loading / data-preparation / augmentation logic that
previously lived inline in AiModelBuilder, with no behavioural change. The facade
continues to be the supported entry point; this class is the audit-2026-05 phase-2a internal
reorganisation that makes the data-pipeline surface testable in isolation and replaceable
without touching the god class.
Implements: IAiModelDataPipeline<T, TInput, TOutput>
Constructors
| Constructor | Summary |
|---|---|
AiModelDataPipeline() |
Properties
| Property | Summary |
|---|---|
AugmentationConfig AugmentationConfig { get; set; } |
|
DataPreparationPipeline<T> DataPreparationPipeline { get; set; } |
|
IDataLoader<T> DataLoader { get; set; } |
|
PostprocessingPipeline<T, TOutput, TOutput> PostprocessingPipeline { get; set; } |
|
PreprocessingPipeline<T, TInput, TInput> PreprocessingPipeline { get; set; } |
|
int? PostprocessingFitMaxRows { get; set; } |
Methods
| Method | Summary |
|---|---|
void AddPreprocessingStep(IDataTransformer<T, TInput, TInput> transformer, string name) |
|
void ConfigureAugmentation(AugmentationConfig config) |
|
void ConfigureAugmentation(AugmentationConfig<T, TInput> config) |
|
void ConfigureDataLoader(IDataLoader<T> dataLoader) |
|
void ConfigureDataPreparation(Action<DataPreparationPipeline<T>> pipelineBuilder) |
|
void ConfigurePostprocessing(Action<PostprocessingPipeline<T, TOutput, TOutput>> pipelineBuilder) |
|
void ConfigurePostprocessing(IDataTransformer<T, TOutput, TOutput> transformer) |
|
void ConfigurePostprocessing(PostprocessingPipeline<T, TOutput, TOutput> pipeline) |
|
void ConfigurePreprocessing(Action<PreprocessingPipeline<T, TInput, TInput>> pipelineBuilder) |
|
void ConfigurePreprocessing(IDataTransformer<T, TInput, TInput> transformer) |
|
void ConfigurePreprocessing(PreprocessingPipeline<T, TInput, TInput> pipeline) |
|
void SetPostprocessingFitMaxRows(int? maxRows) |