Struct ColumnTransformer 

pub struct ColumnTransformer { /* private fields */ }

Apply different transformers to different columns of a dataset, mirroring sklearn.compose.ColumnTransformer.

Output columns are ordered as: each spec in insertion order (categorical specs expand to their one-hot columns), followed by passthrough numeric columns (when remainder = Passthrough).

Implementations

impl ColumnTransformer

pub fn new() -> Self

Creates a new empty column transformer.

pub fn remainder(self, r: Remainder) -> Self

Sets how unselected columns are handled.

pub fn add_numeric<S>( self, name: S, columns: Vec<usize>, transformer: TransformerKind, ) -> Self

where S: Into<String>,

Adds a numeric block applying transformer to columns.

pub fn add_categorical<S>( self, name: S, columns: Vec<usize>, encoder: CategoricalTransformerKind, ) -> Self

where S: Into<String>,

Adds a categorical block encoding columns with a categorical CategoricalTransformerKind wrapper.

pub fn add_target<S>( self, name: S, columns: Vec<usize>, encoder: TargetTransformerKind, ) -> Self

where S: Into<String>,

Adds a target (supervised) category encoding block, using encoder on columns. This spec requires target values during fit; call fit_with_target instead of fit.

pub fn fit(&mut self, table: &Table) -> Result<()>

Fits the column specs and remainder encoders to the table.

Returns an error if any ColumnSpec::Target specs are present, because they require target values. Use fit_with_target instead.

pub fn fit_transform_with_target( &mut self, table: &Table, y: &[f64], ) -> Result<Matrix>

Fits the column specs and Target specs using the provided target values, then returns the transformed result.

Non-target specs are fitted via fit (or re-fitted if already fitted). Target specs are fitted with y.

pub fn fit_with_target(&mut self, table: &Table, y: &[f64]) -> Result<()>

Fits the column specs and Target specs using the provided target values.

Non-target specs are fitted via fit. Target specs are additionally fitted with the given target values.

pub fn transform(&self, table: &Table) -> Result<Matrix>

Transforms the table by applying each fitted spec and concatenating results.

pub fn fit_transform(&mut self, table: &Table) -> Result<Matrix>

Fits the transformer to the table and returns the transformed result.

pub fn transform_to_table(&self, table: &Table) -> Result<Output>

Transforms the table and returns an crate::Output that preserves the separation between numeric and categorical columns.

Unlike transform, remainder categorical columns are passed through as strings instead of being one-hot encoded. This is useful when you want to chain further categorical processing.

pub fn fit_transform_to_table(&mut self, table: &Table) -> Result<Output>

Fits then transforms the table, returning an crate::Output that preserves numeric / categorical column separation.

Trait Implementations

impl Clone for ColumnTransformer

fn clone(&self) -> ColumnTransformer

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ColumnTransformer

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for ColumnTransformer

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for ColumnTransformer

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl FeatureNames for ColumnTransformer

fn feature_names_out(&self, input_features: Option<&[String]>) -> Vec<String>

Returns the output feature names given the optional input feature names.

impl Serialize for ColumnTransformer

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.