tlc.reduction.umap

Dimensionality reduction with the UMAP algorithm.

Module Contents

Classes

Class	Description
UmapReduction	Perform dimensionality reduction on columns of tables using the UMAP algorithm.
UmapTableArgs	Arguments specific to the UMAP reduction method.

API

class UmapReduction(

reducer_args: tlc.reduction.reduction_method.ReducerArgsType | None = None,

)

Bases: tlc.reduction.reduction_method.ReductionMethod[tlc.reduction.umap.UmapTableArgs]

Perform dimensionality reduction on columns of tables using the UMAP algorithm.

Parameters:

reducer_args – A dictionary of arguments which are specific to the reduction method.

apply_reduction_method(

table: Table,

fit_table_url: Url,

column: str,

) → Url | None

default_args() → UmapTableArgs

Returns the default arguments for the UMAP reduction method.

fit_reduction_method(

table: Table,

column: str,

) → Url

Fits a UmapTable and returns the model URL

class UmapTableArgs

Bases: tlc.reduction.reduction_method.ReducerArgs

Arguments specific to the UMAP reduction method.

Pass an instance of this dict to reduce_embeddings() (or one of its variants) via the reducer_args parameter to control the UMAP reduction.

Initialize self. See help(type(self)) for accurate signature.

metric: str = None

The metric to use when calculating distances between points in the input space (default: euclidean).

min_dist: float = None

The minimum distance between points in the reduced space (default 0.1).

n_components: int = None

The number of dimensions in the reduced space (default 2).

n_jobs: int = None

The number of threads to use for the reduction (default -1).

n_neighbors: int = None

The number of neighbors to use when constructing the nearest neighbor graph (default 15).

standard_scaler_normalize: bool = None

Whether to normalize the data before reducing it (default False).