API Reference

These pages contain the reference documentation for the LensKit modules, classes, etc.

Core Abstractions

lenskit.data	Data abstractions and data set access.
lenskit.pipeline	LensKit pipeline abstraction.
lenskit.diagnostics	Warning and error classes and general LensKit diagnostic code.
lenskit.operations	Top-level LensKit functions for recommender operations.
lenskit.types	Types used across LensKit.

Components and Models

These packages provide various recommendation components and models. Many of them need to be installed separately.

lenskit.basic	Basic and baseline pipeline components.
lenskit.knn	k-NN recommender models.
lenskit.als	LensKit ALS implementations.
lenskit.funksvd	FunkSVD (biased MF).
lenskit.implicit
lenskit.hpf

Batch Inference and Evaluation

These package provide evaluation metrics and analysis and batch-inference support. The evaluation code is not directly linked to the rest of LensKit and can be used to evaluate the output of any recommender system implementation.

lenskit.batch	Batch-run recommendation pipelines for evaluation.
lenskit.metrics	Metrics for evaluating recommender outputs.
lenskit.splitting	Splitting data for train-test evaluation.

Implementation Helpers

These modules provide various utilities and helpers used to implement LensKit, and may be useful in building new models and components for LensKit.

lenskit.logging	Logging, progress, and resource records.
lenskit.math	Mathematical helper routines.
lenskit.parallel	LensKit parallel computation support.
lenskit.util	Miscellaneous utility functions.