Changes in version 4.2.2                        

  - Hardened the bootstrap workflow for lcc() and related summaries,
    ensuring parallel workers reliably load the modelling internals.
  - Unified the bootstrap metric outputs so LCC, Pearson correlation,
    and accuracy share a consistent structure across serial and parallel
    runs.
  - Added documented bootstrap confidence intervals to the user workflow
    and refreshed package documentation accordingly.

New Features

  - Support for multiple bootstrap schemes, reproducible seeding, and
    normal, percentile, or BCa confidence intervals through a unified
    builder.
  - Expanded plotting controls, including interval styling, point alpha,
    and axis scaling configured via plotControl helpers.
  - New package options (boot.scheme, ci.method, keep.boot.models,
    boot.seed, numCore) to fine-tune bootstrap execution and
    persistence.

Bug Fixes

  - Added stronger input validation, improved numerical stability in
    degenerate scenarios, and harmonised metric list shapes for
    downstream consumers.

Documentation and Tests

  - Updated README, reference manuals, and CRAN notes to reflect the
    enhanced bootstrap and plotting workflows.
  - Added focused tests covering bootstrap dimensions, metric integrity,
    and plot control behaviour.

Compatibility

  - No breaking user-facing changes; existing APIs continue to work,
    with the major version signalling the enhanced bootstrap
    capabilities and infrastructure improvements.