Changes in version 0.2.1 (2026-03-26)                  

  - Fixed bug where lancaster() was not exported, causing
    match.fun("lancaster") to fail for installed package users
  - Replaced fragile match.fun() lookup with internal resolution in
    compute_p_values()
  - Added aggregation_method parameter to estimate_changepoint() for
    consistency with compute_p_values()
  - Exported lancaster() with documentation
  - Added tests for p-value aggregation methods

                 Changes in version 0.2.0 (2026-03-04)                  

  - Added acat() function for p-value aggregation
  - Added aggregation_method parameter to compute_p_values() to allow
    users to choose between "lancaster" (default) and "acat" methods
  - Improved performance by reusing GAM design matrices in
    compute_p_values() and estimate_changepoint()
  - Changed compute_mvn default to FALSE in estimate_changepoint() to
    skip costly multivariate normal simulation by default
  - Fixed tidyselect deprecation warnings

                 Changes in version 0.1.3 (2025-03-13)                  

  - First release of cpam on CRAN.
  - Implements changepoint detection, smooth and shape-constrained
    trends for time series omics data.
  - Supports transcript- and gene-level analysis, case-only and
    case-control studies, and uncertainty quantification.
  - Includes an interactive 'shiny' interface.