Changelog

Source: NEWS.md

graphicalMCP 0.1.0

• First release

graphicalMCP 0.1.1

• Added compilation of vignettes (#73)
• Removed duplicated columns of “*” in test values (#75)

graphicalMCP 0.1.2

• Updated function documentation
• Updated references

graphicalMCP 0.2.0

• Corrected typos
• Updated function documentation for CRAN release

graphicalMCP 0.2.1

• Corrected typos
• First CRAN release

graphicalMCP 0.2.2

• Updated citations
• Resubmission for first CRAN release

graphicalMCP 0.2.3

• Included cran-comments.ms in .Rbuildignore
• Resubmission for first CRAN release

graphicalMCP 0.2.4

• Updated documentation according to issue #84
• Resubmission for first CRAN release

graphicalMCP 0.2.5

CRAN release: 2024-07-13

• Updated adjust_weights_parametric_util.Rd
• Resubmission for first CRAN release

graphicalMCP 0.2.6

CRAN release: 2024-11-08

• Github repo transferred to openpharma
• Submission for CRAN release

graphicalMCP 0.2.7

• Added Hochberg-based procedures
• Added internal validations
• Expanded example graphs
• Updated vignettes
• Submission for CRAN release

graphicalMCP 0.2.8

CRAN release: 2025-05-07

• Corrected urls for references
• Submission for CRAN release

graphicalMCP 0.2.9

CRAN release: 2026-03-21

• Closed Issue #90 by changing the precision for parametric tests
• Submission for CRAN release

graphicalMCP 0.3.0

CRAN release: 2026-08-27

• Added group sequential testing for graphical MCPs via graph_test_shortcut_gsd(), which extends graph_test_shortcut() to multiple analyses with per-hypothesis spending functions, information fractions, and a look_back option to use repeated or sequential p-values
• Added spending functions spending_of(), spending_pocock(), spending_hsd(), spending_linear(), spending_wt() (Wang-Tsiatis), and spending_with_time() for a custom spending time
• Added gs_boundaries(), gs_corr(), repeated_p(), and sequential_p()
• Added vignettes on group sequential testing and on its validation against gsDesign and rpact
• Renamed the Adj. P-value column to Adj.p in print.graph_report()
• Added a warning in graph_create() for very small transition weights (Issue #97)
• Plotting examples, vignettes, and tests are skipped when igraph or gMCP is not installed

graphicalMCP (development version)

• Gave an informative error when a graph has lost or has inconsistent hypothesis names (e.g. after graph$hypotheses <- c(0.6, 0.4)), instead of a cryptic downstream failure (Issue #110)
• Fixed graph_calculate_power() for boundary marginal power values of 0 or 1: Bonferroni testing no longer hangs, and Hochberg testing no longer over-rejects, when a simulated p-value is exactly 0 or 1 (Issue #47)
• graph_calculate_power() now requires alpha to be strictly between 0 and 1, since marginal power is undefined at alpha of 0 or 1 (previously these produced an error or silently wrong power) (Issue #47)