falsifyr 1.0.0
First CRAN release
- Added the claim-first
attack(model, term) engine and
ranked smallest-kill output.
- Added row deletion (ranked, greedy, beam, and grouped), uncertainty,
covariate, missingness, measurement-error, placebo, specification, and
split attack families.
- Added support for
lm, glm,
aov, lmerMod, glmerMod, and
coxph perturbation workflows, plus explicit limited support
for htest and anova objects that do not retain
refittable source data.
- Added optional
broom::tidy() and
broom::glance() methods for attack leaderboards and claim
summaries.
- Added deterministic fast/normal/deep/insane intensity profiles,
claim and attack registries, and two-worker parallel family
execution.
- Added survival scoring, verdicts, smallest-kill and leaderboard
accessors, console/plot output, standalone HTML reports, and the RStudio
addin.
- Added synthetic fragile/resilient data, vignettes, a GitHub-only
validation corpus, and cross-platform R CMD check automation.