prepare point release: little clean up & fixes

Package: incubate

Title: Parametric Time-to-Event Analysis with Variable Incubation Phases

Version: 1.4.0.9006

Version: 1.4.1

Date: 2026-06-04

Authors@R:

    person(given = "Matthias", family = "Kuhn",

    tidyr,

    withr

URL: https://gitlab.com/imb-dev/incubate/

BugReports: https://gitlab.com/imb-dev/incubate/-/issues/

BugReports: https://gitlab.com/imb-dev/incubate/-/work_items

Language: en-GB

Config/roxygen2/version: 8.0.0

Config/roxygen2/markdown: TRUE

Depends:

    R (>= 3.6.0)

Collate: 

    'S3_integration.R'

    'aaa.R'

    'cpp11.R'

    'data.R'

    'delay.R'

    'delay_test.R'

    'incubate-package.R'

    'integration.R'

    'zzz.R'

VignetteBuilder: knitr

LinkingTo: cpp11

# incubate 1.4.0.9000

* start current development version

# incubate 1.4.1

* `measles_sailer` data: generation numbering now starts at 0 (not 1)

* fixed uniform censoring in `rexp_delayed` and `rweib_delayed`: Lambert W function (implemented in C++) useful for exponential, uniroot approach for Weibull

* enhance documentation

# incubate 1.4.0

* allow for random right-censoring in the response variable

* README: better intro text and more text for example

# incubate 1.1.2

* small changes for CRAN-warnings

* small changes due to CRAN-warnings

* test: bigger data set for delay estimation

# incubate 1.1.1

* bug fix: `plot.incubate_fit` has legible colour legend title

# incubate 1.0.0

* data: add experimental data `stankovic` from a publication Stankovic et al EGFL7 enhances surface expression of integrin doi:10.15252/emmm.201708420 (2018) as example data set

* add data set `stankovic`: experimental data on tumour growth in mice (Stankovic et al, EGFL7 enhances surface expression of integrin (2018))

* This version of package is also published at <https://zenodo.org/record/6462382>

# incubate 0.9.9

#' @param ... further arguments. Not in use here (it is required for generic plot function)

#' @examples

#' # fit a delay-weibull model to serial interval times from historic measles outbreak data:

#' fm <- delay_model(survival::Surv(measles_sailer$serialInterval, measles_sailer$status), distribution = "weibull", method = "MLEw")

#' plot(fm, title = "Serial interval times of measles", subtitle = "Delay-Weibull model fit using weighted MLE (MLEw)")

#' fm <- delay_model(survival::Surv(measles_sailer$serialInterval, measles_sailer$status),

#'                   distribution = "weibull", method = "MLEw")

#' plot(fm,

#'      title = "Serial interval times of measles",

#'      subtitle = "Delay-Weibull model fit using weighted MLE (MLEw)")

#'

#' @export

plot.incubate_fit <- function(x, y, title, subtitle, xlim, ...) {

#'

#' Data stem from a double-blind, placebo-controlled phase 3 trial where n=870

#' patients with completely resected, stage III melanoma with BRAF V600E or

#' V600K mutations were randomly assigned to receive oral dabrafenib plus

#' trametinib (combination therapy, 438 patients) or two matched placebo tablets

#' V600K mutations were randomly assigned to receive oral Dabrafenib plus

#' Trametinib (combination therapy, 438 patients) or two matched placebo tablets

#' (432 patients) for 12 months.

#'

#' @details

      z_y <- if (twoGroup) zF(group = "y", propagateTies = TRUE)

      # How to calculate the weights W1-W3?

      #we use now always 'sdist_median' (also for isSurv-data, cf test-delay_estimation.R, line 802)

      #we use now always 'sdist_median' (also for isSurv-data, cf. test-delay_estimation.R, line 802)

      #method_w <- if (isSurv) "sample" else "sdist_median"

      # Calculates W1-weight (as function of n)

#' For the initial delay parameter we allow to smooth the delay estimate in parametric bootstrap. The initial delay parameter is special because this parameter determines when observations can start to occur.

#' Bootstrapping from a delay model will only produce data that starts never before the given delay. Smoothing adds variability here.

#' The default value for smoothing is `smd_factor=1` which means to add normal noise of 1xstd.dev of the first observation to the first observation.

#' Alternatively, when using the objective function to find a region for delay1 (this branch is turned off in the code, cf USE_OBJFUN hardcoded), the default 1 was an optimal value in a simulation for log-quantile together with log delay-shift = 5.

#' Alternatively, when using the objective function to find a region for delay1 (this branch is turned off in the code, cf. USE_OBJFUN hard-coded), the default 1 was an optimal value in a simulation for log-quantile together with log delay-shift = 5.

#'

#' @param object an `incubate_fit`-object

#' @param bs_data character. Which type of bootstrap method to generate data?

            if (twoGroup) ranFunArgsY[['delay1']] <- delayCandY[ind]

          }

          # cf simulate (but inlined here for performance reasons)

          # cf. simulate (but inlined here for performance reasons)

          x <- rlang::exec(ranFun, !!!ranFunArgsX)

          y <- if (twoGroup) rlang::exec(ranFun, !!!ranFunArgsY)