Open mclements opened 2 months ago
I have added some changes to address the recent R check notes and warnings:
@importFrom R6 R6Class
in ctstm.R -- although this is not strictly necessary.Devin: pkgdown is complaining about the documentation for CohortCtstm in ctstm.R. I can't see the error - would you be able to help here, please?
Some commentary on the implementation:
rehuel
library for CRAN, but found that the package is not well maintained and some of the tests were failing. I have kept the BH dependency. For further discussion?dynamic_cast
to determine (a) if a transmod
was an mstate_list
, and then (b) whether an element of mstate_list->survmod_
is a point mass (after adding an hesim::stats::distribution* get_dist(int trans)
method to mstate_list
). The code would be cleaner and clearer if we did not allow for point masses.I have now fixed the pkgdown issue.
@dincerti, I have now added some tests for this. This is now ready for review -- although admittedly this is a larger pull request than initially anticipated.
Devin:
Devin,
I have been working on a few changes to the hesim package.
For this pull request, I have implemented transition-specific costs, which currently only has an application for iCTSTMs.
Note that this folds in the unresolved inline extension from pull request #122.
As a signpost: I am preparing a presentation for the International Society of Clinical Biostatistics for later this month that implements cCTSTMs using ordinary differential equations. I have a gist and will prepare a pull request. I have a number of questions to ask you about that pull request:).
Sincerely, Mark.