All of the logic in https://github.com/nhejazi/biotmle/blob/master/R/biotmle.R#L128-L152 should be re-written for simplification. Far too many options are set, and it's horribly unclear exactly what the parallelization scheme invoked is. In particular, there's not a good reason to support both of the Future and BiocParallel frameworks, and certainly no reason to invoke both. The constructor of biomarkertmle can be significantly simplified accordingly.
nhejazi
commented
3 years ago
All of the logic in https://github.com/nhejazi/biotmle/blob/master/R/biotmle.R#L128-L152 should be re-written for simplification. Far too many options are set, and it's horribly unclear exactly what the parallelization scheme invoked is. In particular, there's not a good reason to support both of the
FutureandBiocParallelframeworks, and certainly no reason to invoke both. The constructor ofbiomarkertmlecan be significantly simplified accordingly.