aboutsummaryrefslogtreecommitdiff
path: root/inst/rmarkdown/templates/hierarchical_kinetics/skeleton/skeleton.Rmd
diff options
context:
space:
mode:
authorJohannes Ranke <jranke@uni-bremen.de>2023-04-16 13:31:14 +0200
committerJohannes Ranke <jranke@uni-bremen.de>2023-04-16 13:31:14 +0200
commitc99a12f1ccb03b46264c536bc9ded5d5de9f85df (patch)
tree11c14cc54ae0b93df723b3e06ce3a0b2eba39018 /inst/rmarkdown/templates/hierarchical_kinetics/skeleton/skeleton.Rmd
parent76bc0fead33a63168a5e8d77a6c4b82426815d70 (diff)
Make the template work on Windows (with RStudio)
It seems that we cannot use stored DLLs (for pathway fits) on clusters when using Windows, where we use PSOCK clusters. Although it works on winbuilder in the tests in tests/testthat/test_compiled.R. Maybe it is something RStudio specific.
Diffstat (limited to 'inst/rmarkdown/templates/hierarchical_kinetics/skeleton/skeleton.Rmd')
-rw-r--r--inst/rmarkdown/templates/hierarchical_kinetics/skeleton/skeleton.Rmd12
1 files changed, 8 insertions, 4 deletions
diff --git a/inst/rmarkdown/templates/hierarchical_kinetics/skeleton/skeleton.Rmd b/inst/rmarkdown/templates/hierarchical_kinetics/skeleton/skeleton.Rmd
index 71d62677..0baf866a 100644
--- a/inst/rmarkdown/templates/hierarchical_kinetics/skeleton/skeleton.Rmd
+++ b/inst/rmarkdown/templates/hierarchical_kinetics/skeleton/skeleton.Rmd
@@ -220,9 +220,13 @@ Separate evaluations of all datasets are performed with constant variance
and using two-component error.
```{r path-1-sep, dependson = c("path-1-degmod", "ds")}
-cl <- start_cluster(n_cores)
+cl_path <- NULL
+# Uncomment the next line for parallel processing. This does not work on Windows,
+# as stored dlls (that we need for caching to work) do not seem to work on
+# PSOCK clusters generated on Windows.
+# cl_path <- start_cluster(n_cores)
sforb_sep_const <- mmkin(list(sforb_path = m_sforb_sfo2), ds,
- cluster = cl, quiet = TRUE)
+ cluster = cl_path, quiet = TRUE)
sforb_sep_tc <- update(sforb_sep_const, error_model = "tc")
```
@@ -240,7 +244,7 @@ model that was identified for the parent degradation, are attempted below.
path_1 <- mhmkin(list(sforb_sep_const, sforb_sep_tc),
no_random_effect = c("lambda_free_0", "log_k_lambda_free_bound"),
covariates = covariates, covariate_models = list(log_k_lambda_bound_free ~ pH),
- cluster = cl)
+ cluster = cl_path)
```
```{r dependson = "path-1"}
@@ -294,7 +298,7 @@ endpoints(parent_best_pH_2)
```
```{r}
-stopCluster(cl)
+if (!is.null(cl_path)) stopCluster(cl_path)
```
\clearpage

Contact - Imprint