Docker reproducibility

.gitignore

@@ -46,3 +46,4 @@ rsAbsence.pdf
 
 # some folders
 misc/
+submission/

Dockerfile

 ## set R version (https://hub.docker.com/r/rocker/verse/tags)
-FROM rocker/verse:4.2.3
+FROM rocker/verse:4.3
 
 ## name of the manuscript (as in Makefile and paper/Makefile)
 ENV FILE=rsabsence

Makefile

@@ -13,19 +13,10 @@ dbuild: Dockerfile
 	docker build \
     -t $(FILE) .
 
-## run docker image that produces pdf from within docker
+## run docker image that produces *tex* from within docker
+## and compiles to *pdf* outside docker
 docker: dbuild
 	docker run \
-    --rm \
-	--env pdfdocker="true" \
-	--env FILE=$(FILE) \
-	--volume $(CURDIR):/output \
-	$(FILE)
-	mv $(FILE).pdf ./
-
-## run docker image that produces tex from within docker
-docker2: dbuild
-	docker run \
     --rm \
 	--env FILE=$(FILE) \
 	-v $(CURDIR):/output \
@@ -37,3 +28,13 @@ docker2: dbuild
 	rmdir figure/
 	cd paper && make pdf2 clean
 	mv paper/$(FILE).pdf ./
+
+## run docker image that produces *pdf* from within docker
+docker2: dbuild
+	docker run \
+    --rm \
+	--env pdfdocker="true" \
+	--env FILE=$(FILE) \
+	--volume $(CURDIR):/output \
+	$(FILE)
+	mv $(FILE).pdf ./

README.md

@@ -11,35 +11,41 @@ Our results can be reproduced in two ways
 
 ### 1. Reproducing the results locally
 
-Make sure that LaTeX (e.g., texlive-full on Ubuntu), R, and the R packages
-indicated in the `./CRANpackages.txt` file are installed
+
+Make sure that R, and the R packages indicated in the `./CRANpackages.txt` file
+are installed
 
 ```sh
 ## install packages from CRAN by running from a shell
 R -e 'install.packages(read.delim("CRANpackages.txt", header = FALSE)[,1])'
 ```
 
-Then run 
+To only rerun the R code without recompiling the manuscript use the file
+`./paper/rsabsence.R`. To recompile the manuscript, make also sure that Make,
+LaTeX and all necessary LaTeX packages (e.g., the texlive-full distribution on
+Ubuntu) are installed. Then run
 
 ```
 make local
 ```
 
-this should produce `./rsabsence.pdf`
+this should produce `./rsabsence.pdf`.
+
 
 Although the analyses depend on only few dependencies, this approach may lead to
 different results (or not even compile successfully) in the future if R or the
-packages change. The R and R package versions which were used when the thesis
-was successfully compiled the last time are visible in the following output
+packages change. The R and R package versions which were used when the paper was
+successfully compiled before submission are visible in the following output
 
 ```r
 sessionInfo()
 
-#> R version 4.2.3 (2023-03-15)
+#> R version 4.3.0 (2023-04-21)
 #> Platform: x86_64-pc-linux-gnu (64-bit)
 #> Running under: Ubuntu 20.04.6 LTS
+#> 
 #> Matrix products: default
-#> BLAS:   /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.9.0
+#> BLAS:   /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.9.0 
 #> LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.9.0
 #> 
 #> locale:
@@ -50,40 +56,42 @@ sessionInfo()
 #>  [9] LC_ADDRESS=C               LC_TELEPHONE=C            
 #> [11] LC_MEASUREMENT=de_CH.UTF-8 LC_IDENTIFICATION=C       
 #> 
+#> time zone: Europe/Zurich
+#> tzcode source: system (glibc)
+#> 
 #> attached base packages:
 #> [1] stats     graphics  grDevices utils     datasets  methods   base     
 #> 
 #> other attached packages:
-#> [1] reporttools_1.1.3 xtable_1.8-4      dplyr_1.1.1       gridExtra_2.3    
+#> [1] reporttools_1.1.3 xtable_1.8-4      dplyr_1.1.2       gridExtra_2.3    
 #> [5] ggplot2_3.4.2     knitr_1.42       
 #> 
 #> loaded via a namespace (and not attached):
-#>  [1] fansi_1.0.4      withr_2.5.0      utf8_1.2.3       grid_4.2.3      
-#>  [5] R6_2.5.1         lifecycle_1.0.3  gtable_0.3.3     magrittr_2.0.3  
-#>  [9] scales_1.2.1     pillar_1.9.0     rlang_1.1.0      cli_3.6.1       
-#> [13] generics_0.1.3   vctrs_0.6.1      tools_4.2.3      glue_1.6.2      
-#> [17] munsell_0.5.0    xfun_0.38        compiler_4.2.3   pkgconfig_2.0.3 
-#> [21] colorspace_2.1-0 tidyselect_1.2.0 tibble_3.2.1  
+#>  [1] vctrs_0.6.2       cli_3.6.1         rlang_1.1.0       xfun_0.39        
+#>  [5] generics_0.1.3    labeling_0.4.2    glue_1.6.2        colorspace_2.1-0 
+#>  [9] scales_1.2.1      fansi_1.0.4       grid_4.3.0        munsell_0.5.0    
+#> [13] tibble_3.2.1      lifecycle_1.0.3   compiler_4.3.0    pkgconfig_2.0.3  
+#> [17] farver_2.1.1      viridisLite_0.4.1 R6_2.5.1          tidyselect_1.2.0 
+#> [21] utf8_1.2.3        pillar_1.9.0      magrittr_2.0.3    tools_4.3.0      
+#> [25] withr_2.5.0       gtable_0.3.3
 
 cat(paste(Sys.time(), Sys.timezone(), "\n"))
 
-#> 2023-04-06 13:56:17 Europe/Zurich 
+#> 2023-05-08 09:19:12.440643 Europe/Zurich  
 ```
 
 ### 2. Reproducing the results within a Docker container
 
-Make sure that Docker with root rights is installed. Then run
+Make sure that Make, LaTeX with all necessary packages (e.g., the texlive-full
+distribution on Ubuntu), and Docker with root rights are installed. Then run
 
 ```
 make docker
 ```
 
-This may take some time as TinyTeX needs to install several LaTeX packages (run
-`make docker2` to compile only R code within the container but run LaTeX
-locally). The Docker approach reruns the analyses in a Docker container which
-encapsulates the computational environment (R and R package versions) that was
-used in the original analysis. The only way this approach could become
-non-reproducible is when the
-[rocker/verse](https://hub.docker.com/r/rocker/verse/tags) base image becomes
-unavailable and/or the Posit Public Package Manager (RStudio Package Manager,
-RSPM) snapshot of CRAN becomes unavailable.
+The Docker approach reruns the analyses in a Docker container which encapsulates
+the computational environment (R and R package versions) that was used in the
+original analysis. The only way this approach could become non-reproducible is
+when the [rocker/verse](https://hub.docker.com/r/rocker/verse/tags) base image
+becomes unavailable and/or the Posit Public Package Manager (RStudio Package
+Manager, RSPM) snapshot of CRAN becomes unavailable.