We used the provided docker image with R. We then installed the R packages, but some were missing so we had to install them by hand. We also had to change the working directory:
setwd('shceof')
Missing R packages: Installing R Packages: 'bookdown', 'ggnewscale', 'ggpattern', 'ggperiodic', 'ggrepel', 'ppcor', 'shadowtext', 'widyr'
Missing R package: tagger
devtools::install_github("eliocamp/tagger") > requires devtools
1 of the 2 reviewer used R long time ago, the 2nd one never used R.
We had to install docker to use the docker container.
Docker, Rstudio
The main challenges were:
Knit
a file)The README file describes how to install the package and provides the commands to type. It seems that dependencies versions are locked, which is good for reproducibility.
Maybe providing a command to type in a terminal to reproduce the results and paper would be helpful for people that do not have knowledge of Rstudio, but who are familiar with a terminal.
For me (rcaneill) who do not have experience in Rstudio, I was lost in where to type which command.
The authors took time to write the different steps to recompile the document and reproduce the figures. They provide two different methods to install Rstudio: either via apt on the global Ubuntu-like system, or via docker container. The container is very convenient as it does not pollute the rest of the system.
Include more details of the workflow in the README file i.e. which is the path to the manuscript (Rmd file) and how to set up dependencies and inputs (where to store them) to reproduce results (figures, tables, etc). Also, we suggest documenting workarounds for potential errors e.g. use install.package() for missing dependencies.
No licence is provided for the code.
We spent roughly 1 hour on trying to run the code, which is not very long. We believe that with 1) a little bit more experience in Rstudio and 2) some more time we would have been able to reproduce the results. The fact that the paper itself is written in R-markdown (and hence executes the code at compilation time) gives confidence in the reproducibility of the code itself.