This is a somewhat complete example of the paramvalf structure. We have some
data files in data, some task snippets in paramval and a vignette in
vignettes. The actual .func functions do not do anything as this is just to
show the structure.
To inialize your paramvalf project you need to do create a file that marks the root directory.
touch .paramvalf-root
Then execute the depencency scanner from the R installation once, your path might vary slightly.
python3 ~/R/x86_64-redhat-linux-gnu-library/3.5/paramvalf/scan_dependencies.py
From here on you just need to run this:
./paramvalf-run .
This will run all the parts of the analysis that needs running. As a side
effects it creates output/paramvalf-data-flow.pdf:
It will run all the paramval/*.R files in the order that it needs to and
build all the vignettes. If you only want to generate something in particular
you can also execute something like this:
./paramvalf-run . output/corr_boot.Rdata
The dependencies are determined by looking for pv_load('.', …) and
pv_save('.', …) as well as # Depends: for auxiliary files. See the files in
paramval to see how that loading and saving works exactly.