hdaniel7
commented
5 years ago The current way we've been generating these one-off dashboard comparisons has been simply to select the "knit to PDF" button in Rstudio after changing the riv.seg, run.id, and site_number variables within the gage vs. vahydro or vahydro vs. vahydro Rmarkdown files. It's not the most savvy solution, but it's what we've settled on as good enough when making individual comparisons.
I was unaware that you could generate Rmarkdown files using command line scripts such as
Rscript -e "rmarkdown::render('Working_Gage_Vs_Vahydro_Dashboard_2019.Rmd', clean=TRUE)" RU5_6030_0001 121Instead, the workaround I developed when we were generating dashboards en masse for previous analysis is shown in the R file Automated_Dashboard_Creator_2019.R. This script contains a function, automated_dashboard, which reads in a list of all river segments and then generates a pdf for each of these river segments from "Working_Dashboard_2019.Rmd" (a now-defunct markdown file which generated a cbp model scenario 1 vs. cbp model scenario 2 comparison back when that was the only type of comparison we were doing).
So, in short, my workaround was to render a pdf of the Rmarkdown within a R function which I would run from R on the deq2 server -- this is why those parameters were being passed into the Rmarkdown file via the params variable. This function is broken at the moment (and also useless for the type of analysis we're doing now) -- I could fix and expand upon this function so that it is usable for the type of analysis we are doing now (VA Hydro model run vs VA Hydro model run), but your method of generating the scripts seems to be far more useful for our current analysis -- the only upside of the automated_dashboard function was the ability to generate all the dashboards at once, but that doesn't seem like a capability we need for our current analysis which is more focused on specific river segments.
If you agree, I'll try to troubleshoot some of the issues you've been having with generating these dashboards from the Linux command line so that that option is up and running.
rburghol
Overview
Steps
Previous Work
Wanting to pass the river segment and runid into the dashboard scripts so I could run them from command line in linux, I went exploring. First, to pass in arguments in a regular Rscript, is simple using the () command:
Then I check to load arguments if given. There may be a more elegant way, but for this example I just assumed that the user would know the first argument would be riverseg, and the 2nd (if supplied) would be runid:
To get Rscript to run an Rmd is not quite the same as running a regular .R file. You have to explicitly call knitr:
So, this part works! I did not succeed in running the whole script, but that was expected since there are a host of things to do there. But I did create a functioning example that creates a very simple PDF using command line argument input. See below:
Rscript -e "rmarkdown::render('args.Rmd', clean=TRUE)" RU5_6030_0001 121--- header-includes: \pagenumbering{gobble} output: pdf_document pdf_document: default word_document: default html_document: default params: token: token riv.seg: riv.segtitle: "Test" header-includes:
Try it out and check out the resulting pdf.