Add quarto instructions and R dependency management

[platipodium]

Related openjournals/joss-reviews#8845

I just tried to use the quarto notebooks for downloading data and was puzzeld

• for the lack of instructions what quarto is and how to use it.
• for the lack of dependency management with R.

Now I have used both Quarto and R in the past, but getting this to run is very tedious without instructions.

[danielvartan]

Hi @platipodium,

I have reworked all the procedures related to this issue and documented them in the LogoClim user manual (see chapters 1 to 5).

I believe this addresses what you were looking for. Please let me know if anything is missing or if you would like me to go further on any point.

[jamesdamillington]

@danielvartan The latest additions to the manual and use of renv are useful and resolve these issues for me (although @platipodium raised them).

[platipodium]

renv seems to work for me (but see below), thanks for addressing this. I am looking at the project as a non-R person, so there is still instruction missing in ReadMe on how to run renv

So in this part:

To reproduce the analyses in the manual, install the four dependencies
mentioned above and follow these steps:

1.  **Clone** this repository to your local machine.
2.  **Open** the project in your preferred
    [IDE](https://en.wikipedia.org/wiki/Integrated_development_environment).
3.  **Restore the R environment** by running
    [`renv::restore()`](https://rstudio.github.io/renv/reference/restore.html)
    in the R console. This will install all required package
    dependencies.
4.  **Open** the Quarto notebook files (`.qmd`) and run the code as
    described.

Number (2.) IDE is not really needed, because everything can also be done from a command line. But more important, non-R users cannot know what it means to Restore the R environment.. This is missing:

• Make sure to have R installed in your machine
• Run Rscript -e "renv::restore()" (don't know how to do that in an IDE)

[platipodium]

Also add that libgit2 as a dependency. Wow ... renv is taking > 1 hour (no, really, just 46 minutes)
so far on my Mac M5 :=(

[platipodium]

Add as more dependencies:

• libgit2
• imagemagick
• gdal
• curl
• udunits

Makes me wonder whether renv is such a smart move after all, as its installation procedure is tedious and might throw off users. I was much quicker before manually resolving the R dependencies ...

[danielvartan]

Hi, @platipodium,

Thanks for pointing out these issues.

I already made the changes you requested in the README. I'm revising and improving the documentation.

I'm not sure which IDE you're using, but most don't come with an R console built in. If you're using R in the terminal, I suggest installing radian. It's not quite the same as the R console you'd find in an IDE like Positron, but it works really well.

renv can be a little frustrating at times.

About the libs you mentioned: that's strange. When installing these R packages, the system libraries should be installed automatically. Did you have to install them by hand?

I just tested renv::restore() on a Windows machine and it installed everything in under 2 minutes. I use Arch, so I compile everything from source, but on Ubuntu it's also really fast. I don't have a Mac at hand, only the macOS runners from GitHub Actions, so any help you can give me on this is appreciated.

[danielvartan]

Makes me wonder whether renv is such a smart move after all, as its installation procedure is tedious and might throw off users. I was much quicker before manually resolving the R dependencies ...

renv is not required to use LogoClim. It's only needed if you want to render the user manual or run the unit tests. For data download and processing, only the orbis package is required, as described in the user manual:

https://sustentarea.github.io/logoclim/qmd/data-download.html

[platipodium]

Hi @danielvartan, here's what I did.

1. Start from a fresh conda environment with R conda create -n logoclim R
2. Run Rscript -e "renv::restore()"
3. Wait
4. Wait
5. Get to libgit2 error, manually installl it with conda install libgit2
6. Run Rscript -e "renv::restore()" ... wait
7. iterate through 5/6 for installing with conda more of the dependencies
8. still waiting on renv to finish installing (now really > 1 hour)

[danielvartan]

By using conda, you're running one environment inside another, which is probably what causing the issue.

renv already keeps everything in one place, so there's no need for conda on top of it. If you use Python, think of renv as venv on steroids. The R ecosystem is very well organized, and for more complex packages, installing them outside the R infrastructure can get tricky.