Skip to content

Latest commit

 

History

History
180 lines (134 loc) · 6.94 KB

README.md

File metadata and controls

180 lines (134 loc) · 6.94 KB
Raw

VSA_basic_model

The purpose of the VSA Basic Model project is to develop a basic VSA model (VBM1), as an alternative to HRR, FHRR, BSC, MAP, etc. , which is as conceptually simple as possible, and relatively cheap computationally. The motivation for this is that VBM1 would be the default VSA model for my empirical projects, that is, it would be the VSA equivalent of the geneticist’s fruit fly.

All the work is documented in the notebook1 at: https://rgayler.github.io/VSA_basic_model/

Steps I used to make this project from scratch.

They assume that the notebook will be made publicly available via GitHub Pages.

  1. Create a new project in RStudio with renv and a git repository.

  2. Install notestar. This may take a while because renv starts with an empty project library and there are lots of dependencies to install.
    renv::install("tjmahr/notestar")

  3. Create an appropriate license.
    usethis::use_ccby_license().

    • usethis should have been automatically installed as a dependency of notestar.

  4. Create README.Rmd to start documenting these steps.
    usethis::use_readme_rmd()

    • If you’re not planning to create figures in README.Rmd then you should delete the R code chunks. This will stop the directory README_files being created unnecessarily.

  5. Create the notestar directory structure and files.
    notestar::use_notestar()

    • Optionally, edit config.yml if you want to change the file locations.

  6. Configure notestar to use a Makefile.
    notestar:::use_notestar_makefile()

    • The “Build” tab and “Build All” button will not be displayed immediately. They will appear after the project has been closed and re-opened.

  7. Create the reference files, refs.bib and apa.csl.
    notestar::use_notestar_references()

    • WARNING If you use the RStudio visual editor citation tool, the bibliography name in the citation tool defaults to references.bib. You must change the bibliography name in the citation tool to refs.bib.
    • WARNING If you use the RStudio visual editor citation tool to add a citation to a notebook page .Rmd file, the citation tool will add a bibliography: line to the YAML header in that file. This will cause the notestar document build to fail. You must delete the bibliography: line from the YAML header of the notebook page .Rmd file.

  8. Edit _targets.R to set the title, author values, and uncomment the bibliography, and csl lines.

    • Do not edit notebook/index.Rmd (which contains these values)because it is automatically created by targets from the values in _targets.R.

  9. If you are publishing the notebook to GitHub Pages:

    1. Edit _targets.R to add a new target to create a new directory and file ./docs/index.html which is a hard link to the rendered notebook. It needs to be in this location so GitHub Pages can find it. The final list of _targets.R is:

      list(
  targets_main,
  targets_notebook
)

      Edit it to be:

      list(
  targets_main,
  targets_notebook,
  # link rendered notebook to ./docs/index.html for GitHub Pages
  tar_target(
    publish_to_github_pages,
    {
      if(fs::dir_exists("docs")) fs::dir_delete("docs")
      fs::dir_create("docs")
      fs::link_create(tar_read(notebook), "docs/index.html", symbolic = FALSE)
    },
    format = "file"
  )
)

    2. Later, when the GitHub rempte repository has been created, remember to enable GitHub Pages for the repository and to set the source to be the ./docs directory.

  10. Create the first notebook page (i.e dated entry), choosing appropriate values for the date and slug arguments.
    notestar::notebook_create_page(date = "2022-04-03", slug = "design")

    • After the initial setup, this step is where you create new notebook entries.
    • Edit _targets.R and R/functions.R to implement the empirical work to be reported in the notebook page. The assumption is that notebook pages will import (targets::tar_read()) the results of work carried out by targets.

  11. Build the notebook using targets::tar_make() or the “Build All” button (or Ctrl+Shift+B shortcut) in RStudio.

    • After the first build only, create a ./docs directory containing the rendered notebook as the file ./docs/index.html, so the notebook is published on GitHub Pages. ./docs/index.html is created as a hard link to the rendered notebook ./notebook/book/docs/notebook.html. The following shell code assumes the working directory is the project directory.
      mkdir docs
      ln notebook/book/docs/notebook.html docs/index.html

  12. View the notebook using notestar::notebook_browse().

  13. Knit README.Rmd if necessary.

  14. Commit all the files to the local git repository.

  15. Push the local git repository to the remote GitHub repository.

    • On the first time only you can create the remote GitHub repository with usethis::use_github() before pushing.

      • If prompted to store your GitHub PAT in .Renviron - don’t do it.

    • On subsequent occasions just commit and push, as usual.

Iterating on steps 8–14 is the main flow for the notebook. We set up data and modeling things in _targets and R/functions.R, then explore and report them in notebook entries.

Footnotes

  1. The notebook structure is implemented with TJ Mahr’s notestar package. ↩