Misc

Misc

  • Windows

    Shortcut Description
    Ctrl + Tab Change application
    Ctrl + ~ Change window within an application

  • Browser

    Action Shortcut
    To Address Bar Ctrl + L
    Scroll between focus panes (e.g. address to bar to body) f6
    Open a new window Ctrl + n
    Open a new window in Incognito mode Ctrl + Shift + n
    Open a new tab, and jump to it Ctrl + t
    Reopen previously closed tabs in the order they were closed Ctrl + Shift + t
    Jump to the next open tab Ctrl + Tab or Ctrl + PgDn
    Jump to the previous open tab Ctrl + Shift + Tab or Ctrl + PgUp
    Jump to a specific tab Ctrl + 1 through Ctrl + 8
    Jump to the rightmost tab Ctrl + 9
    Open your home page in the current tab Alt + Home
    Open the previous page from your browsing history in the current tab Alt + Left arrow
    Open the next page from your browsing history in the current tab Alt + Right arrow
    Close the current tab Ctrl + w or Ctrl + F4
    Close the current window Ctrl + Shift + w or Alt + F4
    Minimize the current window Alt + Space then n
    Maximize the current window Alt + Space then x
    Quit Google Chrome Alt + f then x
    Move tabs right or left Ctrl + Shift + PgUp or Ctrl + Shift + PgDn

R

  • Search for R packages

  • Changelog for RStudio versions

  • Detach a package

    detach("package:vegan", unload = TRUE)

  • Import functions from package

    # R 4.5.0
use("dplyr", c("filter", "select"))
    • I think this functionality is also available in R4.0+ with library(pkg, include.only = c("fun1", "fun2")

  • Reading in zipped files (source)

    # unz() lets you read one file from a .zip
one_zipped_csv_among_others <- readr::read_csv(
  unz("lotsa_zipped_csvs.zip", "one_csv.csv")
)

# .zip with a single CSV
one_zipped_csv <- readr::read_csv("big_zipped_file.zip")

  • R-devel (>= 4.4.0) gained a command-line option to adjust the connections limit (previous limit was 128 parallel workers)

    $ R
> parallelly::availableConnections()
[1] 128

$ R --max-connections=512
> parallelly::availableConnections()
[1] 512

  • Get installed package names, versions, environment, source (article)

    > package_names <- installed.packages()[, 1]
> package_data <- lapply(
+     package_names,
+     utils::packageDescription,
+     fields = c("Package", "Version", "Built", "Repository")
+ ) |>
+     lapply(as, Class = "list") |>
+     lapply(setNames, c("Package", "Version", "Environment", "Source")) |>
+     dplyr::bind_rows()
    • Environment has OS info, R version, and date installed.

  • Java Dependent Packages

    • {rJavaEnv} - Aims to assist users of all {rJava}-dependent packages by providing functions to quickly install the required Java version and set environment variables.

      • It downloads non-installer archives of Java, extracts them to a cache folder, and links them in the current project or working directory. This way, rJavaEnv does not contaminate the user’s machine with unnecessary installations and configurations.

      • When downloading a Java version, it tells you where it’s stored.

        Java 21 (amazon-corretto-21-x64-windows-jdk.zip) for windows x64 installed at
    C:\Users\ercbk\AppData\Local/R/cache/R/rJavaEnv/installed/windows/x64/21 and symlinked to
    C:/Users/ercbk/Documents/rjavaenv/windows/x64/21

      • Code for the path to the installation is then added to your .Rprofile.

      • “if you are on a Linux system that did not have any ‘Java’ version previoulsy installed and you are not using pre-build ‘R’ package binaries (for example from ‘Posit Package Manager’) and instead install all ‘R’ packages from source, after this step you may have to quit R, follow these steps: https://solutions.posit.co/envs-pkgs/using-rjava/#reconfigure-r to set configure ‘Java’ for ‘R’, and only then install ‘rJava’, as otherwise ‘rJava’ cannot be built from source.”

Hackathon Criteria

Update R

  • Misc

    • https://www.r-project.org/

    • RStudio NEW.md

    • {rig} - r version management system

    • update.packages(checkBuilt = TRUE, ask = FALSE) is supposed to search for packages in other R versions and update them in the new R version, but I haven’t tried it, yet.

    • Install the newest stable version to check it out

      rig add next
R-next
      • R-<ver> runs a R version without it being the default

    • Errors when compiling from source may require installing libraries and they’ll supply code to install via “pacman”

      • Open Start >> scroll down to RTools40 >> RTools Bash
      • Paste pacman code and hit enter to install

    • Problem packages in the past

      • {brms} dependency, {igraph}, didn’t have a binary on CRAN and wouldn’t compile from source even with correct libraries installed.
        • Sol’n: install.packages("igraph", repos = 'https://igraph.r-universe.dev')
          • installs dev version from r-universe
      • Some {easystats} packages had gave {pak} some problems. No difficulties using install.packages with default repo or if they had a r-universe repo though.
  • Steps

    1. Copy user installed packages in current R version

      • In R:

        squirrel <- names(installed.packages(priority = "NA")[, 1]) # user installed packages
readr::write_rds(squirrel, "packages.rds")
        • Then, close RStudio

    2. Check/Update rig version

      • In powershell: rig --version
      • Check current rig release: link
      • Download and install if your version isn’t current
        • Windows:
          • winget install posit.rig
          • Executable (See Assets section)

    3. Install new version of R

      • Close R if not already closed
      • rig add release installs the latest version of R.
      • rig default <new_r_version> sets that version as the default
        • e.g. rig default 4.4.2

    4. RTools: Check to see if you have the latest because you’ll need it to compile some of newest versions of packages.

      • Your rtools folder has the version in it’s folder name.
      • rtools website has the latest version and an .exe to download
      • rig add rtools43 will install the RTools for R 4.3 or you can specify any version
      • rig add rtools will install all RTools versions for all R versions that are installed.

    5. Add R and RTools to path

      • Right-click Windows >> System >> (right panel) Advanced System Settings >> Environment Variables >> Under User Variables, highlight Path, click Edit >> Click Add
        • R: Add path to directory with all the RScript, R exe, etc. e.g. “C:\Program Files\R\R-4.2.3\bin\x64”
        • RTools: e.g. “C:\rtools43\usr\bin”

    6. Open R and confirm new version

      • If RStudio
        • The setting of the new version to the “default” version of R in rig should result in RStudio loading the new version if you have automatic detection enabled.
        • If not, Tools >> Global Options >> General
          • Under “R version”, click “change” button; choose new R version
          • Quit session and restart RStudio

    7. Install “high maintenance” packages

      • I’ve had issues with {pak} installing packages that need to be compiled. Maybe be worth trying {pak} first to see if they’ve fixed it.
      • {cmdstanr} doesn’t live on CRAN, so you have to use: install.packages("cmdstanr", repos = c("https://mc-stan.org/r-packages/", getOption("repos")))
        • Check for latest cmdstan version
          • After loading the package, library(cmdstanr) , it should run a check on your cmdstan version and tell you if there’s a newer version.
          • To update, first check toolchain: check_cmdstan_toolchain()
            • Might tell you to update RTools or that you need some C++ library added
          • Fix C++ toolchain with check_cmdstan_toolchain(fix = TRUE)
          • Update cmdstan: install_cmdstan()
          • May need to install {rstudioapi} and run rstudioapi::restartSession() (programmatically) or just ctrl + shift + f10 so that this package can be used as a dependency for other packages that need to be installed.
      • {rstanarm}: install.packages("rstanarm")

    8. Install other packages

      moose <- readRDS("packages.rds")
moose <- moose[!moose %in% c("cmdstanr", "rstanarm", "ebtools", "translations", "<RStudio add-ins>", "<other pkgs not on CRAN>")]

# Next time, add a try/catch? or maybe purrr::safely, so that it continues through errors. Also, need to log pkgs that do error.
for (i in seq_len(length(moose))) {
  print(moose[i])
  pak::pkg_install(moose[i], ask = FALSE)
}

fs::file_delete("packages.rds")

      • {ebtools} is my personal helper package.

      • {translations} is a system package that shouldn’t have been included when I saved the packages from previous version, but was when I recently updated. Might not be necessary to include it in the excuded packages in the future.

      • If there’s an error in downloading a package, continue on to the next package

        # ex: installing package "mixedup" errored
elt_num <- which(moose == "mixedup")
squirrel <- moose[elt_num:length(moose)]

# carry on
for (i in seq_len(length(squirrel))) {
  print(squirrel[i])
  pak::pkg_install(squirrel[i], ask = FALSE)
}

fs::file_delete("packages.rds")

    9. Check for updates of RStudio (link)

      • Current version under Help >> About Rstudio
      • Possible to check for updates under Help >> Check for Updates, but that’s failed me before.