rstudio-conf-2020 / my-org-first-pkg Goto Github PK

Hi!!

I've created a do-setup branch (https://github.com/pachamaltese/my-org-first-pkg/tree/do-setup) that does some "magic" in three steps:

From terminal:

1. fab build_image -H your_build_droplet_ip_address
2. ssh [email protected] && adduser the_user

From browser:
3. Go to 123.456.789:8787 and login as the_user

1+2+3 creates an image with all the packages used in the workshop and the intendo database, the only difference with respect to the workshop is this:

# old connection
con <- DBI::dbConnect(drv = RMariaDB::MariaDB(), dbname = "intendo", username = "student", password = "old_password", host = "foo.bar/very-long-url", port = 3306)

# new connection
con <- DBI::dbConnect(drv = RMariaDB::MariaDB(), dbname = "intendo", username = "student", password = "new_password", host = "127.0.0.1", port = 3306)

My idea is to send this to DigitalOcean Marketplace later so that users need 3 clicks and in ~1 minute they shall have the full setup to follow the workshop. See as a reference this (scalable) image I created after I attended Hadley's workshop: https://marketplace.digitalocean.com/apps/rstudio-pkgdev at LatinR 2019.

From the instructor's side, this allows each attendee to have his/her own RStudio Server instance, or the instructor can click 3 times on DigitalOcean panel and create a server with, lets say, 0.5GB x "# of attendees" in RAM that is billed only for the duration of the workshop.

First day: R package development

• R package basics, usethis
• testing
• internal data
• whole game: https://r-pkgs.org/whole-game.html

Second day: Topics in internal packages

• Set up live databases to practice connecting to (create MySQL DB with sample data through AWS)
• Report templates (possibly using ymlthis for standard yaml)
• opening and editing files (usethis helpers?)
• Themes (e.g. ggplot, palette functions, css, formatters for style guides (e.g. rounding))
• Where should I host my package? (Git services, S3, tar.gz, RStudio Package Manager?)
• Hosting documentation internally? R Markdown sites, distill, pkgdown
• wrapper functions for plotting and analysis
• automation, particularly in reporting (cron jobs, connect, emailing)
• Shiny app functions (local shiny apps, helper functions for dashboards etc)
• logging (https://github.com/daroczig/logger)
• using httr? set up something with plumbr to play with
• maintaining and feedback?
• slack/google drive etc helpers?

As closing material, perhaps point people to resources like drake, research compendiums, and holepunch for additional strategies

include "helping others contribute" section using usethis pr_*() functions?

package development

https://r-mageddon.netlify.com/post/writing-an-r-package-from-scratch/

internal R packages:

https://rviews.rstudio.com/2017/07/19/supporting-corporate-r-user-groups/
https://masalmon.eu/2017/07/22/chaimagic/
https://support.rstudio.com/hc/en-us/articles/115000239587-Sharing-Internal-R-Packages
https://github.com/gta-r-user-group/gtar/tree/master/2019-06-04/proprietary-r-packages
https://auth0.com/blog/rauth0-internal-r-package-open-source/

managing credentials:

https://db.rstudio.com/best-practices/managing-credentials/
https://db.rstudio.com/best-practices/deployment/
https://httr.r-lib.org/articles/secrets.html
https://github.com/daroczig/dbr
https://github.com/daroczig/botor

httr:

https://www.lucymcgowan.com/talk/asa_joint_statistical_meeting_2018/
https://httr.r-lib.org/articles/api-packages.html

https://emilyriederer.netlify.com/post/resource-roundup-r-in-industry-edition/

• Favicon
• TA details
• Day 2 details
• Links to materials and slides

An R package to download materials and install needed packages. I'll handle this, as I have a template.