• Declarative programming centered around functions and logic (OOP capable)
• Emphasis on statistics and visualization of data
• Used by researchers and industry
• Open sourced, active useRs community (25610 packages)

• R very powerful for creating reproducible workflows
  • Scripted analysis
  • Markdown (“Rmarkdown” + the new kid “Quarto”)
  • https://github.com/FAIReLABS/trirods_ugm2022

Never again wonder what method did I use to center variable “foo” in my regression model … ?

• But what about the data itself?
  • SQLite, MySQL, PostgreSQL, MonetDB with DBI package and ODBC drivers
  • iRODS?

• Store unprocessed data
• Share data and scripts with collaborators
• Reviewer trace back the origin of data and imposed manipulations
• Publish everything = open science

• Authentication
  • connect with iRODS server
  • authenticate

# configuration file
create_irods("<host>", "<zone>")
# authenticate
iauth()

• Navigation
  • icommand like

# current working collection
ipwd()
# change working collection
icd("<path>")
# list
ils()

• Objects/files
  • icommand like

foo <- 1:10
# configuration file
iput(foo)
# authenticate
iget("foo") 
# or 
iget(foo) # ?

• Data discovery
  • imeta vs iquest

# add some metadata
imeta(
  "foo", 
  "data_object", 
  operations = 
    list(operation = "add", attribute = "foo", value = "bar", units = "baz")
)
# discover 
iquery("SELECT COLL_NAME, DATA_NAME WHERE COLL_NAME LIKE '/tempZone/home/%'")

• Curl in R
  • R interface to libcurl curl (Ooms 2022a)
  • Wrapper httr2 (Wickham 2022) for curl and jsonlite (Ooms 2022b)
• Development + Testing
  • irods demo server docker-compose up -d nginx-reverse-proxy
  • Testing with mocking httptest2 (Richardson 2022)
  • Automatic updates of snapshots with GitHub actions
  • R CMD check without internet (simulate CRAN checks)

• Submission to CRAN
• Official release at UGM (summer 2023)
• More R packages: datamanager + panacaea

Help us! https://fairelabs.github.io/webpage/