Release 1.4: Common tasks, Dev acknowledgement, versioning, helper packages/websites, etc

[manciniedoardo]

Master PR

Common Release Tasks

All of these small Release Tasks should be done in one issue and not split out across multiple issues and PRs. I have created a common task list for us to check and hope to develop the task (if needed) below. If we can standardized this stuff, then we could build a release_function that is a wrapper around the https://usethis.r-lib.org/reference/use_release_issue.html

Prepare for release:

• New devs to description file and old devs to acknowledgements
  • Moved to acknowledgments: Junze Zhang, G Gayatri, Daniel Sjoberg, Dinakar Kulkarni, Solveig Holmgaard
  • New authors: @MattBearham - let me know if you want an ORCID added too, Matt
• Review FAQs
• Review Cheatsheet Documentation: Updates to cheatsheet for 1.3 #2758 - as per feedback from last process, doing this in a separate issue as we wish tomerge it right before 1.3 release.
• Polish news for admiral and admiraldev
• Update README schedule for future releases
• Review pharmaverse admiral youtube videos (https://www.youtube.com/channel/UCxQFEv8HNqM01DXzdQLCy6Q)
• Check Size of package ( Code here )
• Check version numbers are correct in DESCRIPTION (i.e. admiraldev, pharmaversesdtm, others?)
• Check Remotes is removed form DESCRIPTION
• Check current CRAN check results
• Scan and remove non-ascii characters - Remove non-ascii characers #2456
• Check if any deprecation processes should be advanced, as described in Gradual deprecation
• urlchecker::url_check() checked as of 19th June
  
• devtools::check(remote = TRUE, manual = TRUE) output as of 19th June:
  
• devtools::check_win_devel() - doesn't work with Remtoes
• Run roche templates
• Use Insight Engineering reverse dependency checks. Requires manually running the Action. Might want to check the Rust one published on pharmaverse slack https://nanx.me/blog/post/revdep-speedrun/
• Review pharmaverse examples for impact (use News to assess)
• Draft blog post

Submit to CRAN:

• Be sure to use [skip vbump] in PR Title to prevent upversioning
• usethis::use_version('major')
• devtools::submit_cran()
• Approve email
• Monitor submission form here: https://r-hub.github.io/cransays/articles/dashboard.html
• Start using CRAN Comments, usethis::??

CRAN Accepted:

• usethis::use_github_release()
• usethis::use_dev_version()
• Finish blog post
• Add Releases and tar.gz file to the Release
• Review admiraldiscovery (presentations and find a function)
• Push forward deprecation

Loose Guidance for particular tasks

New presentations

• Posit Conference
• Something from Edoardo?
• Something from Fanny?

pharmaverse/admiraldiscovery#64

News devs to description File and old devs to acknowledgements - NB: check admiraldev as well

#2748
pharmaverse/admiraldev#492

• Gayatri
• Junze
• Arianna

Devs to moved to acknowledgments (not actively working on admiral anymore) at the bottom of README:

• Dinakar
• Junze
• Solveig

New acknowledgements

• Check closed PRs for mini-PRs and issues raised by users

Review FAQs

• Check for outdated info
• Chances to consolidate information or remove
• Possible to link to other information to help minimize upkeep

Polish News

• Each news item should have a linked issue
• Check for consistency and grammar

[bms63]

Cool Issue Template - https://github.com/insightsengineering/teal/issues/new?assignees=KlaudiaBB%2Ccicdguy&labels=release&projects=&template=release.yml&title=%5BRelease%5D%3A+%3Cversion%3E

[bms63]

Making note here on cheatsheet - needs a little more thought on when to update #2600

[bms63]

I don't know what this means? @manciniedoardo

[ddsjoberg]

@bms63 I wouldn't worry about the qpdf and latex warnings. If you want, you can install it on your machine to avoid the message. Resolving the latex warning will depend on how the latex distribution on your machine, so I would just ignore it. These both arise from building the PDF manual, which CRAN does on their end anyway.

Has the size of the package grown since the last release? Unless the rules have changed, you may be auto-rejected because the build is >5MB. But I know there are packages that exceed this size on CRAN, but I have never gone through that process. 🤷🏼

[bms63]

Thanks @ddsjoberg!!! I was more poking at @manciniedoardo about what we are doing about these warnings and notes - we should be all clean.

• I have a linux installation with qpdf installed so this one doesn't come up for me.
• when i do a manual check on the size for latest admiral dev I get this
  
• Not sure on the tidy Note ?
• @manciniedoardo maybe add to .Rbuildignore the admiral-manual.tex so it doesn't show up as a Note?

[bms63]

Sniff out unicode characters in .Rd files

search_specific_unicode_in_rd_files <- function(folder_path, unicode_char) {

List all .Rd files in the folder

rd_files <- list.files(path = folder_path, pattern = "\.Rd$", full.names = TRUE)

for (file in rd_files) {
# Read each .Rd file
rd_content <- readLines(file, encoding = "UTF-8")

# Identify lines containing the specific Unicode character
has_unicode <- grepl(unicode_char, rd_content)
unicode_lines <- rd_content[has_unicode]

if (length(unicode_lines) > 0) {
  cat("File:", file, "\n")
  cat("Lines with the Unicode character ℹ:\n")
  print(unicode_lines)
  cat("\n")
}

}
}

Define the Unicode character (information source symbol)

unicode_char <- "ℹ"

Run the function with your folder path

search_specific_unicode_in_rd_files("/cloud/project/man", unicode_char)

[bundfussr]

On linux/unix systems you could also use

grep --color=auto $(printf \\U2139) *.Rd

[bms63]

grep forever!! much simpler :)

[manciniedoardo]

@1mgroot @gg106046 @ddsjoberg @cicdguy @SolveigHolmgaard

This is just a courtesy notification that as per our internal guidelines, for the 1.4 release we will be moving your names from the authors section of the DESCRIPTION file to the acknowledgements section of the README, because you have not been involved with the package during the 1.3 release cycle. Of course, if in future your involvement with admiral picks up again, we'll be happy to re-add you as authors.

[manciniedoardo]

Jan 12, 2026 Deep Dive WG - Gemini notes

Summary

@mancinie, @millerg23, @adcascone, @bundfussr, @bms63, @jeffreyad, and other participants discussed the issue of the {admiral} package exceeding the 5 megabyte CRAN limit. The team explored causes such as new vignettes and datasets, with @adcascone suggesting BZIP2 compression as an immediate fix, which @mancinie confirmed would reduce the size sufficiently for the current release. The team also addressed a longer-term strategy, with @bundfussr suggesting reducing the size of admiral_adlb and @bms63 proposing a regular check for package size. Finally, @bms63, @bundfussr, and @millerg23 discussed the need for more flexible Time-to-Event (TTE) functionality, agreeing to gather scenarios to inform updates to the existing derive_param_tte() function or the creation of a new one.

Details

{admiral} Package Size Concern

@mancinie and @millerg23 opened the meeting by addressing the size of the {admiral} package, which had exceeded 5 megabytes, potentially causing issues with CRAN submission. @adcascone inquired about the previous size to determine if the increase was a sudden jump.

Causes for Package Size Increase

@bundfussr and @millerg23 discussed potential reasons for the increased package size, which was unexpected since Excel files were replaced with JSON files. @mancinie calculated that the package gained about 200 kilobytes. @bundfussr suggested that the new vignettes for ADAB might be a factor, with @mancinie identifying a 14 kilobyte file as "by far the biggest," suggesting the package may have just crossed the size barrier.

Strategies for Size Reduction

The participants discussed options for reducing the package size, including seeking an exemption from CRAN or cutting package content. @adcascone proposed moving templates, specifically the ADAB template, to a separate package, though @mancinie noted that ADAB is the only particularly large template. @bundfussr calculated that templates were less than 100 kilobytes, suggesting minimal gain from moving them. @mancinie suggested that the ADLB dataset might be removed or reduced. @jeffreyad and @mancinie discussed deriving the ex_single dataset on the fly to avoid saving a large file.

Immediate Solution with Compression

@adcascone suggested implementing BZIP2 compression for RDA files created from ATOX GR sources. @mancinie ran a test and determined that applying the BZIP2 compression reduced the size by 0.25 megabytes (from 0.49 to 0.24). @bms63 confirmed that this reduction would be enough to bring the package under the 5.1 megabyte limit for the current release.

Long-Term Package Size Strategy

The team agreed that while the compression fix would work for this release, a longer-term strategy is needed. @bundfussr suggested reducing the size of admiral_adlb, and @mancinie noted they could filter variables from datasets like ADLB. @bms63 suggested creating an issue to check the package size regularly and find optimizations outside of the rushed release cycle.

Discussion on Time-to-Event Functionality (TTE)

@bms63 initiated a discussion on potential changes for the 1.5 release, specifically regarding the TTE function, aiming to make it more flexible for estimand strategies. @bundfussr and @bms63 agreed that the estimand (ESTIM) topic should be a separate discussion and vignette from the TTE function updates. @bundfussr emphasized the need to investigate what "more flexible" means, including different scenarios like negative and positive events or composite endpoints. @bms63 proposed gathering more scenarios from Slack to inform decisions on whether to update the existing derive_param_tte() function or create a new one.

Estimand Guidance and Vignettes

@bundfussr mentioned that TTE derivation often involves intermediate datasets like ARS or AD event. @millerg23 stressed identifying scenarios before creating new functions. @bms63 inquired about open-sourcing an internal estimand guidance document, and @mancinie agreed to check with Ross.

Release Process and Next Steps

@bms63 confirmed that the {admiraldev} updates for deprecation and experimental tag guidelines are almost ready. The goal is to submit {admiraldev} first, followed by {admiral}, by the end of the week. This is time-sensitive as all extension packages, including {admiralmetabolic}, {admiralvaccine}, and {admiralophtha}, need to release by February 10th to address CRAN issues regarding relative links.

Suggested next steps

• @mancinie will commit the updates for adding the compress argument and reducing the ADLB dataset size to the release PR.
• The group will create an issue to check the file size of the package and find optimizations after the immediate release fixes.
• @bms63 will split the discussions around the derive_param_tte() function flexibility and the estimand strategies into separate topics.
• @bms63 will canvas the Slack channel to gather more scenarios for the TTE function flexibility.
• @bundfussr will elaborate on the slides and collected information around TTE to use as a basis for a potential new vignette or functions.
• @mancinie will check with Ross regarding open-sourcing the internal document on estimand strategies.
• @bms63 will clean up the deprecation changes and fixes in {admiraldev} to enable its release.
• @bms63 will start the submission process on Wednesday.

[bms63]

Quick note on 404 to check

#2948