What is buffy?

✨ Editorial bots generator coded in Ruby ✨

• GitHub repository: https://github.com/openjournals/buffy

• Docs: https://buffy.readthedocs.io/en/latest/

rOpenSci (Statistical) Software Peer-Review

✨ Transparent, constructive, non adversarial and open review process of R packages✨

• GitHub repository: https://github.com/ropensci/software-review

• Docs: https://ropensci.org/software-review/ & https://ropensci.org/stat-software-review/

GitHub Issue Thread

Why is rOpenSci software review special?

❄️ Specific R package checks;

❄️ Repositories transferred after approval;

❄️ Levels of minting in statistical software review (gold/silver/bronze);

❄️ Airtable database;

❄️ …

Every system is probably special, buffy is extensible! 🎉

What’s an adaptation of buffy?

In the “ropensci” branch of our buffy fork at https://github.com/ropensci-org/buffy/tree/ropensci :

• Configuration;

• Specific responders!

Responders documented at ReadTheDocs.

The big picture

Package submission

Welcome

Package checks at submission

Package checks later

• @ropensci-review-bot check package by editors in the issue;

• pkgcheck::pkgcheck() locally.

Assigning a reviewer, before 🤖

😅 Something like…

• Calculate due date, three weeks from now.

• Assign reviewers by tagging in the issue with the following format:

  Reviewer: @githubname1 
  Reviewer: @githubname2
  Due date: YYYY-MM-DD
  

• Remind reviewers to fill the volunteering form if relevant.
• Store reviewer info in software review database (Airtable).
• Switch numbered label to 3/reviewers-assigned once 2 reviewers are assigned.

Assigning a reviewer, with 🤖

@ropensci-review-bot assign @username as reviewer

Assigning a reviewer, with 🤖

😌 @ropensci-review-bot assign @username as reviewer

• Easier to remember;

• We can tweak what happens in the background!

How we adopted
and adapted buffy

Who are “we”:

Apart from Karthik Ram and Arfon Smith,

• Juanjo Bazán, buffy developer 💎

• 👋

• Mark Padgham, rOpenSci Software Research Scientist.

• All users = patient beta-testers of buffy! Package authors, reviewers, editors, rOpenSci community manager Stefanie Butland…

How we assessed our needs

Reading…

• Upstream buffy’s docs;

• rOpenSci dev guide;

• Plans for rOpenSci statistical peer-review.

How we tested features

• In a test repository:

• In real life. 😉

Commands

@ropensci-review-bot <do-something>

• Meta (help, code of conduct);

• Editorial management;

• Review(er) management;

• Package checking;

• Approval.

Still a humane process 🤗

• The reviews are written by humans!

• Editors write comments personally thanking e.g. reviewers.

• Out-of-scope decisions are also justified with a non automatic comment.

How did we introduce commands?

• Direct communication with editors;

• Release 0.7.0 of our dev guide!

https://ropensci-review-bot.netlify.app/

@ropensci-review-bot thank developers & testers & today's audience

• A very useful tool, less mental overhead and tedium!

• More transparency!