This bot scrapes updates about tidyverse packages from GitHub and posts them to twitter π
-
π’ You can find the bot at @tidyverse_news
-
π¨βπ» This bot is maintained by @_wurli
-
NEWS.md
files are scraped from GitHub for all the packages listed bytidyverse::tidyverse_packages()
-
Annotations are applied to these files using pandoc with a custom Lua filter. Annotations indicate where bullets begin and end so that tweets can be grouped together sensibly.
-
This data is transformed using R to give individual tweets or threads where bullets would otherwise overflow. The twitter-text javascript library is used here to test whether tweets are valid or not.
-
Updates which have already been posted are excluded by checking against a CSV database. Updates don't have to exactly match previously posted ones - if they're very similar to previously posted updates, they won't get posted again.
-
Tweets are posted using the {rtweet} package
This is all automated using GitHub Actions, which simply sources run.R
once every hour.
If you use R a lot it's interesting to see what features are in the works for popular packages. I also wanted to learn more about twitter bots and GitHub actions, so this seemed like a good learning exercise.
See R/news_urls.R
for the full list of packages this
bot follows. If you think any are missing, please open an issue on
GitHub or send me a message on twitter π