jtr13/cc21fall2

Add API tokens to secrets

jtr13 opened this issue · 5 comments

jtr13 commented

Chapters that require API tokens to build will not work unless we add the tokens to GitHub secrets and access them in GitHub actions. For now code chunks are set to eval=FALSE in https://jtr13.github.io/cc21fall2/scrape-twitter-data-using-r.html to circumvent this problem.

@rjr2168 developed a workflow for https://jtr13.github.io/cc21fall2/video-introduction-to-maps-with-ggmap.html using this method which we can use as a model if other authors are interested in going this route.

If you set eval=FALSE for other chapters for the same reason, please add a reference to those chapters in this thread.


Morning prof. Robbins. It seems in this chapter line 121-122 are still causing problems as yesterday...should we add eval = FALSE to all trunks in this chapter? 😿

jtr13 commented

Thanks for the heads-up. I already set eval=FALSE to the top of the document with knitr::opts_chunk$set(eval=FALSE) and am not completely sure why that's not having an effect. I'm trying now some other approaches to get the build to work.

Hello! I will try to do a write up of the exact modifications needed to make API keys work later today, if you decide to go that route. If you'd like an example of code modified from eval=FALSE to using API tokens, you can look in the history of my file and the .github/workflows/bookdown.yaml file to see what it entails. The only other modification made was the creation of a github secret.

jtr13 commented

Working again: https://github.com/jtr13/cc21fall2/actions/runs/1435105154 I had to add cache=FALSE in addition to eval=FALSE to knitr::opts_chunk$set(). Somehow the build was referencing old code for reasons that are not clear to me. Add cache=FALSE forces it to update all the code chunks.

As I mentioned earlier, I have made a brief discussion post detailing the process.