carpentries-incubator / bioc-rnaseq

Analysis and Interpretation of Bulk RNA-Seq Data using Bioconductor
https://carpentries-incubator.github.io/bioc-rnaseq
Other
21 stars 30 forks source link

Transition to workbench #15

Closed csoneson closed 1 year ago

csoneson commented 1 year ago

Hi @zkamvar,

similarly to https://github.com/carpentries-incubator/bioc-intro/issues/76 and https://github.com/carpentries-incubator/bioc-project/issues/48, I would like to ask for your help with transitioning also this repository to the new workbench format.

Thanks a lot in advance, and let me know if I can do anything to help!

All the best, Charlotte

Instructions Thanks for contributing! :heart: If this contribution is for instructor training, please email the link to this contribution to checkout@carpentries.org so we can record your progress. You've completed your contribution step for instructor checkout by submitting this contribution! If this issue is about a specific episode within a lesson, please provide its link or filename. Keep in mind that **lesson maintainers are volunteers** and it may take them some time to respond to your contribution. Although not all contributions can be incorporated into the lesson materials, we appreciate your time and effort to improve the curriculum. If you have any questions about the lesson maintenance process or would like to volunteer your time as a contribution reviewer, please contact The Carpentries Team at team@carpentries.org. You may delete these instructions from your comment. \- The Carpentries
zkamvar commented 1 year ago

I'll add it to the list!

csoneson commented 1 year ago

Hi @zkamvar - thanks for your recent updates concerning the Workbench transition. I followed the discussion over at the bioc-intro lesson, and I just wanted to let you know that similar to what Laurent did there, I also pulled down and tested https://github.com/fishtree-attempt/bioc-rnaseq/ locally, and everything seems to be running smoothly - the output also looks nice 🙂. Please let me know if there's anything else I can do to help in the transition. Thanks!

zkamvar commented 1 year ago

Thank you! I should be getting to the transition today. I was cleaning up a few things yesterday to make the transition a bit smoother (e.g. transforming raw github links to relative paths that will resolve correctly: https://github.com/carpentries/lesson-transition/blob/0478e2f8393bad6ceba70e0de6328d2c6d4f2ba2/carpentries-incubator/bioc-rnaseq.R#L21-L26).

I am going to send a message here notifying everyone on this repository about what to expect and then set all permissions to read except for @csoneson, so that they can give permissions.

Here are some resources in the meantime:

  1. What you can expect from the transition 📹: https://carpentries.github.io/workbench/beta-phase.html#beta
  2. How to update your local clone 💻: https://carpentries.github.io/workbench/beta-phase.html#updating-clone
  3. How to update (delete) your fork (if you have one) 📹: https://carpentries.github.io/workbench/faq.html#update-fork-from-styles
zkamvar commented 1 year ago

This lesson will be converted to use The Carpentries Workbench To prevent accidental reversion of the changes, we are temporarily revoking write access for all collaborators on this lesson:

If you no longer wish to have write access to this repository, you do not need to do anything further.

  1. What you can expect from the transition 📹: https://carpentries.github.io/workbench/beta-phase.html#beta
  2. How to update your local clone 💻: https://carpentries.github.io/workbench/beta-phase.html#updating-clone
  3. How to update (delete) your fork (if you have one) 📹: https://carpentries.github.io/workbench/faq.html#update-fork-from-styles

If you wish to regain write access, please re-clone the repository on your machine and then comment here with I am ready for write access :rocket: and the admin maintainer of this repository will restore your permissions.

If you have any questions, please reply here and tag @zkamvar

zkamvar commented 1 year ago

The deed is done. @csoneson, the infrastructure takes a few minutes to bootstrap and cache the packages for the lesson build). Once the build is done, I will switch github pages to deploy from the gh-pages branch and you will have your workbench lesson.

Thank you all for your enthusiasm and your patience!

zkamvar commented 1 year ago

Note: this is taking a bit longer than usual because there was a 404 error in the package cache setup, so I had to restart the job. I suspect it was ephemeral, but I will keep you posted.

zkamvar commented 1 year ago

It turns out that the 404 error comes from step 6, line 3758 of the run:

  Warning message:
  curl: (22) The requested URL returned error: 404 
  Error: 
  ! in callr subprocess.
  Caused by error: 
  ! failed to find source for 'data.table 1.14.8' in package repositories

Considering that data.table was just released today (less than 12 hours ago: Date/Publication: 2023-02-17 12:20:12 UTC), It is fair to say that this is indeed an ephemeral issue (it could be that it had not yet been propagated to all the mirrors yet) as the build for the beta test worked well: https://fishtree-attempt.github.io/bioc-rnaseq/

I will be travelling over the weekend, but I would suggest to let this wait a day or so and then re-start the build. By then the mirrors should have updated and you will not see the problem.

zkamvar commented 1 year ago

An update: I tried again to run the lesson on CI.

It turns out the problem is ephemeral, but it's not exactly what I thought. I believe the problem is that, data.table has not yet been updated on the Posit Package Manager (see https://github.com/carpentries/actions/issues/69). This normally would not be a problem, but because of the way we have to handle bioconductor packages (by catching the "package not found" errors and re-running), {renv} tries to be helpful and install packages from the PPM, but the PPM is always a week out of date.

At the moment, there are two solutions to this:

  1. downgrade data.table to 1.14.6 using sandpaper::pin_version("data.table@1.14.6")
  2. wait until data.table is updated on PPM

In the meantime, you should still be able to run your lesson locally and (when I am not so jetlagged,) I will try to find a better process for the BioConductor packages.

csoneson commented 1 year ago

Thanks @zkamvar, this is great! I just checked in at PPM and it seems that data.table 1.14.8 is available since 10 minutes or so 🙂 I'll try to kick off the CI again and see if it resolves it.

csoneson commented 1 year ago

Seems to have worked 🎉

zkamvar commented 1 year ago

wunderbar!

The Workbench version is now live: https://carpentries-incubator.github.io/bioc-rnaseq/

In addition, here is map of commits that were changed during the transition