Hosting experiment repo on Github

Earth System Earth System Model
,
21

You shouldn’t need to create the branch first. I think the issue is that you’re trying to push through the git URL using https authentication. Shorter solution which might work:

git remote remove access-esm1-5_IP_m60
git remote add access-esm1-5_IP_m60 https://github.com/ACCESS-Community-Hub/access-esm1.5-sst-perturbation-indo-pacific
git push access-esm1-5_IP_m60 release-preindustrial+concentrations:georgyfalster-60pc_SST_variability_reduction

If you get permission errors again, then you’ll have to setup SSH authentication for Github. That means adding an SSH key and then changing authentication method from https to SSH with:

gh config set -h github.com git_protocol ssh

You can check that it’s worked with gh config get git_protocol. After that, follow those same steps you tried before, and it should get around those permission problems. If that doesn’t fix it, might be worth jumping on a video call and working through it, because it means I’m missing something as well.

22

Success!! All looking good GitHub - ACCESS-Community-Hub/access-esm1.5-sst-perturbation-indo-pacific: Magnitude of tropical Indian and Pacific ocean SST variability altered relative to ACCESS-ESM1.5 preindustrial control

Another thing is that the experiments rely on two input files: a SST restore file (that tells the models what SST anomalies to impose) and a mask file (that tells the model where to impose the anomalies). @spencerwong helped me to make these but it was a bit of a faff - it would be nice to have the process out there available for others to follow easily. If I can get those to my github, is there some ‘official’ way I can link that to these experiment repos?

one final thing: in github, is there some way I can ‘co-own’ the experiment repos? so that they show up under my github repos too?

thanks so much again (:

1 Like
23

Awesome, did you have to change to SSH? Would be nice to know for future issues people have.

I don’t think we have an official way to link preparation steps- it’s something we’re still discussing. Do you have the scripts you used to generate these files in a repository? If so, you could add a description of your process and link to the repository in the README of the experiment repository?

Unfortunately I don’t think co-ownership is possible (although someone like @Aidan might know otherwise?). As a bit of an ad-hoc solution you could fork the repo, and have a copy owned by yourself. Not great though, because any changes you make won’t be included by default in the copy (although it is pretty easy to get those changes)

24

No, I didn’t have to change to SSH - your solution worked.

For anyone else coming here with the same question who doesn’t want to wade through the thread, the steps that ultimately worked for me were:

  • Followed the Authorise with github steps here
  • Manually created new repo on the ACCESS Community Hub ACCESS Community Hub · GitHub
  • Went into my experiment control directory and ran git remote remove origin
  • Ran the below lines to push my experiments to new branches of the repo I created manually on the ACCESS Community Hub (in the second line). You probably won’t need the first line unless you’d run into prior difficulties like I did.
    • access-esm1-5-IP_m60 is the local name for the remote repository
    • https://github.com/ACCESS-Community-Hub/access-esm1.5-sst-perturbation-indo-pacific is the link to the manually-created repo on the ACCESS community Hub
    • release-preindustrial+concentrations is the current branch, identified with git branch as per these instructions
    • georgyfalster-60pc_SST_variability_reduction is the new branch name
git remote remove access-esm1-5_IP_m60
git remote add access-esm1-5_IP_m60 https://github.com/ACCESS-Community-Hub/access-esm1.5-sst-perturbation-indo-pacific
git push access-esm1-5_IP_m60 release-preindustrial+concentrations:georgyfalster-60pc_SST_variability_reduction

And yes @lachlanswhyborn that sounds fine - I’ll put the SST preparation scripts on github and link that in the readme. Also no worries on the ‘no co-ownership’ - I was just thinking of a case where someone might have questions about the experiments but I’ve put my name in the branch name so they should be able to track me down.

2 Likes
25

Thanks for working through the detail to do this.

I think it’s pretty obvious you’re the author, but you can put explicit contact details in the README on the main branch. You can also add a CODEOWNERS file to the main branch, which is a bit of overkill in this case, but is what is is designed for

26

great, thanks - and yes I don’t think there will be any problems with anyone finding me if they have questions about the experiments! I’ll tidy up the readme though.

1 Like