Add anu-tub setup to configurations page

[jbisits]

Closes #4

---

FYI
PR preview is no longer working on this PR

[jbisits]

@chrisb13 this is a draft/skeleton of setting up the ANU-TUB configuration. Whenever you get a chance (no rush at all on this) let me know what you think.

We need to check with Andy Hogg that the relevant restart files can be placed somewhere that can be accessed (the directory you mention in #4 would be just fine).

[jbisits]

Hi @chrisb13,

I have added what I think is most of the info to the ANU-TUB setup page. The main things to be resolved, which are indicated on the webpage currently, are:

• relevant project access and who to get in touch with to get access
• access to input and restart files (I have these from Andy but will need his permission to store them in a directory people can access them)
• copying to /g/data. Andy has a script, which he also shared with me, qsub sync_output_to_gdata.sh which automates this process. This could also be provided (provided Andy is happy for it to be shared) but maybe it is better to let people move data to /g/ on their own? If so can get them to remove the lines in the config.yaml and can provide a warning that they will need to move data to g themselves.

[chrisb13]

@chrisb13 this is a draft/skeleton of setting up the ANU-TUB configuration. Whenever you get a chance (no rush at all on this) let me know what you think.

So awesome, thanks so much @jbisits !

We need to check with Andy Hogg that the relevant restart files can be placed somewhere that can be accessed (the directory you mention in #4 would be just fine).

Yeah I think we put it here: /g/data/ol01/access-iom3-configs it's what it is for.

If you think we'll be doing a lot of this, we can get you access or you can just send me a list of files to put there?

I have added what I think is most of the info to the ANU-TUB setup page. The main things to be resolved, which are indicated on the webpage currently, are:

• access to input and restart files (I have these from Andy but will need his permission to store them in a directory people can access them)

Okay, let's ask. @AndyHoggANU you happy for us to store them somewhere where folks can use them?

• copying to /g/data. Andy has a script, which he also shared with me, qsub sync_output_to_gdata.sh which automates this process. This could also be provided (provided Andy is happy for it to be shared) but maybe it is better to let people move data to /g/ on their own? If so can get them to remove the lines in the config.yaml and can provide a warning that they will need to move data to g themselves.

I think either approach can work. @AndyHoggANU are you okay for us to share the sync_output_to_gdata.sh script?

To my mind, I think we should consider making configurations like these a branch of this repo where it includes all the files. Rather like what is done here. I'll ask for broader views on this (ping @dougiesquire). Ideally a new user can just clone this repo, switch to the right branch and do payu run. If you see what I mean?

[jbisits]

Ideally a new user can just clone this repo, switch to the right branch and do payu run. If you see what I mean?

Yep that would be great. I think I can probably set that up and can make a note at the start of the configuration page that that is also an option.

[dougiesquire]

To my mind, I think we should consider making configurations like these a branch of this repo where it includes all the files. Rather like what is done here. I'll ask for broader views on this (ping @dougiesquire). Ideally a new user can just clone this repo, switch to the right branch and do payu run. If you see what I mean?

Sounds sensible to me. We may want to consider what the scope of this repo is? I don't think I really understand what "idealised-om3" means. Is it just anything that has some relation to OM3, e.g. uses a same component as in the case of ANU-tub?

I recall seeing some discussion that the ANU-tub config wasn't running with the version of MOM currently used in ACCESS-OM3? Has that been resolved? I'm happy to take a look if that's helpful (though, as usual, it might take me some time to get to)?

[chrisb13]

Ideally a new user can just clone this repo, switch to the right branch and do payu run. If you see what I mean?

Yep that would be great. I think I can probably set that up and can make a note at the start of the configuration page that that is also an option.

Cool, let me have a chat to Dougie first. (3pm today if you're interested.)

Sounds sensible to me.

Great.

We may want to consider what the scope of this repo is? I don't think I really understand what "idealised-om3" means. Is it just anything that has some relation to OM3, e.g. uses a same component as in the case of ANU-tub?

Yeah, that's correct but not limited to MOM. It's to encourage folks to build a hierarchy of models in Australia where one can have the same components as our regional/global realistic models. @edoddridge and I got excited about it at the access-nri comm' workshop this year.

It's also to help with our routine node testing.

I recall seeing some discussion that the ANU-tub config wasn't running with the version of MOM currently used in ACCESS-OM3? Has that been resolved? I'm happy to take a look if that's helpful (though, as usual, it might take me some time to get to)?

Yeah that was correct. I think @jbisits has solved all the issues (i.e. this pr) but I'll test when reviewing this PR.

[chrisb13]

Note to self, close when this is merged:

• Testing of ANU-tub using different MOM versions: Jbisits (Issue: #155) ACCESS-NRI/ACCESS-OM3#156
• Testing of ANU-tub using different MOM versions: angus-g (Issue: #155) ACCESS-NRI/ACCESS-OM3#157

[jbisits]

Cool, let me have a chat to Dougie first. (3pm today if you're interested.)

Sorry I could not make it along today!

[chrisb13]

No worries @jbisits

I recall seeing some discussion that the ANU-tub config wasn't running with the version of MOM currently used in ACCESS-OM3? Has that been resolved? I'm happy to take a look if that's helpful (though, as usual, it might take me some time to get to)?

Yeah that was correct. I think @jbisits has solved all the issues (i.e. this pr) but I'll test when reviewing this PR.

@dougiesquire and I just had another quick skim of the PR (see summary of chat here). It rather looks like you're using Angus' executable @jbisits? So perhaps you haven't been successful getting anu-tub working with a spack built mom6 access-nri fork executable not just Angus' one?

To be specific, here is the relevant lines:

model: mom6
shortpath: /scratch/x77                       # replace with NCI project you are part of
exe: /g/data/x77/ahg157/exes/MOM6/AG-07f0a144 # replace with path to MOM6 executable

So for it to "working", one would need:

exe: /path/to/spack/built/exe

[jbisits]

No worries @jbisits

I recall seeing some discussion that the ANU-tub config wasn't running with the version of MOM currently used in ACCESS-OM3? Has that been resolved? I'm happy to take a look if that's helpful (though, as usual, it might take me some time to get to)?

Yeah that was correct. I think @jbisits has solved all the issues (i.e. this pr) but I'll test when reviewing this PR.

@dougiesquire and I just had another quick skim of the PR (see summary of chat here). It rather looks like you're using Angus' executable @jbisits? So perhaps you haven't been successful getting anu-tub working with a spack built mom6 access-nri fork executable not just Angus' one?

To be specific, here is the relevant lines:

model: mom6
shortpath: /scratch/x77                       # replace with NCI project you are part of
exe: /g/data/x77/ahg157/exes/MOM6/AG-07f0a144 # replace with path to MOM6 executable

So for it to "working", one would need:

exe: /path/to/spack/built/exe

No I have been using executables that I have built (with both spack and mom6-ninja-nci). The above is a copy of the anu-tub config.yaml after cloning. This then needs to be edited so that exe points to point to the executable you want to use. In my case I have three executables I have been using:

• /scratch/e14/jb2381/mom6-ninja-nci-nm/ocean_only_symmetric/MOM6-nm
• /scratch/e14/jb2381/mom6-ninja-nci/ocean_only_symmetric/MOM6_202507
• /g/data/e14/jb2381/spack/0.22/environments/MOM_2025-07/.spack-env/view/bin/mom6

So the idea is that someone who is following this setup edits exe in config.yaml to point to the exectuable they want to use. I should make this clearer!

[jbisits]

@chrisb13 let me know what you think is best to do here. We could merge here with skeleton details in the section Using a pre-built ANU-TUB configuration and update when done? Otherwise happy to wait until the pre-built configuration is setup. Let me know if I can help with that!

[chrisb13]

@jbisits thanks for the ping on this and sorry for my delay, I haven't forgotten about it just that there's been some progress over here, so waiting for the dust to settle.

[chrisb13]

fyi @jbisits, @atteggiani has just moved this website to a new home:
https://access-om3-idealised-configs.readthedocs.io/
I wanted to do this sooner rather than later to minimise disruption.

We'll no longer use github previews but read the docs ones instead. (This is part of a large-scale move across the org'.) This shouldn't really change much from your perspective but do let us know if you have any issues.

[atteggiani]

Also, you might see ads in the new readthedocs-hosted websites.
This is temporary and the ads will likely be removed in the near future.

[chrisb13]

Also, you might see ads in the new readthedocs-hosted websites. This is temporary and the ads will likely be removed in the near future.

@atteggiani so I notice this isn't working:
#10 (comment)

Where is the RTD-preview? Perhaps something is missing?

(Also, can we delete this comment?)

[atteggiani]

The preview comment can be removed.

I think RTD build is not available for this PR because the RTD project was set up after this PR was created, so the webhook for this PR to RTD was not sent.
Not sure if we can get around that unfortunately.

[chrisb13]

I think RTD build is not available for this PR because the RTD project was set up after this PR was created, so the webhook for this PR to RTD was not sent.

Thanks @atteggiani. Does look like it's working with new PRs.

[jbisits]

I think RTD build is not available for this PR because the RTD project was set up after this PR was created, so the webhook for this PR to RTD was not sent.

Thanks @atteggiani. Does look like it's working with new PRs.

@chrisb13 do you want to merge and fix in a new PR with previews?

[chrisb13]

@chrisb13 do you want to merge and fix in a new PR with previews?

Yeah I was just wondering that. It's a bit of a pain not being able to view it but I guess I can build it locally.

Moreover, sorry for my delay. We're likely still some time away from having the backend fix I've discussed, it's waiting on the upstream update that we're currently doing. Once that's done, mom6 solo will be available or available to build. Are you happy to wait or would you like to see this merged? If it's the latter, we can perhaps soften some of the wording on the build page.

[jbisits]

@chrisb13 do you want to merge and fix in a new PR with previews?

Yeah I was just wondering that. It's a bit of a pain not being able to view it but I guess I can build it locally.

Moreover, sorry for my delay. We're likely still some time away from having the backend fix I've discussed, it's waiting on the upstream update that we're currently doing. Once that's done, mom6 solo will be available or available to build. Are you happy to wait or would you like to see this merged? If it's the latter, we can perhaps soften some of the wording on the build page.

I am happy with whatever you think is best; there is no hurry!