Update credentials management

[Marie59]

Hi all !

I'm trying to do the update of this tool. I don't know how to inject the secret into the configfile to then get it the .cdsapirc.
Any ideas ?

Ping @bgruening @yvanlebras @PaulineSGN

[yvanlebras]

Hi Marie ! THANK YOU! If I am not saying a mistake, looking at this example https://github.com/galaxyecology/tools-ecology/pull/214/changes#diff-135d08b2c88f38c565c73c2f2ee959d66a4520c676eaa880c6b0f8d9a336d7e1 , one can remove the use of .cdsapirc and maybe configfile. Am I wrong ?

[PaulineSGN]

Hi !
Yes, you don't need configfile anymore, since the api key is directly injected in the environment variable CDS_API_KEY.

[Marie59]

Yes but then we also need to inject the URL one way or another no ?

[PaulineSGN]

Yes but then we also need to inject the URL one way or another no ?

Yes sorry, I didn't see that. Does it work manage in the python script ?

[Marie59]

For now no i'm still testing ;) But if you have any ideas don't hesitate

[Marie59]

@annefou Apparently the python version is not up to date (if I understand correctly). Does this repo need a full update ? If yes could it break other tools ?

What do you think ?
@bgruening do you have an opinion on this too ?

Btw, the tool is working locally :)

[bgruening]

Please update the CI script in a different PR. Copy the one from IUC to here and adopt it to this repo, e.g. changing the galaxyproejct/iuc

[bgruening]

Ok, now all the errors and warnings seems legit and should be worked on ideally.

[Marie59]

I don't get the error about the tests. Do you have any ideas ?

[yvanlebras]

So the idea here is to use this ""error message"" to test the execution of the command line who provides an error message from cdsapi COpernicus Python API package, waiting for Planemo to allow to use the new manner to manage credentials in Galaxy (if I am not wrong) so Planemo is not finding the Galaxy env variables related to credentials and try to find the cdsapi file and return an error on it. Am I ok @PaulineSGN @Marie59 ? Is it ok for you @bgruening ?

[bgruening]

should we not add here a proper try/except and raise an error that we can assert in the XML tests?

[bgruening]

This will also improve the situation at some point: galaxyproject/galaxy#21643

[Marie59]

Okay I guess everything is fix ! Except fort this @bgruening https://github.com/NordicESMhub/galaxy-tools/actions/runs/22145319976/job/64020879425?pr=70#step:6:20
Even thought I added the file it does not seem happy

[annefou]

I am not sure how to help @Marie59 We can definitely update the repo. We could also have only one joint repo for all our tools, (similar to what we now have with. the element channel)

[Marie59]

I am not sure how to help @Marie59 We can definitely update the repo. We could also have only one joint repo for all our tools, (similar to what we now have with. the element channel)

Hi @annefou I started to update this repo already (It should be up to date now). Yes at least for earth and climate this would be great to have a joint repo. Just need to think on how to that nicely ;)

[Marie59]

I don't know how to fix the issue here. It's working locally for me. Any ideas ? @bgruening

[bgruening]

Expected text 'CDS retrieval failed, make sure you filled in your CDS API Key' in output ('')

You assert for this given error message, but there is no error message. Maybe you are printing it to stdout? Or its not triggered?

[Marie59]

ohoh ! What do you think about this ? I think the function c.retrieve was still running even though the api key was none so I added a check to make sure the api key was not none