Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Naming convention for modified standard recipes #1

Open
TomasTorsvik opened this issue Mar 10, 2021 · 1 comment
Open

Naming convention for modified standard recipes #1

TomasTorsvik opened this issue Mar 10, 2021 · 1 comment

Comments

@TomasTorsvik
Copy link
Collaborator

How to name modified standard recipes in the most sensible way?

The "tested_recipes" directory contains some recipes that have been modified based on standard recipes. I think it makes sense to keep the main part of the recipe name and add a label, i.e. <standard_recipe_name>_<new_label>.yml, but it's not clear what is the best approach. Here is a suggestion based on some typical situations

  1. The recipe compares a single model output with observational data. In this case we could use the model name as label
    ex.: recipe_modes_of_variability.yml --> recipe_modes_of_variability_NorESM1-M.yml
  2. A recipe constructs a multi-model product. In this case we probably want dataset from the same CMIP
    ex.: recipe_ocean_example.yml --> recipe_ocean_example_CMIP5.yml
  • Problem: "CMIP5" is already used as a label in perfmetrics recipes, maybe "CMIP6" will also appear, causing confusion between standard and modified recipes
  1. A recipe may contain dataset from both CMIP5 and CMIP6, in particular if we want to look at the evolution of model performance, so CMIP5/6 may not be appropriate in this case.

NorESM1 and NorESM2 does not correspond to CMIP5 and CMIP6, respectively (CMIP6 contains NorESM1-F datasets), so this can be confusing as well.

Are there other labels we could use that would make more sense?

@YanchunHe
Copy link
Collaborator

YanchunHe commented Mar 10, 2021

The ESMValTool is designed for multiple models comparison (>=1), and can work across CMIP5/CMIP6 dataset. Therefore, it is not sensible to have model name or CMIP5/6 as suffix in the recipe names. Except in the case that it is not able to combine different models (which should not be the case) or CMIP phases (that is possible at some point).

Therefore, I would not have any suffix. It is good to include as many as possible different model output and observational data activated in the tested recipes, but it can also also included (activated) one or several models. There are many options in one recipe, so it is the end-users' choice to switch on and off some of them. We put the 'tested_recipes" so that it can run without any modification. They are just a start for other users to adapt to their own needs, e.g., use more model, switch on/off additional diagnostics.

In some cases, we can have, e.g., CMIP5/CMIP6 in the suffix, as I found that it may take quite some time to combine them, as the ESMValTool does not support this well. In the long run, this should also be fixed.

In short, I would suggest no suffix is used but keep the same name as the standard recipes; Use only CMIP5/CMIP6 where it is necessary.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants