Datasets name changes

[Cmurilochem]

Hi @scarlehoff.
When testing the calculation of phi in the hyperopt_loss branch, I have the following:

def get_experimental_data(dataset_name="NMC", theoryid=399):
    """Get experimental data set using validphys.

    Returns a list defined by the data set as
    `validphys.core.DataGroupSpec`.
    """
    loader = Loader()
    ds = loader.check_dataset(dataset_name, theoryid=theoryid, cuts="internal")
    return loader.check_experiment("My DataGroupSpec", [ds])

After rebasing this morning, I have an error:

E               validphys.loader.LoaderError: Trying to use a new theory with an old commondata format, surely it must be a mistake: NMC

/usr/share/miniconda/envs/test/lib/python3.12/site-packages/validphys/loader.py:728: LoaderError
----------------------------- Captured stderr call -----------------------------
Warning: : Error trying to read NMC, falling back to the old format reader
Error: : Trying to read NMC in the old format. Note that this is deprecated and will be removed in future releases

Do you have any clue on how to solve this?
Thank you!

[Radonirinaunimi]

Hi @Cmurilochem, with the new theories, the dataset names should be replaced by their new names according to the following mapping. So in your particular example, it should be instead NMC_NC_NOTFIXED_P_EM-SIGMARED.

Edit: that being said, I am not sure if in the case above raising an error is the intended behavior.

[Cmurilochem]

Thanks @Radonirinaunimi. With the new name, my test still fails with a new error: validobj.errors.ValidationError: Missing data_uncertainties for NMC_NC_NOTFIXED_P_EM
Is this unexpected indeed ?

[scarlehoff]

Hi @Cmurilochem this is due to these datasets being only legacy datasets.

You can this runcard instead: 240301-jcm-001_new_new (vp-get fit 240301-jcm-001_new_new)
This is the whole set of datasets of the current standard fit with the new names.

[scarlehoff]

The new names are now in the runcard in #1986