inplace usage of include leads to unexpected behavior

[younesStrittmatter]

here is a code snippet to explain the issue (maybe it is the expected behavior, but could lead to errors down the line):

# define include (e.g., to be used for multiple models)
include = [
        {
            "name": "v",
            "prior": {
                "Intercept": {"name": "Uniform", "lower": -3.0, "upper": 3.0},
                "coherence": {"name": "Uniform", "lower": 0., "upper": 1.0},
            },
            "formula": "v ~ 1 + coherence",
            "link": "identity",
        }
    ]
    print('before')
    print(include)
    model_ddm = hssm.HSSM(
        data=data.copy(),
        model='ddm',
        include=i_,
    ).sample(sampler="nuts_numpyro", chains=chains, cores=cores, draws=draws, tune=tune)
    print('after')
    print(include)

leads to:

    before
    [{'name': 'v', 'prior': {'Intercept': {'name': 'Uniform', 'lower': -3.0, 'upper': 3.0}, 'coherence': {'name': 'Uniform', 'lower': 0.0, 'upper': 1.0}}, 'formula': 'v ~ 1 + coherence', 'link': 'identity'}]
    
    after
    [{'name': 'v', 'prior': {'Intercept': {'name': 'Uniform', 'lower': -3.0, 'upper': 3.0}, 'coherence': {'name': 'Uniform', 'lower': 0.0, 'upper': 1.0}}, 'formula': 'v ~ 1 + coherence', 'link': 'identity', 'bounds': (-inf, inf)}, {'name': 'p_outlier', 'prior': 0.05, 'bounds': None}]

If you want to use the same inclue in a different model, this then leads to an error:

ValueError: Please do not specify `p_outlier` in `include`. Please specify it with `p_outlier` instead.

[digicosmos86]

Thanks for bringing this to our attention. Yes, the list used for include will be modified in the initializer of hssm.HSSM inplace and cannot be reused. This is a bug. We'll start an issue and fix this.