Uq[Py]Lab v1.0 - How to assign different options for inference of two (or more) marginal distributions

UQ with UQLab Community Q&A and How To
,
1

Hello.
I wrote a test script in Python which I wanted to perform inference of marginals based on data generated from defined distributions (in analogy to the example in uqlab docs).
There are two random variables and I would like to set different parameters of the inference analysis for each of them.
According to the instructions, this can be done with syntax e.g.

"iOpts['Marginals'][ii][Bounds]".

When I try to do this, I get a Python error:

InputOpts2["Marginals"][1] = {

IndexError: list assignment index out of range

which means that the index ii=1 was not found.
Can I ask for help on how to define different parameters for more than one input variable?

Here’s the script:

from uqpylab import sessions, display_util

mySession = sessions.cloud()
uq = mySession.cli
mySession.reset()

display_util.load_plt_defaults()
uq_colors = display_util.get_uq_color_order()

uq.rng(100, 'twister')

InputOpts1 = {
    'Marginals': [
        {
            'Type': 'Lognormal',
            'Moments': [3,1.2],
            'Bounds': [0,3.6]
        },
        {
            'Type': 'Beta',
            'Parameters': [6,4],
            'Bounds': [0,0.8]
        }
    ]
}

myInput = uq.createInput(InputOpts1)
uq.print(myInput)
uq.display(myInput)

X_LHS = uq.getSample(N=1000, Method='LHS')

InputOpts2 = {
    "Copula": {"Type": "Independent"},
    "Inference": {
        "Data": X_LHS.tolist()
        },
    "Marginals": [{
        "Inference": {
            "Criterion": 'ML'
        }
    }]
}

InputOpts2["Marginals"][0] = {
        "Type": ['Gaussian', 'Exponential', 'Weibull', 'LogNormal'],
        "Bounds": [0,3.6]
}

InputOpts2["Marginals"][1] = {
        "Type": ['Gaussian', 'LogNormal'],
        'Bounds': [0,0.8]
}

InputHat1 = uq.createInput(InputOpts2)
fig = uq.display(InputHat1)
uq.print(InputHat1)

mySession.quit()
2

Ok, I have managed to solve this.

The syntax

iOpts['Marginals'][1]['Bounds']

will not work unless both marginals are specified in analogy to defining input marginals:

InputOpts2 = {
    "Copula": {"Type": "Independent"},
    "Inference": {
        "Data": X_LHS.tolist()
        },
    "Marginals": [{
        "Type": ['Gaussian', 'Beta', 'Exponential', 'Weibull', 'LogNormal'],
        'Bounds': [0,3.6]
    },
    {
        "Type": ['Gaussian', 'LogNormal'],
        'Bounds': [0,0.8]
    }]
}