oat_dashboard - if parameter is not a uniform distribution, range function returns always nan

[ntropy-esa]

Hello,
We continue our work with lca_algebraic and found a few things to fix. Not worth making a pull request for that, though.

1/ Just noticed a typo in the current master branch

In file: params.py
Line: 42
"min": data.get("mininum"),
Should read: minimum

2/ When using loadParams() for the type 7 - Discrete uniform uncertainty

I have defined a boolean (not an enumeration) but it gets detected as an enumeration

File: params.py
Line: 476
if data.get('scale') is not None :
# Enum param : group them by common prefix

In my case, even though it's not an enum param, I don't have a None as scale, but a nan
print(data.get('scale')) > returns nan

Just think we need to change that if statement, for me this works:

             if not np.isnan(data.get('scale')) and data.get('scale') is not None:

3/ And finally, when using the function oat_dashboard in stats.py with a parameter having a normal distribution, the method .range(n) returns always nan, because normal params don't have a min and max. But it works when I use a param with uniform distribution.

File: stats.py
Line: 118

# Compute range of values for given param
params[varying_param.name] = varying_param.range(n)

File: params.py
Line: 160

def range(self, n):
    '''Used for parametric analysis'''
    step = (self.max - self.min) / (n - 1)
    return list(i * step + self.min for i in range(0, n))

I think this function should distinguish between the distribution types, and if it does not have a min/max set, we can find a way of defining of min/max: e.g. for a normal distribution, min/max could be mean -/+ 2*standard deviation ? Similar for other distribution types? What do you think?

Worth noting: we work with parameters that have been defined in the activity-browser, and then imported in the notebook with the new function loadParams()

/Elias