Scalar Functions throw error with Parameters

[artpelling]

Scalar Functions throw error with Parameters

[pmli]

What do you want to use this for? To create a constant time-dependent parameter? That can be done with the following:

from pymor.parameters.base import Parameters

n = 2
mu = Parameters(t=1, f=n).parse({'t': 1.5, 'f': str(n * [1])})
mu       # Mu({'f': ExpressionFunction('[1, 1]', variable='t'), 't': array([1.5])})
mu['t']  # array([1.5])
mu['f']  # array([1, 1])

Although, if the time-dependent parameter is constant, it could just be a regular parameter.

from pymor.parameters.base import Parameters

n = 2
mu = Parameters(t=1, f=n).parse({'t': 1.5, 'f': n * [1]})
mu       # Mu({'f': array([1, 1]), 't': array([1.5])})
mu['t']  # array([1.5])
mu['f']  # array([1, 1])

It would probably still be good for parse to give better error messages or even make it work with ConstantFunctions.

[artpelling]

What do you want to use this for? To create a constant time-dependent parameter?

Thanks for the snippets. I am using it here as an input to compute which only accepts a NumPy array or Function.

[sdrave]

What do you want to use this for? To create a constant time-dependent parameter?

Thanks for the snippets. I am using it here as an input to compute which only accepts a NumPy array or Function.

I'm confused. Wouldn't it be more straightforward (and efficient) to use input = np.zeros(self.dim_input)?

[artpelling]

I'm confused. Wouldn't it be more straightforward (and efficient) to use input = np.zeros(self.dim_input)?

Good point. In this case yes. In impulse_resp, however, I need a Function.