Time-dependent Envelope

[filliong]

These commits implement a new class Envelope that is used to generate a time envelope on the electromagnetic field. Each Envelope model is a function with values varying in time in [0,1]. This functions multiplies the electromagnetic field. The targeted application of the envelope is to ensure that the electromagnetic is zero at the initial time. Right now, there are two envelope models: NoEnvelope and EnvelopeHann (cos squared). Other models can be simply added using inheritance.

[joeydumont]

The envelope should NOT be added in the Particle constructor. While it is badly named, in retrospect, Particle describes the motion of particle in a given electromagnetic field. The envelope is a property of the electromagnetic field, NOT of the motion of the particle.

This breaks backwards compatibility, and forces us to make unnecessary changes to all of the drivers and tests.

As an additional property of an electromagnetic field, Envelope should be a full-fledged electromagnetic field in its own right. This can be achieved through composition (see the StrattoCalculator's field models for examples). In short, you'd have something like this

template <class FieldModel>
class EnvelopeFieldModel
{
public:
    EnvelopeFieldModel(FieldModel & my_field_model,
                                      my_envelope_field_model_parameters...)
    {
    ...
    }

    std::array<double,6> ComputeFieldComponents(t,x,y,z) const
    {
        my_field = my_field_model.ComputeFieldComponents(t,x,y,z);
        
        for (auto&& element : my_field)
        {
            element *= envelope_value(t)
        }
        ....
    }
};

[joeydumont]

The new plot seems like a good idea. Would you mind showing an example?

[filliong]

The new plot seems like a good idea. Would you mind showing an example?