``DynamicalSystem`` as a function

[chaoming0625]

Please:

• Check for duplicate requests.
• Describe your goal, and if possible provide a code snippet with a motivating example.

Currently, a DynamicalSystem should define update function which receive two arguments:

• s: shared arguments
• x: individual inputs

However, it is possible to call a DynamicalSystem as a function. Because shared arguments can be got through a global function like brainpy.math.get_current_step_data().

At each step, data for the next time step can be set through:

brainpy.math.set_current_step_data()

[chaoming0625]

Through this way, there is no need to pass shared arguments s into the submodule. And new models in BrainPy can be called like a function. How do you think? @ztqakita

[ztqakita]

In my opinion, models in BrainPy can be called like a function by passing shared argument s and input x, why bother changing the update function?

[ztqakita]

However, I agree that we can use another way to deal with shared arguments s instead of treating them as update function arguments.