Subclass from sympy.Function with predefined variable-CodePudding

I'm trying to make a new class called MyFunction which is an One Variable Function.

It inherits from sympy.Function to get the same properties, but I'm having problems with the __new__ method.

My goal is to make it work

import sympy as sp
# class definition
x = MyFunction("x")
dx = x.dt

The equivalent code, which already works:

import sympy as sp
t = sp.symbols("t")
x = sp.Function("x")(t)
dx = sp.diff(x, t)

The class that inherits from sp.Function

import sympy as sp

class MyFunction(sp.Function):

    def __new__(cls, name):
        t = sp.symbols("t")
        # Now we do just like 'sp.Function("x")(t)'
        self = super().__new__(cls, name)(t)  # Error
        self.time = t
        return self

    @property
    def dt(self):
        return sp.diff(self, self.time)


x = MyFunction("x")
dx = x.dt
print(x)  # expected print 'x(t)'
print(dx)  # expected print 'Derivative(x(t), t)'

The received error is

self = super().__new__(cls, name)(t)  # Error
TypeError: 'MyFunction' object is not callable

I tried to inherit from an another class (like sp.core.function.UndefinedFunction), but it gives the error

AttributeError: 'x' object has no attribute 'dt'

Now I have no idea how to proceed and solve it.

CodePudding user response：

If I understand you correctly:

import sympy as sp

class MyFunction(sp.Function):

    def __new__(cls, name):
        t = sp.symbols("t")
        self = sp.Function(name)(t)
        self.time = t
        self.dt = sp.diff(self, self.time)
        return self

x = MyFunction("x")
dx = x.dt

print(x)
print(dx)

Result:

x(t)
Derivative(x(t), t)