Custom printing functions with sympy - sympy

Question

I would like to do some math with sympy. However, working with functions gives a very messy output.
Take the example of the chain rule.
import sympy as sym
def ddt (s):
return sym.diff(s, t)
x, y, z, t = sym.symbols("x y z t")
a = sym.Function('a')(x,y,z,t)
b = sym.Function('b')(x,y,z,t)
c = sym.Function('c')(a, b)
print (ddt(c))
This prints all the dependencies of the variables:
Derivative(a(x, y, z, t), t)*Derivative(c(a(x, y, z, t), b(x, y, z, t)), a(x, y, z, t)) + Derivative(b(x, y, z, t), t)*Derivative(c(a(x, y, z, t), b(x, y, z, t)), b(x, y, z, t))
Anyway of removing the dependencies? (x,y,z,t) do not appear in the output:
And get something like:
Derivative(a, t)*Derivative(c, a) + Derivative(b, t)*Derivative(c, b)
https://latex.codecogs.com/svg.image?\frac{\partial&space;a}{\partial&space;t}\frac{\partial&space;c}{\partial&space;a}&space;+&space;\frac{\partial&space;b}{\partial&space;t}\frac{\partial&space;c}{\partial&space;b}&space;

Answer 1

You have to create a custom printer, which quickly become a tedious job. A basic implementation is this:
from sympy.printing import StrPrinter
from sympy.core.function import AppliedUndef
class MyStrPrinterPrinter(StrPrinter):
""" Extended Latex printer with new options:
1. applied_no_args=False: wheter applied undefined function should be
rendered with their arguments. Default to False.
"""
def __init__(self, settings=None):
self._default_settings.update({
"applied_no_args": False,
}
)
super().__init__(settings)
def _print_Function(self, expr):
if isinstance(expr, AppliedUndef) and self._settings["applied_no_args"]:
return expr.func.__name__
return super()._print_Function(expr)
def my_str(expr, **settings):
return MyStrPrinterPrinter(settings).doprint(expr)
def myprint(expr):
print(my_str(expr, applied_no_args=True))
Then, on your code:
import sympy as sym
sym.init_printing(
def ddt (s):
return sym.diff(s, t)
x, y, z, t = sym.symbols("x y z t")
a = sym.Function('a')(x,y,z,t)
b = sym.Function('b')(x,y,z,t)
c = sym.Function('c')(a, b)
myprint(ddt(c))
# out: Derivative(a, t)*Derivative(c, a) + Derivative(b, t)*Derivative(c, b)
Theoretically, you could execute this line of code at the beginning of your code block, and then it should use the new printer:
sym.init_printing(str_printer=my_str)
Unfortunately, I'm unable to get it working on Jupyter. That's why I created the myprint function.