What does "loop of ufunc does not support argument 0 of type Mul" mean?-CodePudding

I have the following problem:
I need to run the following code, but it always throws an error:

import sympy as sp
import numpy as np
import scipy.constants as const

g, t, theta, t_max, v, x_0, y_0, y_f = sp.symbols("g t theta t_max v x_0 y_0 y_f")
# Bahnkurve
x = x_0   v*sp.cos(theta)*t
y = y_0   v*sp.sin(theta)*t   sp.Rational(1,2)*g*t**2

t_end = sp.solve(sp.Eq(y,0),t)[1] # Verwerfen der ersten Möglichkeit, da diese eine negative Zeit liefert
t_end_func = sp.lambdify([y_0,v, theta], t_end.subs(g,const.g), modules="numpy")
x_func = sp.lambdify([x_0, v, theta, t], x, modules=np)
y_func = sp.lambdify([y_0, v, theta, t], y.subs(g, -const.g), modules=np)


fig1, ax1 = plt.subplots()
ax1.set_title(f"Bahnkurven mit normalverteilten $\\Theta$ und $v$")
ax1.set_xlabel("$x$")
ax1.set_ylabel("$y$")

anz_kurven = int(1e2)
for i in range(anz_kurven):
    rand_v = np.random.normal(v_0, sigma_v)
    rand_theta = np.random.normal(theta_0, sigma_theta)

    zeitpunkte = np.linspace(0, t_end_func(rand_v, rand_theta, float(50)))
    ax1.plot(x_func(100, rand_v, rand_theta, zeitpunkte), y_func(50, rand_v, rand_theta, zeitpunkte), color="#0000FF50")

Which throws the error "loop of ufunc does not support argument 0 of type Mul which has no callable sin method"

This is a problem that I encountered a few times over the last exercises I needed to solve, but sometimes it went away without doing something I remember. The same problem occurs at other points in this script I write, but there I was able to just write the function by hand, here I really want to do it the correct way.

Because @David_sd asked, here are the values I have given:

v_0 , sigma_v = 200, 1 # m/s
theta_0, sigma_theta = 45*sp.pi/180, 1*sp.pi/180 # Degrees converted to radians

CodePudding user response：

I'm still not getting any error, but I got a warning from inside the for loop:

<lambdifygenerated-22>:2: RuntimeWarning: invalid value encountered in sqrt
  return -0.101971621297793*v*sin(theta) - 0.451600755751787*sqrt(0.0509858106488964*v**2*sin(theta)**2 - y_0)

The line raising the warning is this one:

zeitpunkte = np.linspace(0, t_end_func(rand_v, rand_theta, float(50)))

Looking at how you created t_end_func, this is a function which requires the following arguments: y_0, v, theta (with this order). Looking at what you wrote, it appears you switched the arguments to v, theta, y_0.

If I modify that line of code with:

zeitpunkte = np.linspace(0, t_end_func(float(50), rand_v, rand_theta))

The warning goes away, and the plot comes out.

CodePudding user response：

The 3 lambdify create functions with this python code:

In [48]: help(t_end_func)
    ...
    def _lambdifygenerated(y_0, v, theta):
        return -0.101971621297793*v*sin(theta) - 0.451600755751787*sqrt(0.0509858106488964*v**2*sin(theta)**2 - y_0)

In [49]: help(x_func)
    ...
    def _lambdifygenerated(x_0, v, theta, t):
         return t*v*cos(theta)   x_0

In [50]: help(y_func)
    ...
    def _lambdifygenerated(y_0, v, theta, t):
        return -4.903325*t**2   t*v*sin(theta)   y_0

All have a sin(theta) call.

Called with all number arguments:

In [51]: y_func(0, 1,2, 3)
Out[51]: -41.4020327

But called with a symbol:

In [52]: y_func(0, 1,theta, 3)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
AttributeError: 'Symbol' object has no attribute 'sin'

The above exception was the direct cause of the following exception:

TypeError                                 Traceback (most recent call last)
Input In [52], in <cell line: 1>()
----> 1 y_func(0, 1,theta, 3)

File <lambdifygenerated-3>:2, in _lambdifygenerated(y_0, v, theta, t)
      1 def _lambdifygenerated(y_0, v, theta, t):
----> 2     return -4.903325*t**2   t*v*sin(theta)   y_0

TypeError: loop of ufunc does not support argument 0 of type Symbol which has no callable sin method

Or with a sympy expression:

In [53]: y_func(0, 1,2*theta, 3)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
AttributeError: 'Mul' object has no attribute 'sin'

The above exception was the direct cause of the following exception:

TypeError                                 Traceback (most recent call last)
Input In [53], in <cell line: 1>()
----> 1 y_func(0, 1,2*theta, 3)

File <lambdifygenerated-3>:2, in _lambdifygenerated(y_0, v, theta, t)
      1 def _lambdifygenerated(y_0, v, theta, t):
----> 2     return -4.903325*t**2   t*v*sin(theta)   y_0

TypeError: loop of ufunc does not support argument 0 of type Mul which has no callable sin method

Just using np.sin:

In [55]: np.sin(2*theta)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
AttributeError: 'Mul' object has no attribute 'sin'

When np.sin is given a non-array, it first converts the input to array:

In [56]: np.array(2*theta)
Out[56]: array(2*theta, dtype=object)

Math on object dtype arrays is performed iteratively, applying the appropriate operator or method to each element. In the case of ufunc like this, it tries obj.sin(), which almost never works. You have to make sure lambdified functions are called with numbers or numeric dtype arrays, never sympy objects (or arrays containing such).