I'm writing a function using numpy and matplotlib.pyplot to generate a slope field based on some code I found in an old blog. That code works fine, but I was trying to make a function for it. The function takes a range for x and y, a frequency, and finally some differential equation in the form dy/dx = (some function of x and y). The differential equation is itself a function. Of course, depending on the equation, division by zero can occur. When it does, I get a
RuntimeWarning: divide by zero encountered in double_scalars
and the rest of the slope field displays in pyplot as expected, absent a value at whatever coordinate(s) caused division by zero. Finding the actual value of the equation at those coordinates is beyond the scope of the simple function, but I would obviously like to handle such cases.
The differential equation I used to test different methods of dealing with this can be described in python as:
def fun3(x, y):
return (x + y) / (x - y)
For this particular equation, the only point causing the issue is the point (0, 0). As such, the slope field returned by my function looks like:
No line at (0, 0), but that's not an issue. The problem is that before the field is displayed, I get the RuntimeWarning described above.
The code that actually generates the field is here:
for j in x:
for k in y:
slope = diff(j, k)
domain = np.linspace(j - 0.07, j + 0.07, 2)
def fun(x1, y1):
z = slope * (domain - x1) + y1
return z
field += plt.plot(domain, fun(j, k))
Where x and y are ndarrays containing the x-values and y-values, respectively, where the lines are to be drawn, including in this case (0, 0). field is what is returned by the function, a list of artists added to the plot.
My first attempt at fixing the issue was to just catch the ZeroDivisionError within the loop, before assigning slope, using:
try:
diff(j, k)
except ZeroDivisionError:
continue
Where diff is the equation described by fun3 above. However, this failed to produce the desired result and in fact changed nothing at all about the output. When I tried recreating this result outside of the function using something like,
try:
fun3(0, 0)
except ZeroDivisionError:
print('oops')
'oops' is printed as expected. What makes this different than the code inside the function that didn't work? It ultimately does the same thing by calling fun3 and checking for the ZeroDivisionError. When the nested loops eventually reach the value (0, 0), If the ZeroDivisionError does not appear, how can I handle the RuntimeWarning?
