Converting for loop from matlab to python

I am converting some MATLAB code in to python and have the encountered the error "ValueError Traceback (most recent call last) in 1 for ig in range(nbas): ----> 2 psi[:,ig] = np.polyval(np.array(pp[ig,:ig]),nodes)

ValueError: could not broadcast input array from shape (56,1) into shape (56) " I cannot find the issue with following code:

import numpy as np
import scipy.special as scl

##basis parameters (nbas<ngrid+1)
nbas = 54 #Basis set size in one dimension

#Hermite matrix
def hermipol(n):
p = np.zeros((n+1,n+1))
p[0][0] = 1
if n == 0:
p = np.array([[1,0],[2,0]], dtype = float)
if n > 0:
p[1][range(0,2)] = np.array([2,0])
if n >=1:
for k in range(2,n+1):
p[k][range(n)] = 2*p[k-1][range(0,n)]
p[k][range(2,n+1)] += -2*(k-1)*p[k-2][range(0,n-1)]
for i in range(0,n+1):
p[i,:] /= np.sqrt(np.sqrt(np.pi)*2**(i)*scl.factorial(i))

return(p)
#Generation Gauss-Hermite Quadrature nodes and weights
return(np.polynomial.hermite.hermgauss(n))

pp = hermipol(ngrid)

#Evaluation and store psi_i()(x,y)_k)
nodes = np.array([nodes]).T
weights = np.array([weights]).T
print(weights)
print(nodes)
np.shape(nodes)
psi = np.zeros((ngrid,nbas))
np.shape(psi)

for ig in range(nbas):
psi[:,ig] = np.polyval(np.array(pp[ig,:ig]),nodes)



Any guidance would be greatly appreciated.

Thanks!

This post is better suited for Stackoverflow I think. Anyway you can simply solve your problem by changing psi[:,ig]to psi[:,ig:ig+1]. Then the left-hand side is truly a nx1 matrix, not just a vector of size n.
Or you can remove the line nodes = np.array([nodes]).T which is useless here, and causes Numpy to transform the array "nodes" (size n), into a 1xn array. Then the original psi[:,ig] works.