I am compiling this Cython code in Sage Cell Server and I get the following error.
undeclared name not builtin: array
It displays the same error in Sage Notebook. I think it is not recognizing numpy array
but it
is strange cause I have imported numpy already.
cython('''
cimport numpy as np
ctypedef np.int DTYPE
def computeDetCy(np.ndarray[DTYPE, ndim=2] matrix):
return determ(matrix,len(matrix))
cdef inline int determ(np.ndarray[DTYPE, ndim=2] matrix, int n):
cdef int det = 0
cdef int p=0
cdef int h
cdef int k
cdef int i=0
cdef int j=0
cdef np.ndarray[DTYPE, ndim=2] temp=np.zeros(4,4)
if n == 1:
return matrix[0][0]
elif n == 2:
return matrix[0][0]*matrix[1][1] - matrix[0][1]*matrix[1][0]
else:
for p in range(0, n):
h = 0
k = 0
for i in range(1, n):
for j in range(0, n):
if j==p:
continue
temp[h][k] = matrix[i][j]
k+=1
if k ==(n-1):
h+=1
k=0
det= det + matrix[0][p] * (-1)**p * determ(temp, n-1)
return det
computeDetCy(array([[13,42,43,22],[12,67,45,98],[23,91,18,54],[34,56,82,76]]))
''')
Yeah, but you imported it as np
, not importing *
(which would be a bad idea anyway) and didn't do a regular Python import. (Sometimes you have to do both a cimport
and import
, see this SO question for an example.)
However, even after
import numpy as np
and using np.array
, I still get some errors
ValueError: Buffer dtype mismatch, expected 'DTYPE' but got 'long'
So this solves your question, but isn't the whole story, and the things I tried didn't work to fix this new issue.