numpy常用操作

print("######demo0 python-based array op ######")
z0 = [[1, 2], [3, 4]]
z1 = [[1, 1], [1, 2]]
print("add:", z0 + z1)
# print("multiply:", z0 * z1)  # not supported

print("######demo1 basic array op ######")
X0 = np.array([[1, 2], [3, 4]])
X1 = np.array([[1, 1], [1, 2]])
print("add:", X0 + X1)
print("multiply:", X0 * X1)

print("######demo2 basic matrix op ######")
Y0 = np.matrix([[1, 2], [3, 4]])
Y1 = np.matrix([[1, 1], [1, 2]])
print("add:", Y0 + Y1)
print("multiply:", Y0 * Y1)

print("######demo3 stack ######")
arr = [np.random.randn(1, 2) for _ in range(3)]
print("shape:", np.shape(arr), arr)
print("stack axis 0 , shape:", np.shape(np.stack(arr, axis=0)))
print("stack axis 1 , shape:", np.shape(np.stack(arr, axis=1)), np.stack(arr, axis=1))
print("stack axis 2 , shape:", np.shape(np.stack(arr, axis=-1)))

print("######demo4 vstack ######")
arrv1 = [1, 2]
arrv2 = [3, 4]
print("vstack, shape", np.shape(np.vstack((arrv1, arrv2))), np.vstack((arrv1, arrv2)))

print("######demo5 hstack ######")
arrh1 = [1, 2]
arrh2 = [3, 4]
print("hstack, shape", np.shape(np.hstack((arrv1, arrv2))), np.hstack((arrv1, arrv2)))

print("######demo6 concatenate ######")
a1 = [[[1, 2, 3, 4]]]
a2 = [[[4, 5, 6, 7]]]
print("concatenate by axis 0:")
concate0 = np.concatenate((a1, a2))
print(concate0, ", shape:", np.shape(concate0))
concate1 = np.concatenate((a1, a2), axis=1)
print(concate1, ", shape:", np.shape(concate1))
concate2 = np.concatenate((a1, a2), axis=-1)
print(concate2, ", shape:", np.shape(concate2))

print("######demo7 multivariate_normal ######")
mean = np.array([0, 1])
cov = np.eye(2)
print("multivariate_normal: ", np.random.multivariate_normal(mean, cov, 3))

print("######demo8 argmax ######")
array8 = np.array([[2, 0], [5, 1], [6, 3]])
print("src array:", array8)
print("argmax by axis 0:", np.argmax(array8, axis=0))
print("argmax by axis 1:", np.argmax(array8, axis=1))

print("######demo9 count_nonzero ######")
array9 = np.array([[3, 0], [5, 1], [3, 0]])
print("array9 - array8 (axis=0):", np.count_nonzero(array9 - array8, axis=0))
print("array9 - array8 (axis=1):", np.count_nonzero(array9 - array8, axis=1))

print("######demo10 zero ######")
print(' np.zeros(3):', np.zeros(3))
print(' np.zeros([1,2]):', np.zeros([1,2]))