将一串浮点数组转换为 ndarray

Question

想象一个人必须将 ndarray 转换为字符串，然后他必须将字符串转换回 ndarray。例如，你有这个字符串：

[1.1463377809323188, -5.631205546574898, 2.2669977232047676][4.579898563059968, 7.903827618691004, -6.286433358859447]

如何将其转换为像这样的 ndarray：

[[ 1.14633778 -5.63120555  2.26699772]
 [ 4.57989856  7.90382762 -6.28643336]]

我正在考虑将 str 转换为列表，然后将列表转换为 ndarray。然而，str.split() 方法给我这个：

['[1.1463377809323188,', '-5.631205546574898,', '2.2669977232047676][4.579898563059968,', '7.903827618691004,', '-6.286433358859447]']

当我尝试split('[]') 时，列表中只剩下一个元素：

['[2.182687926670855, -6.165595264315509, 2.0514527771233135][-1.9916686860482642, 6.353279016100898, -2.237430822379459]']

Answer 1

这可行：

string = "[1.1463377809323188, -5.631205546574898, 2.2669977232047676][4.579898563059968, 7.903827618691004, -6.286433358859447]"
>>> [[float(x.strip()) for x in row.replace("[","").split(",")] for row in string.split("]") if len(row.strip())>0]
[[1.1463377809323188, -5.631205546574898, 2.2669977232047676],
 [4.579898563059968, 7.903827618691004, -6.286433358859447]]