numpy subclassing

jpecci

I would like to ask you a quick question:
I am facing the opposite problem described here:
http://docs.scipy.org/doc/numpy/user...bclassing.html

In short, I want to create a class Dummy(np.ndarray) which returns a plain array whenever is sliced or viewed. I cannot figure out how.

I would really appreciate you help, here is a chunk of code, what is commented was an attempt which didn’t work.
Thanks a lot,
Jacopo

Expand|Select|Wrap|Line Numbers

 
import numpy as np 
 
class Dummy(np.ndarray): 
 
    def __del__(self): 

        print "__del__" 

    def __new__(cls, array): 

        print "__new__" 

        obj=array.view(cls) 

        return obj 

    def __array_finalize__(self, obj): 

        print "__array_finalize__" 

        #self=self.view(np.ndarray) 

        #self=np.asarray(self) 

    def __repr__(self): 

        return "%s"%np.asarray(self) 
 
p=Dummy(np.ones(5)) 

print type(p)

bvdet

I don't know much about Numpy. If I understand you correctly, you want a slice of an instance of your subclass to return a list or tuple. Try defining a __getslice__() method.

Expand|Select|Wrap|Line Numbers

 import numpy as np
 
class N(np.ndarray):
 
    def __new__(cls, obj): 

        return obj.view(cls) 
 
    def __getslice__(self,i=None,j=None):

        return tuple(self.view()[i:j:])
 
p = N(np.ones(9))

print p

print p[1:3]

print p[:3]

print p[3:]

print p[1:8:3]

Output:

Expand|Select|Wrap|Line Numbers

 
>>> [ 1.  1.  1.  1.  1.  1.  1.  1.  1.]

(1.0, 1.0)

(1.0, 1.0, 1.0)

(1.0, 1.0, 1.0, 1.0, 1.0, 1.0)

[ 1.  1.  1.]

It does not work for a slice with a stride. The __getslice__() method has been deprecated since version 2.0.

jpecci

thanks a lot. I think I will do something on these lines even if
I was hoping in something which didnt involve getslice() since it has been deprecated.
jacopo

numpy subclassing

re: numpy subclassing

re: numpy subclassing

Similar Python bytes

/bytes/development

/bytes/it

/bytes/about