Forum Archive

So, I am writing some code using my Mac, that I am then saving on iCloud and reading into my iPad. All works well... alas...

I am using the numpy library, and though this runs perfectly on the Mac, when the script reaches the third line here:

import numpy as np

[my code here]

print(np.flip(board, 0))

I get the following error in Pythonista

Attribute error:
Module numpy has no attribute 'flip'

Can anyone offer any advice for me, please?

The method flip was introduced in numpy 1.12.0.
Pythonista ships with numpy 1.8.0, so this feature is not included. You have to wait for a new Pythonista version, which will most likely include a later version of numpy.

@themusicman Temporary, source code is here

Many thanks both! Still learning, so very grateful for your expertise.

So @cvp - may I ask how I would install that source code into Pythonista please?

@themusicman I'm really sorry, I am very far from being a Python specialist and I am sure that many others in this forum will help you, but what I could do is:

import numpy as np

def flip(m, axis):
    if not hasattr(m, 'ndim'):
        m = asarray(m)
    indexer = [slice(None)] * m.ndim
    try:
        indexer[axis] = slice(None, None, -1)
    except IndexError:
        raise ValueError("axis=%i is invalid for the %i-dimensional input array"
                                         % (axis, m.ndim))
    return m[tuple(indexer)]
np.flip = flip

board = np.arange(8).reshape((2,2,2))
print(board)
print(np.flip(board, 0))

Thanks @cvp, I simply used this instead:

import numpy as np

     reverse_board = board[::-1]
     print(reverse_board)

which reversed the individual items in the board array such that as there are 7 items in the array, board[0] becomes board[6], board[1] becomes board[5] etc...

Now to work on SpriteNodes - something I have never ever used. Need to figure out how to draw shapes et al.

@themusicman I never used it also but this forum is full of examples

I have been hitting the old NumPy wall a lot of times lately, as it is a very frequent dependency in other packages (for instance astropy and its affiliated packages like astroplan). NumPy 1.8.0 is now about 6 years old, which makes it very like to be an issue for almost any packages that requires it. Simply too old.