Passing array range as argument to a function?

Python accepts the 1:5 syntax only within square brackets. The interpreter converts it into a slice object. The __getitem__ method of the object then applies the slice.

Look at numpy/lib/index_tricks.py for some functions that take advantage of this. Actually they aren't functions, but rather classes that define their own __getitem__ methods. That file may give you ideas.

But if you aren't up to that, then possibilities include:

blah(arr, slice(1, 5))
blah(arr, np.r_[1:5])

nd_grid, mgrid, ogrid extend the 'slice' concept to accept an imaginary 'step' value:

mgrid[-1:1:5j]
# array([-1. , -0.5,  0. ,  0.5,  1. ])

Just be aware that anything which expands on a slice before passing it to your blah function, won't know about the shape of the other argument. So np.r_[:-1] just returns [].

And None can be used in slice: e.g. slice(None,None,-1) is equivalent of [::-1].

you can try like this:

def blah(ary, arg):
    arg = map(int, arg.split(":"))
    print ary[arg[0]:arg[1]]

blah([1,2,3,4,5,6],"2:5")

output:

[3, 4, 5]

You can use the slice function

>>> def blah(ary,arg1):
...     print ary[arg1]
>>> blah(range(10), slice(1, 5))
[1, 2, 3, 4]