iter for single value....

[jklymak]

Perhaps related to #55, but I can't see where that was resolved...

I have the input to a function that can either be a single value or an array. I just want the first value of the array or the value. Note that this has to work with or without pint.

If I do

import pint

ureg = pint.UnitRegistry()
ureg.setup_matplotlib(True)

width = 0.25 * ureg.cm
print(np.iterable(width))
print(width[0])

I get:

True
Traceback (most recent call last):
  File "/Users/jklymak/pint/pint/quantity.py", line 1400, in __getitem__
    value = self._magnitude[key]
TypeError: 'float' object is not subscriptable

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "testBar.py", line 13, in <module>
    print(width[0])
  File "/Users/jklymak/pint/pint/quantity.py", line 1404, in __getitem__
    "supports indexing".format(self._magnitude))
TypeError: Neither Quantity object nor its magnitude (0.25)supports indexing

Note that iter(width) also works fine...

So maybe I'm just missing something, but how do I check if width is a singleton?

[hgrecco]

This because pint (currently) uses a single type for iterable and non-iterable object. Quantity does has a __iter__ method and a __getitem__ method which is implemented by dispatching to the underlying type. I think that iter just checks that these exists and therefore works. You should check the magnitude print(np.iterable(width.m))

[jklymak]

Ok, but if I do

b = array(3)
iter(b)
<ipython-input-10-4e62f6ba0650> in <module>
----> 1 iter(b)

TypeError: iteration over a 0-d array

If I do the same with a pint object...

In [11]: a = 3 * ureg.cm

In [12]: iter(a)
Out[12]: <iterator at 0x116c39390>

I'd argue pint should do the same thing as numpy...

Agreed that a.m doesn't iterate, but we can't assume a has a method m. We assume that a is something like a list, or numpy array, or a singleton, and that generally works for pint except for the singleton.

[hgrecco]

True. According to the iter docs

object must be a collection object which supports the
iteration protocol (the iter() method), or it must support the sequence
protocol (the getitem() method with integer arguments starting at 0).
If it does not support either of those protocols, TypeError is raised.

Therefore, I think that could be accomplished by doing for __getitem__ the same trick we do for __iter__ here: https://github.com/hgrecco/pint/blob/master/pint/quantity.py#L136

[crusaderky]

The fix seems straightforward - just need to change __iter__ to return an iterator, as opposed to being an iterator function itself.

class Wrapper:
    def __init__(self, inner):
        self.inner = inner

    def __iter__(self):
        it_inner = iter(self.inner)

        def it_outer():
            for elem in it_inner:
                yield (elem, "wrapped")

        return it_outer()


a = Wrapper([1, 2])
print(iter(a))
print(list(iter(a)))
b = Wrapper(1)
iter(b)

Output:

<generator object Wrapper.__iter__.<locals>.it_outer at 0x7fcc4ffb2468>
[(1, 'wrapped'), (2, 'wrapped')]
Traceback (most recent call last):
  File "iter_wrap.py", line 20, in <module>
    iter(b)
  File "iter_wrap.py", line 6, in __iter__
    it_inner = iter(self.inner)
TypeError: 'int' object is not iterable

[jthielen]

@crusaderky Thanks for catching that! I can't believe I missed that. I put in #891 with your suggested fix.