Type, np.ndarray shape

[tsoernes]

Feature request for having ict print type in addition, and icnp (or just plain ic) ( print shape/dtype if argument is a numpy array

[woodm1979]

Not entirely sure it's necessary, but I'd +1 like the 2nd command. I'd not want to prevent people from seeing a numpy array if they wanted to see one.

[gruns]

I'm not intimately familiar with numpy, but I'll play around and see how ic() handles np.ndarrays.

But, so we're on the same page, please provide an example of what you'd like to see ic() output when provided with an np.ndarray.

[tsoernes]

Sure: \>>> a = np.random.uniform(size=(2,2)) \>>> a array([[0.96552581, 0.00289987], [0.35171792, 0.53700621]]) \>>> type(a) <class 'numpy.ndarray'> \>>> a.shape (2, 2) \>>> a.dtype dtype('float64') The shape is the array shape, i.e. the size of each dimension, and dtype the datatype of the elements. These two are often more interesting to print than the array itself when debugging (though I'm not suggesting to view *only* this information, just in addition) because functions on np arrays often change either property. 2018-02-18 7:46 GMT+01:00 gruns <notifications@github.com>:

…

I'm not intimately familiar with numpy. I'll play around with it and see how ic() handles np.ndarrays. But, so we're on the same page, please provide an example of what you'd like to see ic() output when provided with an np.ndarray. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#1 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AGd9xP5Or0JamkcNN2IvWQ2ooyn_ORE2ks5tV8c3gaJpZM4SGGzW> .

[gruns]

Digging into numpy a bit more: does numpy include a way to seraliaze np.ndarrays to strings in such a way that their type and/or shape is included?

str() and repr() (which ic() uses) don't do it.

>>> import numpy as np
>>> a = np.ndarray(shape=(2,2), dtype=float, order='F')
>>> str(a)
'[[  6.90652664e-310   1.18609510e-316]\n [  1.25709272e-316   1.18609272e-316]]'
>>> repr(a)
'array([[  6.90652664e-310,   1.18609510e-316],\n       [  1.25709272e-316,   1.18609272e-316]])'

Unfortunately I'm not familiar enough with numpy to deftly navigate dir(np.ndarray(...)).

[tsoernes]

No, I don't think that's built-in 2018-03-07 10:06 GMT+01:00 gruns <notifications@github.com>:

…

Digging into numpy a bit more: does numpy include a way to seraliaze np.ndarrays to strings in such a way that their type and/or shape is included? str() and repr() (which ic() uses) don't do it. >>> import numpy as np >>> a = np.ndarray(shape=(2,2), dtype=float, order='F') >>> str(a) '[[ 6.90652664e-310 1.18609510e-316]\n [ 1.25709272e-316 1.18609272e-316]]' >>> repr(a) 'array([[ 6.90652664e-310, 1.18609510e-316],\n [ 1.25709272e-316, 1.18609272e-316]])' Unfortunately I'm not familiar enough with numpy to deftly navigate dir(np.ndarray(...)). — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#1 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AGd9xK1Krb6opvN_b78-QuJH0dr0ES9Hks5tb6L1gaJpZM4SGGzW> .

[gruns]

numpy.ndarrays are now supported in IceCream v1.2.

>>> import numpy as np
>>> from icecream import ic
>>> 
>>> a = np.ndarray(shape=(2,2), dtype=float, order='F')
>>> ic(a)
ic| a: array([[  6.90246722e-310,   4.67090452e-317],
              [  6.90246722e-310,   4.67028793e-317]])

Also new in v1.2 is the ability to customize the argument to string serialization function. So, if you want to include the datatype and shape of ndarrays, you can now do so like

>>> import numpy as np
>>> from icecream import ic, argumentToString
>>> 
>>> def numpyAwareToString(obj):
>>>     s = argumentToString(obj)
>>>     if isinstance(obj, np.ndarray):
>>>         return f'<ndarray: type {obj.dtype}, shape {obj.shape}>\n{s}'
>>>     return s
>>> 
>>> ic.configureOutput(argToStringFunction=numpyAwareToString)
>>> 
>>> a = np.ndarray(shape=(2,2), dtype=float, order='F')
>>> ic(a)
ic| a: <ndarray: type float64, shape (2, 2)>
       array([[ 6.94547807e-310, -8.91366976e+303],
              [ 6.94547807e-310,  2.21091692e-236]])

Does this suffice for your needs, @tsoernes?

[tsoernes]

Cool, thanks! 2018-03-12 8:02 GMT+01:00 gruns <notifications@github.com>:

…

numpy.ndarrays are now supported in IceCream v1.2. >>> import numpy as np>>> from icecream import ic>>> >>> a = np.ndarray(shape=(2,2), dtype=float, order='F')>>> ic(a) ic| a: array([[ 6.90246722e-310, 4.67090452e-317], [ 6.90246722e-310, 4.67028793e-317]]) Also new in v1.2 is the ability to customize the argument to string serialization function. So, if you want to include the datatype and shape of ndarrays, you can now do so like >>> import numpy as np>>> from icecream import ic, argumentToString>>> >>> def numpyAwareToString(obj):>>> s = argumentToString(obj)>>> if isinstance(obj, np.ndarray):>>> return f'<ndarray: type {obj.dtype}, shape {obj.shape}>\n{s}'>>> return s>>> >>> ic.configureOutput(argToStringFunction=numpyAwareToString)>>> >>> a = np.ndarray(shape=(2,2), dtype=float, order='F')>>> ic(a) ic| a: <ndarray: type float64, shape (2, 2)> array([[ 6.94547807e-310, -8.91366976e+303], [ 6.94547807e-310, 2.21091692e-236]]) — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#1 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AGd9xNltZDp2gTgQ8TThQMln2HAolPzAks5tdh2UgaJpZM4SGGzW> .