bug-report jupiter's ipython notebook

[ethan1987-2]

i was scripting with spyder 3 and using ipython and noticed that jupiter's nb can't work this:
import scipy as sp
...
sp.integrate.nquad(f,intervs)

it yields and error stating that there is no integrate function in module sp

but if you do:
import integrate from scipy
...
integrate.nquad(f,intervs)

then there is no problem at all!

[Carreau]

sp.integrate is a submodule of SciPy, so you need to import it once for it to be available. That's how python work so I'd be worried about changing the behavior. The confusion is likely that many modules do implicitly import their submodules, which makes things like `import foo; foo.bar.func()` appear to be correct, though it is likely not. One classic example is matplotlib that used to have a `matplotlib.finance` submodule. It is - I agree - a behavior of python that may be confusing at first sight but is the reasonable one. Sorry I'm on my phone so code examples are hard to write.

…

On Jun 8, 2017 11:06, "ethan1987-2" ***@***.***> wrote: i was scripting with spyder 3 and using ipython and noticed that jupiter's nb can't work this: import scipy as sp ... sp.integrate.nquad(f,intervs) it yields and error stating that there is no integrate function in module sp but if you do: import integrate from scipy ... integrate.nquad(f,intervs) then there is no problem at all! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#10646>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAUez7PEdc_kHmgrFnkKy9a2fbTntC_xks5sCDg3gaJpZM4N0bX_> .

[ethan1987-2]

but then how come there's no problem doing:
import scipy as sp
...
sp.integrate.nquad(f,intervs)
in spyder 3 but it is a problem in jupyter...it's like you're coding on the same language but not at the same time. I do run code like that in spyder with no problems at all and then it doesn't work on jupyter. it doesn't make sense

[Carreau]

Well because something else does import it.
I don't know what, I don't have the code to inspect, but it's easy to replicate:

$ cat foo.py
import scipy.integrate

$ python
Python 3.6.0 |Continuum Analytics, Inc.| (default, Dec 23 2016, 13:19:00)
[GCC 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import scipy as sp
>>> sp.integrate
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: module 'scipy' has no attribute 'integrate'
>>> import foo
>>> sp.integrate
<module 'scipy.integrate' from '/Users/bussonniermatthias/anaconda/lib/python3.6/site-packages/scipy/integrate/__init__.py'>
>>>

So if something else is importing it, it will be available, and as you see it's Python behavior not IPython or Jupyter. Spyder may do imports as well, maybe it import scipy.integrate. @ccordoba12 an idea ofwhere this difference could come from on the Spyder Side ?