Re: [Visualpython-users] Let's hash out the documentation for restrictive imports

[James M. <mu...@pi...>]

Bruce,

On Oct 7, 2010, at 3:25 PM, Bruce Sherwood wrote:

> The issue is that importing visual also imports math and numpy, and
> does some tricky stuff to make sure that, for example, sqrt(scalar)
> uses the fast math.sqrt routine, not the numpy.sqrt routine which is
> much slower when acting on scalars.

Is this really an issue?  If you do

> import numpy as np
> import scipy as sp
> import visual as vp
> 
> a=np.sqrt(5)

Does the fact that the vp namespace also includes numpy and math affect this (other than bloat)?

on the other hand

> from visual import *
> from numpy import *

Now you have replaced the visual version of sqrt with its check, with just the numpy one.

I think what you are getting at is a questions of whether it is OK for visual to import math and numpy into its own namespace, and if not, how should visual do this branch?  I don't really consider myself enough of an expert to comment.  Perhaps someone else does. 

Actually it looks like visual imports cmath as well.
> >>> a=-1
> >>> vp.sqrt([a])
> array([ NaN])
> >>> vp.sqrt(a)
> 
> Traceback (most recent call last):
>   File "<pyshell#16>", line 1, in <module>
>     vp.sqrt(a)
>   File "/Library/Frameworks/Python.framework/Versions/Current/lib/python2.6/site-packages/visual/__init__.py", line 45, in _uf
>     if type(x) in mathtypes: return math(x)
> ValueError: math domain error
> >>> b=-1+0j
> >>> vp.sqrt(b)
> 1j
> >>> vp.sqrt([b])
> array([ 0.+1.j])

-Jim