[SciPy-user] strange interaction between scipy.stats and scipy.optimize.minpack

[Nils Wagner]

On Mon, 3 Dec 2007 23:32:07 +0100
  "Giulio Bottazzi" <giulio.bottazzi at gmail.com> wrote:
> Hi all,
> I don't know if it has been already reported but I've 
>noticed a
> strange behavior. Consider the following script
> 
> #START of fit.py ----------------------
> 
> from numpy import array
> import scipy.stats
> from scipy.optimize.minpack import leastsq
> 
> empfreq=array([.31,.57,.12])
> 
> def thempdiff(p):
>    thfreq   = array([(1-p[0])*(1-p[0]), 
>2*(1-p[0])*p[0],p[0]*p[0] ])
>    return thfreq - empfreq
> 
> print leastsq(thempdiff,x0=[.5], full_output=1)
> 
> #END of fit.py -------------------------
> 
> now running it I obtain a seg fault
> 
>> python fit.py
> zsh: segmentation fault  python fit.py
> 
> If instead
> 
> 1) I remove the line "import scipy.stats" or
> 
> 2) I remove the option "full_output=1" from leastsq
> 
> the script runs smoothly. I don't understand what's 
>going on. Is
> somebody else able to reproduce it? Any help?
> 
> Giulio.
> 
> -- 
> Giulio Bottazzi <giulio.bottazzi at gmail.com>
> http://giulio.bottazzi.googlepages.com
> PGP Key ID:BAB0A33F
> _______________________________________________
> SciPy-user mailing list
> SciPy-user at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user

  
I cannot reproduce your problem.

>>> numpy.__version__
'1.0.5.dev4528'
>>> scipy.__version__
'0.7.0.dev3610'


python  fit.py
(0.42369157713188349, array([[ 0.46734154]]), {'qtf': 
array([ -4.51744596e-06]), 'nfev': 17, 'fjac': array([[ 
1.46279277, -0.20867347, -0.5792871 ]]), 'fvec': array([ 
0.0221314 , -0.08164595,  0.05951455]), 'ipvt': array([1], 
dtype=int32)}, 'Both actual andpredicted relative 
reductions in the sum of squares\n  are at most 0.000000', 
1)


Nils