[Neuroimaging] positivity constraint in mapmri crashes

[Rutger Fick]

Hi Tugan,

This actually has to do with upgrading to the latest version of cvxpy. It
has to do with the following line:

https://github.com/nipy/dipy/blob/9994c2a83f49a3e909b569051326b58f04a14bb2/dipy/reconst/mapmri.py#L405


In an older version of the maprmi code this was a 'strict inequality' as in
only a '>' sign, but cvxpy now now only accepts larger or equal than '>='.

Updating dipy to the latest version should fix this line to what it should
be.

best,
Rutger


On 9 July 2018 at 18:29, Muftuler, L. Tugan <lmuftuler at mcw.edu> wrote:

>
> Hi,
>
> I am running mapmri on ex-vivo brain data. When I choose "positivity
> constraint" i get a cvxpy error :
>
>
>
> File "/data/data/Anaconda/lib/python2.7/site-packages/cvxpy/expressions/expression.py",
> line 502, in __gt__
>     raise NotImplementedError("Strict inequalities are not allowed.")
> NotImplementedError: Strict inequalities are not allowed.
>
> I updated cvxpy to the latest, but it still fails. Did anyone else
> encounter this? I would be grateful if anyone could offer a solution.
>
> best,
> -tugan
> _______________________________________________
> Neuroimaging mailing list
> Neuroimaging at python.org
> https://mail.python.org/mailman/listinfo/neuroimaging
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/neuroimaging/attachments/20180710/fcbdc9cf/attachment.html>