[Nipy-devel] attributes error

[Brian Hawthorne]

jonathan,
for the declaration of the slice attribute, set implements=list,slice.

regarding modifying the value of slice, i highly recommend simply
constructing
a new SliceIteratorNext with the new slice value instead of modifying the
existing one.
in general, it's a good idea to modify objects as little as possible and
favor copying.
a good bit of the work i've done one the codebase so far is oriented toward
shifting from a modify to a copy pattern.  modification can generate hidden
dependencies which are
hard to debug and untangle, making the code harder to scale and maintain.

-brian


On 5/28/06, Jonathan Taylor <jonathan.taylor at stanford.edu> wrote:
>
> I wanted to change a read-only attribute (the slice attribute of
> SliceIteratorNext) to
> not read-only but "untyped". I changed "readonly" to "attribute" and got
> the following message:
>
> --------------
> Traceback (most recent call last):
> File "run.py", line 173, in ?
>    FIACrun(subj=subj, run=run)
> File "run.py", line 134, in FIACrun
>    OLS.fit()
> File
>
> "/home/jtaylo/python/lib/python2.4/site-packages/neuroimaging/fmri/fmristat/__init__.py",
> line 107, in fit
>    LinearModelIterator.fit(self, **keywords)
> File
>
> "/home/jtaylo/python/lib/python2.4/site-packages/neuroimaging/statistics/regression.py",
> line 48, in fit
>    output.next(data=out)
> File
>
> "/home/jtaylo/python/lib/python2.4/site-packages/neuroimaging/fmri/regression.py",
> line 40, in next
>    value.slice = value.slice[1]
> File "/home/jtaylo/python/lib/python2.4/site-packages/attributes.py",
> line 256, in set
>    self.validate(value)
> File "/home/jtaylo/python/lib/python2.4/site-packages/attributes.py",
> line 211, in validate
>    if not self.isvalid(value): raise ProtocolOmission(
> protocols.ProtocolOmission: attribute slice implements ([slice(0, 191,
> 1), slice(0, 1, 1)],), value slice(0, 1, 1) of type <type 'slice'> does
> not implement: ('__getslice__', 'pop', 'remove', '__rmul__', '__lt__',
> 'append', 'index', '__delslice__', '__getitem__', '__contains__',
> '__len__', 'sort', '__ne__', 'extend', 'insert', '__iter__', '__add__',
> '__gt__', '__eq__', 'reverse', 'count', '__delitem__', '__reversed__',
> '__imul__', '__setslice__', '__setitem__', '__iadd__', '__le__',
> '__mul__', '__ge__')
>
> ---------------------------
>
> I am not sure how to get this to work. I want it to behave like a slice or
> a list of slices.
>
> There is a very good example when you want to be able to change its value
> later: when the "iterator" driving the voxelwise analysis comes from an fMRI
> image, it spits out
> slices of time series, but if you want to output slices of floats (i.e. a
> 3d image of means), you want to ignore the first part of this (python) slice
> but you can still use the same "SliceIteratorNext" instance to drive the
> next method of the images....
>
> Thanks,
>
> Jonathan
>
> --
> ------------------------------------------------------------------------
> I'm part of the Team in Training: please support our efforts for the
> Leukemia and Lymphoma Society!
>
> http://www.active.com/donate/tntsvmb/tntsvmbJTaylor
>
> GO TEAM !!!
>
> ------------------------------------------------------------------------
> Jonathan Taylor                           Tel:   650.723.9230
> Dept. of Statistics                       Fax:   650.725.8977
> Sequoia Hall, 137                         www-stat.stanford.edu/~jtaylo
> 390 Serra Mall
> Stanford, CA 94305
>
> _______________________________________________
> Nipy-devel mailing list
> Nipy-devel at neuroimaging.scipy.org
> http://neuroimaging.scipy.org/mailman/listinfo/nipy-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://projects.scipy.org/pipermail/nipy-devel/attachments/20060530/64771371/attachment-0002.html