[pymvpa] feature id --> feature coordinate

[Matthias Ekman]

> indeed we seems to be missing explicit instructions on that, so look
> into
>
> http://www.pymvpa.org/modref/mvpa.mappers.base.html?highlight=getinid#mvpa.mappers.base.Mapper.getInId
>
> so, for a mapped dataset ds, try:
>
> ds.mapper.getInId(10)
>
> e.g:
>
> n [4]: print dataset
> ------>  print(dataset)
> Dataset / int16 1452 x 530 uniq: 12 chunks 9 labels labels_mapped
>
> In [5]: print dataset.mapper.getInId(1)
> ------>  print(dataset.mapper.getInId(1))
> [ 0  1 17]

Thank you very much!
Now I remember that this is used in a similar way in the 
eventrelated_example.py... so it seems to be documented (but users are 
lazy) :-)

In [10]: coor=dataset.mapper.getInId(1)

In [11]: coor
Out[11]: array([ 0,  1, 17])

In [12]: coor_reverse=np.array([coor[2],coor[1],coor[0]])

In [13]: coor_reverse
Out[13]: array([17,  1,  0])

>
> In 0.5 the world becomes simpler -- arbitrary feature attributes allow to embed
> original coordinates from the beginning:
>
> http://dev.pymvpa.org/tutorial_datasets.html#loading-fmri-data
> look for voxel_ids

Cool - this is a great improvement.

Best,
  Matthias

-- 
_____________________________________________________

Matthias Ekman, PhD Student
Radboud University | Donders Center for Cognition
Montessorilaan 3
6525 HR Nijmegen, Netherlands

E-mail: m.ekman at donders.ru.nl
Phone: +31-(0)24 365 5934
_____________________________________________________