William Stein wrote:
> Hi Andrew (cc: sage-devel),
>
> I just tried out rpy2 (a superfast C-level Python<-->R interface) and
> the user interface is amazingly ridiculously painfully to actually
> interactively use. Evidently rpy was removed from sage and replaced
> by rpy2 sometime this summer. Anyway, with rpy we had:
>
> sage: import rpy; rpy.sd([1..100]) # no longer works in sage
>
> to compute the standard deviation of a list. With rpy2 the same thing
> becomes:
>
> sage: import rpy2.robjects # using sage-4.2.1
> sage: rpy2.robjects.r['sd'](rpy2.robjects.IntVector([1..100])).r_repr()
> '29.011491975882'
>
> Just look at that carefully. Ouch.
>
> Anyway, any thoughts? I wonder if it would be easy to provide a
> lightweight wrapper around rpy2 that would
> provide a much more sensible interface. At least the docs at
> http://rpy.sourceforge.net/rpy2/doc/html/index.html are pretty good.
>
Like this?
import rpy2.rpy_classic as rpy
r=rpy.r
rpy.set_default_mode(rpy.BASIC_CONVERSION)
r.sd([int(i) for i in [1..100]])
[29.011491975882016]
(the only problem here is that the internal function doesn't know about
Sage types, which is very easily fixed...)
Or do you mean a wrapper like this?
import rpy2
import rpy2.robjects as ro
import numpy
import rpy2.robjects.numpy2ri as numpy2ri
def my_py2ri(o):
if isinstance(o,(list,tuple)):
o=ro.IntVector(o)
else:
o=ro.default_py2ri(o)
return o
ro.conversion.py2ri=my_py2ri
def my_ri2py(obj):
res = ro.default_ri2py(obj)
if isinstance(res, ro.RVector):
if len(res) == 1:
res = res[0]
else:
res = [res[i] for i in range(len(res))]
return res
ro.conversion.ri2py = my_ri2py
r=rpy2.robjects.r
r.sd([1..100])
29.011491975882016
-Jason
--
Jason Grout
--
To post to this group, send an email to [email protected]
To unsubscribe from this group, send an email to
[email protected]
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
