Hi Ben,
I see what you're doing now. There are a couple things I should mention.
First of all, you are right: getattr(r,'generic.scatter.plot') is not the
same as typing r.generic.scatter.plot -- instead, it invokes the
__getattr__ method of r, and looks for something by that name in the R
namespace. Having found nothing, it raises an exception. If it did find
something -- you'd have to rplot.source_for_function() or
r.source("/path/to/generic.scatter.plot.R") first -- it would return the
function, and assign it to r.generic_scatter_plot (rpy automatically
translates R dots . to python underscores _ in variable names).
It also looks like our package versions are significantly different. RPy
is notoriously touchy about what particular package versions you are
using, so that may be the cause of django.contrib.dataplot's failure on
your system. But despite the differences in package versions, I bet that
if you source the R code first, you will be able to execute the getattr
properly, i.e.
>>> from rpy import r
>>> r.source("/usr/local/lib/python2.4/site-packages/django/contrib/dataplot/R/generic.scatter.plot.R")
{'visible': False, 'value': <Robj object at 0xb7e11190>}
>>> r.generic_scatter_plot
<Robj object at 0xb7e111a0>
However, the whole point of django.contrib.dataplot is that it takes care
of these details for you ---
>>> from django.contrib import dataplot
>>> plot=dataplot.Scatter('myscatterplot',fun_to_get_scatter_data)
>>> plot.to_html()
Have you tried to get the demo bike app (django.contrib.dataplot.bike) to
work? If you are able to get those simple examples working, it should be
straightforward to translate the design paradigm to your problems.
Sincerely,
Toby Dylan Hocking
http://www.ocf.berkeley.edu/~tdhock
On Thu, 12 Jul 2007, Ben Ford wrote:
> Hi Toby,
> My versions are:
> R: 2.4.0
> Python: 2.5.1
> Rpy: 1.0-RC2
>
> Sorry I wasn't very clear earlier... I imported r into ipython to have a
> play with it and see what happens. I tried getattr(r,' generic.scatter.plot')
> which didn't work - I'm under the impression that getattr won't work with
> dotted notation like that if you're doing normal object lookup, am I right
> in saying that r does it differently to python? So i tried getattr(r,
> 'generic') and that didn't work either. The error I get is:
>> Error in get(x, envir, mode, inherits) : variable "generic.time.series"
> was not found
> I tried it with underscores too and I got exactly the same message...
>
> In fact just having looked through the source and having a play with rpy it
> seems that there are bugs all over the place:
>
>>>> from rpy import r
> RHOME= /usr/lib/R
> RVERSION= 2.4.0
> RVER= 2040
> RUSER= /home/ben
> Loading Rpy version 2040 .. Done.
> Creating the R object 'r' .. Done
>>>> r
> Traceback (most recent call last):
> File "<stdin>", line 1, in <module>
> File "/usr/lib/python2.5/site-packages/rpy.py", line 307, in __repr__
> Rver = self.__getitem__('R_version_string')
> File "/usr/lib/python2.5/site-packages/rpy.py", line 290, in __getitem__
> obj = self.__dict__[name] = self.__dict__.get(name, self.get(name))
> rpy.RException: Error in get(x, envir, mode, inherits) : variable
> "R_version_string" was not found
>
> It looks like the logic in R.__getattr__ isn't working properly! I'll
> reinstall it at some point and see how it works :-)
> Cheers
> Ben
>
> On 12/07/07, Toby Dylan Hocking <[EMAIL PROTECTED]> wrote:
>>
>>
>> Hi Ben,
>>
>> I'm excited to hear that you got a copy of django.contrib.dataplot and you
>> are trying it out. I will try to help you debug the problem, but the
>> full traceback may be more useful. Can you send it?
>>
>> The purpose of the get_r_fun method is to look at the current R
>> environment and check if the desired R function has been source'ed yet.
>> This is one of the first steps before passing the data before R.
>>
>> I take the "r has no attribute generic" to refer to one of the generic
>> plotting functions named in one of the RPlot subclasses -- you must be
>> dealing with Scatter, TimeSeries, or NormalQQPlot, right? Are you sure it
>> doesn't say something like "r has no attribute generic.scatter.plot"? You
>> might try changing the dots . to underscores _ in the
>> r_fun_name='generic.scatter.plot' line in the definition of the
>> ScatterPlot subclass.
>>
>> Maybe your version of RPy isn't translating python names into R names the
>> same way as mine is? I'm using RPy 0.4.6-3ubuntu2 Python 2.4.2-0ubuntu3
>> and R 2.2.1-2 on ubuntu dapper. What are your versions?
>>
>> Sincerely,
>> Toby Dylan Hocking
>> http://www.ocf.berkeley.edu/~tdhock<http://www.ocf.berkeley.edu/%7Etdhock>
>>
>> On Wed, 11 Jul 2007, Ben Ford wrote:
>>
>>> Hi Toby
>>> I've grabbed a copy of your code but the RPlot.get_r_fun method isn't
>>> working for me... Could you perhaps explain how it does the mapping to
>> the R
>>> function? What my testing is saying to me is that r (the one that's
>> imported
>>> at the top of __init__.py) has no attribute generic... I've had a very
>> brief
>>> look at rpy before, but I'm not exactly familiar with it's inner
>>> workings....
>>> Great work, I've been meaning to have a crack at something like this for
>>> ages!!
>>> Ben
>>>
>>> On 11/07/07, Toby Dylan Hocking <[EMAIL PROTECTED]> wrote:
>>>>
>>>>
>>>> Hi there,
>>>>
>>>> If you still need help with making data-driven plots, why don't you
>> check
>>>> out my new Django package --- I'd like some testers and comments.
>>>> Basically it is an interface to the R programming language (for
>> statistics
>>>> and graphics) through the RPy package. You can download my latest
>> release:
>>>>
>>>> http://www.ocf.berkeley.edu/~tdhock/dataplot-0.2.tgz<http://www.ocf.berkeley.edu/%7Etdhock/dataplot-0.2.tgz>
>>>>
>>>> I'm a statistician who has been using this system for some time at my
>>>> work, but I finally am getting around to generalizing it and packaging
>> it
>>>> for general use with the Django framework.
>>>>
>>>> The installation instructions are in INSTALL.txt in the archive.
>>>> Documentation is mostly present in docstrings at the moment -- I'm
>> working
>>>> on more tutorials, but the .txt files and the example app should be
>> enough
>>>> to get you started.
>>>>
>>>> Sincerely,
>>>> Toby Dylan Hocking
>>>> http://www.ocf.berkeley.edu/~tdhock<http://www.ocf.berkeley.edu/%7Etdhock>
>>>>
>>>>
>>>>>
>>>>
>>>
>>>
>>> --
>>> Regards,
>>> Ben Ford
>>> [EMAIL PROTECTED]
>>> +628111880346
>>>
>>>>
>>>
>>
>>>
>>
>
>
> --
> Regards,
> Ben Ford
> [EMAIL PROTECTED]
> +628111880346
>
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Django users" group.
To post to this group, send email to django-users@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/django-users?hl=en
-~----------~----~----~----~------~----~------~--~---
