[Rcpp-devel] Help with an Rcpp and CppBugs example

[Shige Song]

Very nice, thanks.

Shige

On Sat, Oct 1, 2011 at 5:16 PM, Whit Armstrong <armstrong.whit at gmail.com> wrote:
> Please have a look here.
>
> It's a simple linear model using inline and cppbugs:
> https://github.com/armstrtw/CppBugs/tree/master/test/r.inline.example
>
> -Whit
>
>
> On Sat, Oct 1, 2011 at 3:29 PM, Dirk Eddelbuettel <edd at debian.org> wrote:
>>
>> On 1 October 2011 at 15:03, Shige Song wrote:
>> | Any examples showing how CppBugs and Rcpp work together will be good.
>> | I am particularly interested in knowing how GLM models and GLMM models
>> | can be estimated that way.
>> |
>> | Thanks in advance.
>>
>> Well, maybe you should really try these two things:
>>
>>  i)  take the five or six small examples in Whit's git repo, and read up on
>>      the inline package and its cxxfunction() to wrap them -- given the
>>      /working example/ I just provided you yesterday, this is not all that hard.
>>
>>  ii) take glm / glmm code you and try to make it work with CppBugs in
>>      standalone mode as per Whit's examples in hit repo.  Then revisit i)
>>      and wrap it too.
>>
>> Dirk
>>
>> | Best,
>> | Shige
>> |
>> | On Sat, Oct 1, 2011 at 1:52 PM, Whit Armstrong <armstrong.whit at gmail.com> wrote:
>> | > I'm happy to provide more examples of cppbugs with inline and Rcpp.
>> | >
>> | > Is there something in particular you had in mind?
>> | >
>> | > -Whit
>> | >
>> | >
>> | > On Sat, Oct 1, 2011 at 7:05 AM, Shige Song <shigesong at gmail.com> wrote:
>> | >> Dear Whit,
>> | >>
>> | >> I have been playing with other examples you provided in the github
>> | >> repository. The one Dirt sent, however, is the only example that I can
>> | >> find from the internet showing how CppBugs works with Rcpp (and R). As
>> | >> I see it, such a combination has great potential providing a flexible
>> | >> yet powerful Bayesian computational tool.
>> | >>
>> | >> Very nice work, and thanks for the suggestion.
>> | >>
>> | >> Best,
>> | >> Shige
>> | >>
>> | >> On Fri, Sep 30, 2011 at 10:06 PM, Whit Armstrong
>> | >> <armstrong.whit at gmail.com> wrote:
>> | >>> Shige,
>> | >>>
>> | >>> That example is quite dated at this point.  The CppBugs api has
>> | >>> changed a lot since then and is likely to change more in the near
>> | >>> future.
>> | >>>
>> | >>> Please git pull the latest from github, and ping me if you have any issues.
>> | >>>
>> | >>> There are also quite a few pure c++ examples the the 'test' dir to get
>> | >>> you started.
>> | >>>
>> | >>> In the next major release of CppBugs you will be able to declare the
>> | >>> objects directly in R, but give me a few months to get that working.
>> | >>>
>> | >>> -Whit
>> | >>>
>> | >>>
>> | >>> On Fri, Sep 30, 2011 at 9:40 PM, Shige Song <shigesong at gmail.com> wrote:
>> | >>>> Dear Dirk,
>> | >>>>
>> | >>>> Thank you very much for the suggestions and the upated file. Your file
>> | >>>> actually works flawlessly on my system. It looks really interesting
>> | >>>> and educational.
>> | >>>>
>> | >>>> Thanks also for the great work on Rcpp, really amazing piece of
>> | >>>> software you got there.
>> | >>>>
>> | >>>> Best,
>> | >>>> Shige
>> | >>>>
>> | >>>> On Fri, Sep 30, 2011 at 9:11 PM, Dirk Eddelbuettel <edd at debian.org> wrote:
>> | >>>>>
>> | >>>>> Shige,
>> | >>>>>
>> | >>>>> There is no way to sugarcoat this: you have to learn to live with, and learn
>> | >>>>> from, the compiler errors and relate them to the actual code. Using Rcpp
>> | >>>>> still means programming in the context of a C++ compiler.
>> | >>>>>
>> | >>>>>
>> | >>>>> You also need Whit's CppBugs repo from github _installed somewhere_ so that
>> | >>>>>
>> | >>>>>  #include <cppbugs/cppbugs.hpp>
>> | >>>>>
>> | >>>>> works. Plus the same for Conrad's Armadillo as we have
>> | >>>>>
>> | >>>>>  #include <armadillo>
>> | >>>>>
>> | >>>>> And to top it all off, you probably need a bunch of Boost installed as
>> | >>>>> CppBugs uses it.  If all that is a given, then you can run the attached file
>> | >>>>> 'whit.r' as I do below. This file served as in example in the Rcpp workshop
>> | >>>>> in April and I just fetched it from my sources. The version posted then is
>> | >>>>> likely a little outdated.  But this one works:
>> | >>>>>
>> | >>>>> $ r whit.R
>> | >>>>> Loading required package: methods
>> | >>>>>   user  system elapsed
>> | >>>>>  0.220   0.020   0.236
>> | >>>>> $b
>> | >>>>> [1] -0.3303790  0.5276294
>> | >>>>>
>> | >>>>> $ar
>> | >>>>> [1] 0
>> | >>>>>
>> | >>>>> $
>> | >>>>>
>> | >>>>> Whether you use Rscript or r (from littler) does not matter.  The updated
>> | >>>>> whit.r is attached.  It builds and runs, I have no idea if it makes any
>> | >>>>> sense... I think it regresses y ~ X with both being noise so there.
>> | >>>>>
>> | >>>>> Hope this helps,  Dirk
>> | >>>>>
>> | >>>>>
>> | >>>>>
>> | >>>>>
>> | >>>>> --
>> | >>>>> New Rcpp master class for R and C++ integration is scheduled for
>> | >>>>> San Francisco (Oct 8), more details / reg.info available at
>> | >>>>> http://www.revolutionanalytics.com/products/training/public/rcpp-master-class.php
>> | >>>>>
>> | >>>>>
>> | >>>> _______________________________________________
>> | >>>> Rcpp-devel mailing list
>> | >>>> Rcpp-devel at lists.r-forge.r-project.org
>> | >>>> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
>> | >>>
>> | >>
>> | >
>> | _______________________________________________
>> | Rcpp-devel mailing list
>> | Rcpp-devel at lists.r-forge.r-project.org
>> | https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
>> --
>> New Rcpp master class for R and C++ integration is scheduled for
>> San Francisco (Oct 8), more details / reg.info available at
>> http://www.revolutionanalytics.com/products/training/public/rcpp-master-class.php
>>
>