Just to update everyone that I've found the bug in the package, and now the
RcppDend package works fine. You can see the relevant commit here:
https://github.com/talgalili/RcppDend/commit/25d456571f2f1e52f1f9aa8ac6bc3a7fcb832e65
The problem was that my NAMESPACE file was missing the two lines:
useDynLib(RcppDend)
exportPattern("^[[:alpha:]]+")
The reason it was missing them was because I am using Rstudio+roxygen2, and
had NAMESPACE editing turned on, and it appears to erase these lines. This
had caused R to not find the Rcpp functions, and resulted in the error:
Error in .Call("RcppDend_Rcpp_labels_dendrogram", PACKAGE = "RcppDend") :
"RcppDend_Rcpp_labels_dendrogram
" not available for .Call() for package "RcppDend"
I have submitted a bug report for RStudio/roxygen2:
http://support.rstudio.org/help/discussions/problems/8418-rcpp-packages-fails-when-using-roxygen2-to-handle-namespace
https://github.com/klutometis/roxygen/issues/130
Future tweaks/commits/suggestions to RcppDend
<https://github.com/talgalili/RcppDend>are most welcomed.
Best,
Tal
----------------Contact
Details:-------------------------------------------------------
Contact me: tal.gal...@gmail.com |
Read me: www.talgalili.com (Hebrew) | www.biostatistics.co.il (Hebrew) |
www.r-statistics.com (English)
----------------------------------------------------------------------------------------------
On Fri, Aug 16, 2013 at 9:48 PM, Tal Galili <tal.gal...@gmail.com> wrote:
> Dear Dirk, Romain, Kevin, and others.
>
> Since I am currently actively working on extending the dendrogram object
> in R, I have a real interest in finding and solving some of the major
> bottlenecks of dendrogram manipulation.
> Thank you all for all of your help and suggestions thus far.
>
> I took the liberty of starting a github repo for a new R package
> "RcppDend", you can see it here:
> https://github.com/talgalili/RcppDend
>
> It currently contains the function for label extraction from dendrogram (a
> mixture of the code Romain and Dirk wrote, with some minor modifications by
> me).
> The Rcpp code is here:
> https://github.com/talgalili/RcppDend/blob/master/src/labels_dendrogram.cpp
> And the R code (wrapping it a bit, and adding documentation), is here:
> https://github.com/talgalili/RcppDend/blob/master/R/RcppDend.r
>
> Currently, I can get the labels.dendrogram function to work when I'm
> compiling it in a local session, but *the function fails to work when
> loaded from this package*. If you have suggestions on how to fix it, I'd
> be happy to read.
>
> And in general, I welcome any code contributions/commits.
>
> Best,
> Tal
>
>
>
>
>
>
> ----------------Contact
> Details:-------------------------------------------------------
> Contact me: tal.gal...@gmail.com |
> Read me: www.talgalili.com (Hebrew) | www.biostatistics.co.il (Hebrew) |
> www.r-statistics.com (English)
>
> ----------------------------------------------------------------------------------------------
>
>
>
> On Fri, Aug 16, 2013 at 9:04 PM, Kevin Ushey <kevinus...@gmail.com> wrote:
>
>> Just a suggestion re: recursion with Rcpp, yet another nice way we can
>> recurse is through using the sugar lapply:
>>
>> -----
>>
>> #include <Rcpp.h>
>> using namespace Rcpp;
>>
>> // [[Rcpp::export]]
>> RObject add_attr(RObject x_) {
>> RObject x = clone(x_);
>> if (is<List>(x)) {
>> x = wrap( lapply( as<List>(x), add_attr ) );
>> } else {
>> x.attr("type") = "Other";
>> }
>> return x;
>> }
>>
>> /*** R
>> add_attr( list(x=list(a=1, b=2, c=3), y=list( list( d=4 ) ) ) )
>> */
>>
>> -----
>>
>> If we want to operate on specific elements as well, we can use is<T> and
>> as<T> together, and then operate on objects as needed.
>>
>> It lets us avoid an explicit for loop, and 'feels' more like R.
>>
>> -Kevin
>>
>
>
_______________________________________________
Rcpp-devel mailing list
Rcpp-devel@lists.r-forge.r-project.org
https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
