Hi Jeremy,
Thanks for your quick response.
I'll try switching from using 'tools/plotting/r_wrapper.sh' from
Galaxy distribution, to my own wrapper using python, so I can use your
advice.
I still wonder if this is the best way to run R's scripts? I would
have thought there would be a simpler option. I wonder if
"r_wrapper.sh" could be improved to absorb stderr messages, I would
guess Bash has these capabilities.
Thanks,
Carlos
On Mon, Jan 30, 2012 at 1:41 PM, Jeremy Goecks <jeremy.goecks(a)emory.edu> wrote:
Carlos,
You'll want to have your wrapper collect output to stderr and report only if the tool
fails. See the Cufflinks/compare/diff wrappers for examples of how this can be done.
Good luck,
J.
On Jan 30, 2012, at 1:01 PM, Carlos Borroto wrote:
> Hi all,
>
> I'm working on integrating cummeRbund[1] into Galaxy. This is a very
> helpful tool for analysis and visualization of cuffdiff output.
>
> My problem is any text output from running R code, puts the galaxy
> output in error state. Even if everything is working correctly and the
> outputs are generated as intended. I tried this code from
> 'tools/plotting/xy_plot.xml':
> ## Setup R error handling to go to stderr
> options( show.error.messages=F,
> error = function () { cat( geterrmessage(), file=stderr() ); q(
> "no", 1, F ) } )
>
> But it didn't help. Even for loading the library I had to go with:
> suppressPackageStartupMessages(library("cummeRbund"))
>
> Nothing else would work. Now the tool is outputting some messages
> while running R functions, which is putting the output in galaxy in
> error state. I'm thinking about asking the upstream developer to
> provide a way to run the functions in the library in 'quite' mode. Is
> there any other option?
>
> Right now I'm using the path of creating a R script in the XML file
> and using r_wrapper.sh to run it. I tried first with rpy, as it would
> give me more freedom, but cummeRbund use ggplot2 which seems to be
> supported only on rpy2.
>
> [
1]http://compbio.mit.edu/cummeRbund/
>
> Thanks in advance,
> Carlos
> ___________________________________________________________
> Please keep all replies on the list by using "reply all"
> in your mail client. To manage your subscriptions to this
> and other Galaxy lists, please use the interface at:
>
>
http://lists.bx.psu.edu/