[Bioc-devel] Guidelines for shiny packages

[Thomas Dybdal Pedersen]

Unfortunately this approach is limited to apps where the UI of the app is defined in R code. For more involved GUI development it is much preferred to write the HTML yourself and for this to work, there need to be a server.R file. I don’t think we should have recommendations that limits the developer to such an extend…

best

Thomas


On 29 Sep 2014, at 21:59, Laurent Gatto <lg390 at cam.ac.uk> wrote:

> 
> On 29 September 2014 20:49, Thomas Dybdal Pedersen wrote:
> 
> Dear Thomas, 
> 
>> Hi All
>> 
>> I’m in the process of getting package submitted to Bioconductor. The
>> package is a pure shiny package and this has led to some - especially
>> one question. For those not familiar with shiny apps, the code that
>> makes these work resides in their own .R files outside of the R
>> directory in a package. Consequently they are not ‘build’ and whatever
>> is in there is not part of the package namespace. The functionality of
>> my GUI requires several packages, but for the reasons given above they
>> are not imported into the package namespace. Still they are needed for
>> the package to function properly. The main question is thus: How to
>> state this hard dependency in the DESCRIPTION file without getting
>> knocked on the head for not importing your dependencies during check
>> and BiocCheck?
> 
> When developing a package, I do not use the server.R and ui.R files, but
> define the shiny app in R/code.R as follow.
> 
>  myapp <- function(object) { 
>    app <- list(
>      ui = bootstrapPage( ... ),
>      server = function(input, output) { ... }
>    )
>  }
> 
> and dependencies are stated in DESCRIPTION and NAMESPACE.
> 
> Hope this helps.
> 
> Laurent
> 
>> Another related question/proposal is: Based on the interest in Shiny
>> at BioC 2014 it seems as we should anticipate a lot of shiny based
>> packages. As one of the core principles of the Bioconductor project is
>> to promote good coding style and high quality packages I was wondering
>> whether it would make sense to begin working on a set of guidelines
>> for developing shiny based GUIs, as their structure and style differs
>> a lot from standard R packages. Just a suggestion - I’ll happily join
>> such an effort…
>> 
>> best
>> 
>> Thomas
>> _______________________________________________
>> Bioc-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
> 
> 
> -- 
> Laurent Gatto
> http://cpu.sysbiol.cam.ac.uk/