Instead of having a huge file with all the shiny app fully separated from the package as we have now, Bioconductor proposes a package architecture (the GUI is under the R/ folder), with some conventions with respect to the observables, the user interface the server code, etc.
I believe that:
Those Bioconductor conventions can help having a GUI that is easier to maintain than what we currently have, I'm happy to have those directives to follow.
This new requirement on the shiny architecture will prevent us from submitting the GUI in its current state to Bioconductor.
Having the GUI in the GCIMS package would require that we have many more Imports/Suggests. Instead of increasing even more that number, I would suggest splitting the GCIMS GUI into a separate package (e.g GCIMSgui)
What do you think @Luis-Fernandez-Romero @celiamallafre?
On Sept. 16 (so basically after Anita finished her work), a new chapter on shiny apps was added to the Bioconductor package guidelines book: https://contributions.bioconductor.org/shiny.html?#shiny
Instead of having a huge file with all the shiny app fully separated from the package as we have now, Bioconductor proposes a package architecture (the GUI is under the
R/folder), with some conventions with respect to the observables, the user interface the server code, etc.I believe that:
GCIMSgui)What do you think @Luis-Fernandez-Romero @celiamallafre?
If you agree I'll move it