Open anthonysena opened 2 years ago
As I'll discuss during a HADES meeting (hopefully the next one), I'd like all HADES packages to follow the same pattern:
The data model of the CSV files should be documented like this. Ideally, all Shiny apps should be able to work both with CSV files (possibly pre-processed for efficiency), as well as a (Postgres) database backend for when data get big. Also, and I'm not yet sure how to do this, but the Shiny app would really be a module, meaning it could be integrated into a bigger Shiny app, for example a study that uses the CohortGenerator package, but also has other results to show.
Also, and I'm not yet sure how to do this, but the Shiny app would really be a module, meaning it could be integrated into a bigger Shiny app, for example a study that uses the CohortGenerator package, but also has other results to show.
Perhaps adding the module(s) required for displaying cohort stats should live in: https://github.com/OHDSI/OhdsiShinyModules and then the CohortGenerator package would list this in the Suggests
section of the DESCRIPTION file?
I was thinking the analytic code, data model for the analytics results, and Shiny module for displaying those results would all live in the same package. But I can see why having the different modules in the same package can help improve integration of the modules. I'm honestly not sure what the right approach would be.
Why not use the module version of cohort page in cohort diagnostics diagnostic explorer?
It gives everything we need, including human readable text and counts
It would be useful to have a small shiny app for reviewing the cohort generation results