Extending OmopViewer

library(OmopViewer)

Introduction

Do yo want OmopViewer to support more packages? This is a developer focused vignette explaining the process to add new default panels to the package.

1 What lives in panels.R

panels.R builds one list – omopViewerPanels – that the package ships as data. Every element of that list is a panel definition. If you’ve read the other vignette - Panels defined in OmopViewer - you’ve already seen this structure in action. Here’s a quick recap:

key	purpose
title	label shown in the sidebar
icon	any free FontAwesome icon
data$result_type	string returned by your ETL / analyser
automatic_filters	vector of column groups auto-exposed
filters	named list of picker / switch definitions
content	tabs such as tidy, table, plot, …

str(omopViewerPanels$summarise_omop_snapshot, max.level = 2)
#> List of 6
#>  $ title            : chr "Snapshot"
#>  $ icon             : chr "clipboard-list"
#>  $ data             :List of 1
#>   ..$ result_type: chr "summarise_omop_snapshot"
#>  $ automatic_filters: chr "variable_name"
#>  $ filters          :List of 1
#>   ..$ cdm_name:List of 7
#>  $ content          :List of 2
#>   ..$ tidy :List of 6
#>   ..$ table:List of 5
#>  - attr(*, "class")= chr "omopviewer_panel"

2 Add a new panel

Edit panels.R above the “all panels” block. Here I use SamplePkg::tablefunction and SamplePkg::plotfunction as a placeholder for the plot and table functions developer should add.

## my custom summary ----
myCustomPanel <- list(
  # --- metadata -----------------------------------------------------------
  title  = "My Custom",
  icon   = "wand-magic-sparkles",
  data   = list(result_type = "summarise_my_custom"),
  automatic_filters = c("group", "strata", "variable_name", "settings"),
  filters = list(cdm_name = cdmFilter),
  
  # --- UI tabs ------------------------------------------------------------
  content = list(
    tidy = tidyContent,
    
    table = list(
      title       = "Table Sample",
      output_type = "gt",
      reactive    = "<filtered_data> |> SamplePkg::tablefunction(",
      render      = "<reactive_data>",
      filters = rankTableButton(
        none        = c("<strata>", "estimate_name"),
        header      = "variable_name",
        groupColumn = "cdm_name",
        hide        = "variable_level"
      ),
      download = downloadGtTable("table_sample")
    ),
    
    plot = list(
      title       = "Plot Sample",
      output_type = "ui",
      reactive    = "<filtered_data> |> SamplePkg::plotfunction(facet = input$facet)",
      render      = "renderInteractivePlot(<reactive_data>, input$interactive)",
      filters = list(
        interactive = list(
          button_type = "materialSwitch",
          label = "Interactive",
          value = TRUE
        ),
        facet = list(
          button_type = "pickerInput",
          label = "Facet",
          choices = c(
            'cdm_name', '<group>', '<strata>', '<additional>', '<settings>'
          ),
          selected = 'cdm_name',
          multiple = TRUE
        )
      ),
      download = downloadPlot("plot_sample.png")
    )
  )
)

Key points for developers

• Always add both table and plot nodes (unless your package really has no plot).
• Use the standard helper objects: tidyContent, rankTableButton(), downloadGtTable(), downloadPlot().
• Place the new list above the omopViewerPanels <- list(...) block, then register it:

omopViewerPanels <- list(
  # existing entries …
  summarise_my_custom = myCustomPanel,
  # default
  default = defaultPanel
) |> purrr::map(\(x) newOmopViewerPanel(x))

Run devtools::load_all(); the static viewer will now include My Custom.

---

Commit, tag, release. Done.