Data Pipeline, Parameter Reactivity & Isolated Testing

Runtime Data Initialization (`foundrSetup`)

Before launching server modules or standalone test apps, runtime dataset objects are initialized via foundrSetup():

foundrSetup(
  data_instance = "Liver",
  data_subset   = c("Physio", "MixMod"),
  custom_settings = TRUE,
  dirpath       = "~/founder_diet_study/HarmonizedData"
)

Global Runtime Objects

foundrSetup() generates five core global objects required by module servers:

Object	Data Structure	Purpose
`traitData`	`tibble` / `data.frame`	Individual observation measurements
`traitSignal`	`tibble` / `data.frame`	Cell means and normalized trait metrics
`traitStats`	`list` / `tibble`	ANOVA estimates and statistical contrast results
`traitModule`	`list`	WGCNA module assignments
`customSettings`	`list`	App deployment metadata, dataset lists, and help file paths

Three-Tier Parameter Reactivity

Parameters are managed as standard Shiny reactiveValues across three distinct tiers:

# Server initialization pattern
main_par  <- mainParServer("main_par", traitSignal)
panel_par <- panelParServer("panel_par", main_par, traitSignal)
plot_par  <- plotParServer("plot_par", contrast_table)

flowchart TD
    main_par["main_par (Global Parameters: dataset, trait order, plot or table toggle, height)"]
    panel_par["panel_par (Panel Parameters: strains, genotypes, sex B/F/M/C, faceting)"]
    plot_par["plot_par (Plot Parameters: volcano thresholds volsd and volvert, interact toggle, row names)"]

    main_par --> panel_par
    panel_par --> plot_par

    classDef param fill:#2ca02c,stroke:#333,stroke-width:2px,color:#fff
    class main_par,panel_par,plot_par param

Parameter Persistence & Sub-Panel Reactivity

Managing input persistence across Shiny modules requires careful scoping using reactiveVal() and namespaced inputs:

Top-Level Parameter Flow: main_par$dataset is passed from mainParServer to each panel module.
Panel Scoping: Local selections (e.g., keydataset inside traitOrderApp) use reactiveVal() to maintain state while navigating sub-panels within the same tab.
Sub-Panel Parameter Sharing:
- In contrastServer, selection changes in the Sex sub-panel persist when switching to the Group/Module sub-panel (and vice-versa).
- In traitServer, selecting a key trait updates choices available for related traits via traitNamesApp.
Cross-Tab Behavior: Switching top-level tab panels re-evaluates scoped panel parameters against global defaults unless explicitly wired to main_par.

Unit Testing with Standalone Module Apps

Each file in R/*.R exports a standalone *App() function that acts as an isolated test harness. This allows developers to test individual UI components or plots without running the full multi-tab app.

library(foundrShiny)

# 1. Initialize data
foundrSetup(data_instance = "Liver", data_subset = "Physio")

# 2. Test correlation table module in isolation
corTableApp(traitSignal)

# 3. Test volcano plot module in isolation
volcanoApp(traitStats)

# 4. Test full trait panel in isolation
traitApp(traitData, traitSignal, customSettings)

Data Pipeline, Parameter Reactivity & Isolated Testing