Back to Skills

build-shiny-module

pjt222
Updated 2 days ago
4 views
17
2
17
View on GitHub
Metareactaitestingdesign

About

This skill helps developers build reusable Shiny modules with proper namespace isolation using NS(). It covers creating UI/server pairs, managing reactive returns, and enabling inter-module communication and nesting. Use it to extract reusable components, encapsulate complex logic, or compose larger applications from testable units.

Quick Install

Claude Code

Recommended
Primary
npx skills add pjt222/agent-almanac -a claude-code
Plugin CommandAlternative
/plugin add https://github.com/pjt222/agent-almanac
Git CloneAlternative
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/build-shiny-module

Copy and paste this command in Claude Code to install this skill

Documentation

Build Shiny Module

Create reusable Shiny UI/server module pairs with proper namespace isolation, reactive communication, and composability.

When to Use

  • Extracting a reusable component from a growing Shiny app
  • Building a UI widget that will be used in multiple places
  • Encapsulating complex reactive logic behind a clean interface
  • Composing larger applications from smaller, testable units

Inputs

  • Required: Module purpose and functionality description
  • Required: Input/output contract (what the module receives and returns)
  • Optional: Whether the module nests other modules (default: no)
  • Optional: Framework context (golem, rhino, or vanilla)

Procedure

Step 1: Define the Module Interface

Before writing code, define what the module accepts and returns:

Module: data_filter
Inputs: reactive dataset, column names to filter on
Outputs: reactive filtered dataset
UI: filter controls (selectInput, sliderInput, dateRangeInput)

Got: Clear contract specifying reactive inputs, reactive outputs, and UI elements.

If fail: If the interface is unclear, the module is probably too broad. Split it into smaller modules with single responsibilities.

Step 2: Create the Module UI Function

#' Data Filter Module UI
#'
#' @param id Module namespace ID
#' @return A tagList of filter controls
#' @export
dataFilterUI <- function(id) {
  ns <- NS(id)
  tagList(
    selectInput(
      ns("column"),
      "Filter column",
      choices = NULL
    ),
    uiOutput(ns("filter_control")),
    actionButton(ns("apply"), "Apply Filter", class = "btn-primary")
  )
}

Key rules:

  • Function name follows <name>UI convention
  • First argument is always id
  • Create ns <- NS(id) at the top
  • Wrap every inputId and outputId with ns()
  • Return a tagList() to allow flexible placement

Got: UI function that creates namespaced input/output elements.

If fail: If IDs collide when using the module twice, check that every ID is wrapped with ns(). Common miss: IDs inside renderUI() or uiOutput() — these need ns() too.

Step 3: Create the Module Server Function

#' Data Filter Module Server
#'
#' @param id Module namespace ID
#' @param data Reactive expression returning a data frame
#' @param columns Character vector of filterable column names
#' @return Reactive expression returning the filtered data frame
#' @export
dataFilterServer <- function(id, data, columns) {
  moduleServer(id, function(input, output, session) {
    ns <- session$ns

    # Update column choices when data changes
    observeEvent(data(), {
      available <- intersect(columns, names(data()))
      updateSelectInput(session, "column", choices = available)
    })

    # Dynamic filter control based on selected column
    output$filter_control <- renderUI({
      req(input$column)
      col_data <- data()[[input$column]]

      if (is.numeric(col_data)) {
        sliderInput(
          ns("value_range"),
          "Range",
          min = min(col_data, na.rm = TRUE),
          max = max(col_data, na.rm = TRUE),
          value = range(col_data, na.rm = TRUE)
        )
      } else {
        selectInput(
          ns("value_select"),
          "Values",
          choices = unique(col_data),
          multiple = TRUE,
          selected = unique(col_data)
        )
      }
    })

    # Return filtered data as a reactive
    filtered <- eventReactive(input$apply, {
      req(input$column)
      col <- input$column
      df <- data()

      if (is.numeric(df[[col]])) {
        req(input$value_range)
        df[df[[col]] >= input$value_range[1] &
           df[[col]] <= input$value_range[2], ]
      } else {
        req(input$value_select)
        df[df[[col]] %in% input$value_select, ]
      }
    }, ignoreNULL = FALSE)

    return(filtered)
  })
}

Key rules:

  • Function name follows <name>Server convention
  • First argument is always id
  • Additional arguments are reactive expressions or static values
  • Use moduleServer(id, function(input, output, session) { ... })
  • Use session$ns for dynamic UI created inside the server
  • Return reactive values explicitly

Got: Server function that processes inputs and returns reactive output.

If fail: If reactive values don't update, check that inputs from dynamic UI use session$ns (not the outer ns). If the module returns NULL, ensure return() is the last expression inside moduleServer().

Step 4: Wire the Module into the Parent App

# In app_ui.R or ui
ui <- page_sidebar(
  title = "Analysis App",
  sidebar = sidebar(
    dataFilterUI("filter1")
  ),
  card(
    DT::dataTableOutput("table")
  )
)

# In app_server.R or server
server <- function(input, output, session) {
  # Raw data source
  raw_data <- reactive({ mtcars })

  # Call module — capture its return value
  filtered_data <- dataFilterServer(
    "filter1",
    data = raw_data,
    columns = c("cyl", "mpg", "hp", "wt")
  )

  # Use the module's returned reactive
  output$table <- DT::renderDataTable({
    filtered_data()
  })
}

Got: Module appears in the UI and its returned reactive flows into downstream outputs.

If fail: If the module UI doesn't render, verify the id string matches between UI and server calls. If the returned reactive is NULL, check that the server function returns a value.

Step 5: Compose Nested Modules (Optional)

For modules that contain other modules:

analysisUI <- function(id) {
  ns <- NS(id)
  tagList(
    dataFilterUI(ns("filter")),
    plotOutput(ns("plot"))
  )
}

analysisServer <- function(id, data) {
  moduleServer(id, function(input, output, session) {
    # Call inner module with namespaced ID
    filtered <- dataFilterServer("filter", data = data, columns = names(data()))

    output$plot <- renderPlot({
      req(filtered())
      plot(filtered())
    })

    return(filtered)
  })
}

Key rule: In the UI, nest with ns("inner_id"). In the server, call with just "inner_id"moduleServer handles the namespace chaining.

Got: Inner module renders correctly within the outer module's namespace.

If fail: If the inner module's UI doesn't appear, you likely forgot ns() around the inner module's ID in the outer UI function. If server communication breaks, check that the inner module ID matches (no ns() in the server call).

Step 6: Test the Module in Isolation

# Quick test app for the module
if (interactive()) {
  shiny::shinyApp(
    ui = fluidPage(
      dataFilterUI("test"),
      DT::dataTableOutput("result")
    ),
    server = function(input, output, session) {
      data <- reactive(iris)
      filtered <- dataFilterServer("test", data, names(iris))
      output$result <- DT::renderDataTable(filtered())
    }
  )
}

Got: Module works correctly in the minimal test app.

If fail: If the module fails in isolation but works in the full app (or vice versa), check for implicit dependencies on global variables or parent session state.

Validation

  • Module UI function accepts id as first argument and uses NS(id)
  • Every input/output ID in the UI is wrapped with ns()
  • Module server uses moduleServer(id, function(input, output, session) { ... })
  • Dynamic UI in server uses session$ns for IDs
  • Module can be instantiated multiple times without ID collisions
  • Reactive return values are accessible to the parent app
  • Module works in a minimal standalone test app

Pitfalls

  • Forgetting ns() in renderUI(): Dynamic UI created inside the server must use session$ns — the outer ns is not available inside moduleServer().
  • Passing non-reactive data: Module arguments that change over time must be reactive expressions. Pass reactive(data) not data.
  • ID mismatch: The id string in the UI call must exactly match the id in the server call.
  • Not returning reactives: If the module computes something the parent needs, it must return() a reactive. Forgetting this is a silent bug.
  • Namespace in nested modules: In UI: ns("inner_id"). In server: just "inner_id". Mixing these up causes namespace double-wrapping or missing prefixes.

Related Skills

  • scaffold-shiny-app — set up the app structure before adding modules
  • test-shiny-app — test modules with testServer() unit tests
  • design-shiny-ui — bslib layout and theming for module UIs
  • optimize-shiny-performance — cache and async patterns within modules

GitHub Repository

pjt222/agent-almanac
Path: i18n/caveman-lite/skills/build-shiny-module
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

Related Skills

content-collections

Meta

This skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.

View skill

polymarket

Meta

This skill enables developers to build applications with the Polymarket prediction markets platform, including API integration for trading and market data. It also provides real-time data streaming via WebSocket to monitor live trades and market activity. Use it for implementing trading strategies or creating tools that process live market updates.

View skill

creating-opencode-plugins

Meta

This skill helps developers create OpenCode plugins that hook into 25+ event types like commands, files, and LSP operations. It provides the plugin structure, event API specifications, and implementation patterns for JavaScript/TypeScript modules. Use it when you need to intercept, monitor, or extend the OpenCode AI assistant's lifecycle with custom event-driven logic.

View skill

sglang

Meta

SGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.

View skill