Customizing Module Output

Introduction

teal is a powerful shiny-based framework with built-in modules for interactive data analysis. This document outlines the customization options available for modifying the output of teal modules. You will learn how to use teal_transform_module() to modify and enhance the objects created by teal::modules(), enabling you to tailor the outputs to your specific requirements without rewriting the original module code.

Decorators

In programming, decoration refers to the process of modifying an object while preserving its original class. For instance, given an object x of class "my_class", a function foo(x) is considered a decorator function if it modifies x and returns an object that retains the same class. In this context, x is referred to as the decorated object, and foo() is the decorator function or decorator. Decorators can perform a variety of operations, such as adding new methods or modifying data, while ensuring the object remains compatible with its original usage.

In the context of teal applications, decoration is specifically used to modify module outputs, such as plots or tables. For example, consider a decorator function add_title(x, <plot_title>) that takes a ggplot2 plot object (x) as input, applies a title modification, and returns a modified ggplot2 plot object. This function qualifies as a decorator because it preserves the original class of the input object. Conversely, a function like create_plot(<data>, <x_axis>, <y_axis>), which generates a new plot object, is not a decorator, as it produces an output of a different class.

Preserving the object’s class during decoration is essential for compatibility. It ensures that the subsequent “display” logic can seamlessly handle both decorated and non-decorated objects.

The decoration process can vary in complexity:

Simple Decorations: Single-step modifications, such as a single method call that does not require additional data.
Complex Decorations: Multi-step operations that may involve interdependent transformations, potentially requiring input from dedicated shiny UI elements.

This powerful functionality empowers application developers to significantly customize outputs beyond the default capabilities provided by existing module parameters. Decorations allow for advanced modifications, enabling highly tailored and dynamic user experiences in teal applications.

Requirements and Limitations

To use decorators effectively, certain requirements must be met:

Module Support: While teal provides the core functionality for decorators, the module must explicitly support this functionality. Developers should ensure that the module has been designed to work with decorators.
Matching Object Names: Decorators must reference object names that align with the internal naming conventions of the module. Each module may use different names for its output objects, such as plot or table. This alignment is critical for successful decoration.

It is recommended to review the module documentation or source code to understand its internal object naming before applying decorators.

Decorators in `teal`

One of ways of adjusting input data or customizing module outputs in teal is the usage of transformators created through teal_transform_module.

In below chapter we will present how to create the simplest static decorator with just a server part. Later, we will present examples on more advanced usage, where decorators contain UI. You will also learn about a convenience function that makes it easier to write decorators, called make_teal_transform_server(). The chapter ends with an example module that utilizes decorators and a snippet that uses this module in teal application.

Server

The simplest way to create a decorator is to use teal_transform_module() with only server argument provided (i.e. without UI part). This approach adds functionality solely to the server code of the module. In the following example, we assume that the module contains an object (of class ggplot2) named plot. We modify the title and x-axis label of plot:

library(teal)

## Loading required package: shiny

## Loading required package: teal.data

## Loading required package: teal.code

## Loading required package: teal.slice

## Registered S3 method overwritten by 'teal':
##   method        from      
##   c.teal_slices teal.slice

## 
## You are using teal version 0.15.2.9095

## 
## Attaching package: 'teal'

## The following objects are masked from 'package:teal.slice':
## 
##     as.teal_slices, teal_slices

static_decorator <- teal_transform_module(
  label = "Static decorator",
  server = function(id, data) {
    moduleServer(id, function(input, output, session) {
      reactive({
        req(data())
        within(data(), {
          plot <- plot +
            ggtitle("This is title") +
            xlab("x axis")
        })
      })
    })
  }
)

To simplify the repetitive elements of writing new decorators (e.g., function(id, data), moduleServer, reactive, within(data, ...)), you can use the make_teal_transform_server() convenience function, which takes a language as input:

static_decorator_lang <- teal_transform_module(
  label = "Static decorator (language)",
  server = make_teal_transform_server(
    expression(
      plot <- plot +
        ggtitle("This is title") +
        xlab("x axis title")
    )
  )
)

UI

To create a decorator with user interactivity, you can add (optional) UI part and use it in server accordingly (i.e. a typical shiny module). In the example below, the x-axis title is set dynamically via a textInput, allowing users to specify their preferred label. Note how the input parameters are passed to the within() function using its ... argument.

interactive_decorator <- teal_transform_module(
  label = "Interactive decorator",
  ui = function(id) {
    ns <- NS(id)
    div(
      textInput(ns("x_axis_title"), "X axis title", value = "x axis")
    )
  },
  server = function(id, data) {
    moduleServer(id, function(input, output, session) {
      reactive({
        req(data())
        within(data(),
          {
            plot <- plot +
              ggtitle("This is title") +
              xlab(my_title)
          },
          my_title = input$x_axis_title
        )
      })
    })
  }
)

As in the earlier examples, make_teal_transform_server() can simplify the creation of the server component. This wrapper requires you to use input object names directly in the expression - note that we have xlab(x_axis_table) and not my_title = input$x_axis_title together with xlab(my_title).

interactive_decorator_lang <- teal_transform_module(
  label = "Interactive decorator (language)",
  ui = function(id) {
    ns <- NS(id)
    div(
      textInput(ns("x_axis_title"), "X axis title", value = "x axis")
    )
  },
  server = make_teal_transform_server(
    expression(
      plot <- plot +
        ggtitle("This is title") +
        xlab(x_axis_title)
    )
  )
)

Handling Various Object Names

teal_transform_module relies on the names of objects created within a module. Writing a decorator that applies to any module can be challenging because different modules may use different object names. It is recommended to create a library of decorator functions that can be adapted to the specific object names used in teal modules. In the following example, focus on the output_name parameter to see how decorator can be applied to multiple modules:

gg_xlab_decorator <- function(output_name) {
  teal_transform_module(
    label = "X-axis decorator",
    ui = function(id) {
      ns <- NS(id)
      div(
        textInput(ns("x_axis_title"), "X axis title", value = "x axis")
      )
    },
    server = function(id, data) {
      moduleServer(id, function(input, output, session) {
        reactive({
          req(data())
          within(data(),
            {
              output_name <- output_name +
                xlab(x_axis_title)
            },
            x_axis_title = input$x_axis_title,
            output_name = as.name(output_name)
          )
        })
      })
    }
  )
}

Decorator failures are managed by an internal teal mechanism called trigger on success, which ensures that the data object within the module remains intact. If a decorator fails, the outputs will not be shown, and an appropriate error message will be displayed.

failing_decorator <- teal_transform_module(
  label = "Failing decorator",
  ui = function(id) {
    ns <- NS(id)
    div(
      textInput(ns("x_axis_title"), "X axis title", value = "x axis")
    )
  },
  server = function(id, data) {
    moduleServer(id, function(input, output, session) {
      reactive(stop("\nThis is an error produced by decorator\n"))
    })
  }
)

Decorating Plots

Example Module

To include decorators in a teal module, pass them as arguments (ui_args and server_args) to the module’s ui and server components, where they will be used by ui/srv_teal_transform_module.

Please find an example module for the sake of this article:

tm_decorated_plot <- function(label = "module", transformators = list(), decorators = NULL) {
  checkmate::assert_list(decorators, "teal_transform_module", null.ok = TRUE)

  module(
    label = label,
    ui = function(id, decorators) {
      ns <- NS(id)
      div(
        selectInput(ns("dataname"), label = "select dataname", choices = NULL),
        selectInput(ns("x"), label = "select x", choices = NULL),
        selectInput(ns("y"), label = "select y", choices = NULL),
        ui_transform_teal_data(ns("decorate"), transformators = decorators),
        plotOutput(ns("plot")),
        verbatimTextOutput(ns("text"))
      )
    },
    server = function(id, data, decorators) {
      moduleServer(id, function(input, output, session) {
        observeEvent(data(), {
          updateSelectInput(inputId = "dataname", choices = names(data()))
        })

        observeEvent(input$dataname, {
          req(input$dataname)
          updateSelectInput(inputId = "x", choices = colnames(data()[[input$dataname]]))
          updateSelectInput(inputId = "y", choices = colnames(data()[[input$dataname]]))
        })

        dataname <- reactive(req(input$dataname))
        x <- reactive({
          req(input$x, input$x %in% colnames(data()[[dataname()]]))
          input$x
        })

        y <- reactive({
          req(input$y, input$y %in% colnames(data()[[dataname()]]))
          input$y
        })
        plot_data <- reactive({
          req(dataname(), x(), y())
          within(data(),
            {
              plot <- ggplot2::ggplot(dataname, ggplot2::aes(x = x, y = y)) +
                ggplot2::geom_point()
            },
            dataname = as.name(dataname()),
            x = as.name(x()),
            y = as.name(y())
          )
        })

        plot_data_decorated_no_print <- srv_transform_teal_data(
          "decorate",
          data = plot_data,
          transformators = decorators
        )
        plot_data_decorated <- reactive(
          within(req(plot_data_decorated_no_print()), expr = plot)
        )

        plot_r <- reactive({
          plot_data_decorated()[["plot"]]
        })

        output$plot <- renderPlot(plot_r())
        output$text <- renderText({
          teal.code::get_code(req(plot_data_decorated()))
        })
      })
    },
    ui_args = list(decorators = decorators),
    server_args = list(decorators = decorators)
  )
}

Application

library(ggplot2)

## Error in get(paste0(generic, ".", class), envir = get_method_env()) : 
##   object 'type_sum.accel' not found

app <- init(
  data = teal_data(iris = iris, mtcars = mtcars),
  modules = modules(
    tm_decorated_plot("identity"),
    tm_decorated_plot("no-ui", decorators = list(static_decorator)),
    tm_decorated_plot("lang", decorators = list(static_decorator_lang)),
    tm_decorated_plot("interactive", decorators = list(interactive_decorator)),
    tm_decorated_plot("interactive-from lang", decorators = list(interactive_decorator_lang)),
    tm_decorated_plot("from-fun", decorators = list(gg_xlab_decorator("plot"))),
    tm_decorated_plot("failing", decorators = list(failing_decorator))
  )
)

if (interactive()) {
  shinyApp(app$ui, app$server)
}

Multiple Decorators

Example Module

It is possible to pass any number of decorators (n) to a module. The example below demonstrates how to handle a dynamic number of decorators, allowing the user to choose which decorator to apply from a list. This makes the module more flexible and capable of accommodating various customization requirements.

library(ggplot2)
tm_decorated_plot <- function(label = "module", decorators = NULL) {
  checkmate::assert_list(decorators, "teal_transform_module", null.ok = TRUE)
  module(
    label = label,
    ui = function(id, decorators) {
      ns <- NS(id)
      div(
        selectInput(ns("dataname"), label = "Select dataset", choices = NULL),
        selectInput(ns("x"), label = "Select x-axis", choices = NULL),
        selectInput(ns("y"), label = "Select y-axis", choices = NULL),
        selectInput(
          ns("decorator_choice"),
          "Choose decorator",
          choices = names(decorators),
          selected = names(decorators)[1]
        ),
        div(
          id = ns("decorate_wrapper"),
          lapply(names(decorators), function(decorator_name) {
            div(
              id = ns(paste0("decorate_", decorator_name)),
              ui_transform_teal_data(
                ns(paste0("decorate_", decorator_name)),
                transformators = decorators[[decorator_name]]
              )
            )
          })
        ),
        plotOutput(ns("plot")),
        verbatimTextOutput(ns("text"))
      )
    },
    server = function(id, data, decorators) {
      moduleServer(id, function(input, output, session) {
        observeEvent(data(), {
          updateSelectInput(inputId = "dataname", choices = names(data()))
        })

        dataname <- reactive(req(input$dataname))

        observeEvent(dataname(), {
          updateSelectInput(inputId = "x", choices = colnames(data()[[input$dataname]]))
          updateSelectInput(inputId = "y", choices = colnames(data()[[input$dataname]]))
        })

        observeEvent(input$decorator_choice, {
          # Dynamically show only the selected decorator's UI
          lapply(names(decorators), function(decorator_name) {
            if (decorator_name == input$decorator_choice) {
              shinyjs::show(paste0("decorate_", decorator_name))
            } else {
              shinyjs::hide(paste0("decorate_", decorator_name))
            }
          })
        })

        x <- reactive({
          req(input$x, input$x %in% colnames(data()[[dataname()]]))
          input$x
        })

        y <- reactive({
          req(input$y, input$y %in% colnames(data()[[dataname()]]))
          input$y
        })
        plot_data <- reactive({
          req(dataname(), x(), y())
          within(data(),
            {
              plot <- ggplot2::ggplot(dataname, ggplot2::aes(x = x, y = y)) +
                ggplot2::geom_point()
            },
            dataname = as.name(dataname()),
            x = as.name(x()),
            y = as.name(y())
          )
        })

        selected_decorator <- reactive({
          req(input$decorator_choice)
          input$decorator_choice
        })

        decorated_data_no_print <- srv_transform_teal_data(
          sprintf("decorate_%s", selected_decorator()),
          data = plot_data,
          transformators = decorators[[selected_decorator()]]
        )
        decorated_data <- reactive(within(req(decorated_data_no_print()), expr = plot))

        output$plot <- renderPlot(decorated_data()[["plot"]])
        output$text <- renderText({
          req(input$decorator_choice)
          teal.code::get_code(req(decorated_data()))
        })
      })
    },
    ui_args = list(decorators = decorators),
    server_args = list(decorators = decorators)
  )
}

By order of the decorator we will:

Change the x axis title
Change the y axis title
Replace the x axis title

interactive_decorator_1 <- teal_transform_module(
  label = "Interactive decorator 1",
  ui = function(id) {
    ns <- NS(id)
    div(
      textInput(ns("x_axis_title"), "X axis title", value = "x axis 1")
    )
  },
  server = function(id, data) {
    moduleServer(id, function(input, output, session) {
      reactive({
        req(data())
        within(data(),
          {
            plot <- plot +
              xlab(title)
          },
          title = input$x_axis_title
        )
      })
    })
  }
)

interactive_decorator_2 <- teal_transform_module(
  label = "Interactive decorator 2",
  ui = function(id) {
    ns <- NS(id)
    div(
      textInput(ns("y_axis_title"), "Y axis title", value = "y axis 1")
    )
  },
  server = function(id, data) {
    moduleServer(id, function(input, output, session) {
      reactive({
        req(data())
        within(data(),
          {
            plot <- plot +
              ylab(title)
          },
          title = input$y_axis_title
        )
      })
    })
  }
)

interactive_decorator_3 <- teal_transform_module(
  label = "Interactive decorator 3",
  ui = function(id) {
    ns <- NS(id)
    div(
      textInput(ns("x_axis_title"), "X axis title", value = "x axis 3")
    )
  },
  server = function(id, data) {
    moduleServer(id, function(input, output, session) {
      reactive({
        req(data())
        within(data(),
          {
            plot <- plot +
              xlab(title)
          },
          title = input$x_axis_title
        )
      })
    })
  }
)

Application

As you might have noted, the x axis title from the first decorator will be used but won’t show up on the resulting plot:

app <- init(
  data = teal_data(iris = iris, mtcars = mtcars),
  modules = modules(
    tm_decorated_plot(
      "dynamic_decorators",
      decorators = list(
        decorator_1 = interactive_decorator_1,
        decorator_2 = interactive_decorator_2,
        decorator_3 = interactive_decorator_3
      )
    )
  )
)

if (interactive()) {
  shinyApp(app$ui, app$server)
}

Modules with Multiple Outputs

In this section, we demonstrate how to extend a teal module to handle multiple outputs and allow separate decoration for each. Specifically, the module will have two outputs:

a ggplot plot
and a table

We will apply independent decorators to each output.

Example Module with Two Outputs

The following module generates both a scatter plot and a summary table. Each of these outputs can be decorated independently using decorators passed to the module:

tm_decorated_plot_table <- function(label = "module with two outputs", decorators = list()) {
  checkmate::assert_list(decorators, "teal_transform_module", null.ok = TRUE)

  module(
    label = label,
    ui = function(id, decorators) {
      ns <- NS(id)
      div(
        selectInput(ns("dataname"), label = "Select dataset", choices = NULL),
        selectInput(ns("x"), label = "Select x-axis", choices = NULL),
        selectInput(ns("y"), label = "Select y-axis", choices = NULL),
        ui_transform_teal_data(ns("decorate_plot"), transformators = decorators$plot),
        ui_transform_teal_data(ns("decorate_table"), transformators = decorators$table),
        plotOutput(ns("plot")),
        tableOutput(ns("table")),
        verbatimTextOutput(ns("text"))
      )
    },
    server = function(id, data, decorators) {
      moduleServer(id, function(input, output, session) {
        observeEvent(data(), {
          updateSelectInput(inputId = "dataname", choices = names(data()))
        })

        dataname <- reactive(req(input$dataname))

        observeEvent(dataname(), {
          updateSelectInput(inputId = "x", choices = colnames(data()[[input$dataname]]))
          updateSelectInput(inputId = "y", choices = colnames(data()[[input$dataname]]))
        })

        x <- reactive({
          req(input$x, input$x %in% colnames(data()[[dataname()]]))
          input$x
        })

        y <- reactive({
          req(input$y, input$y %in% colnames(data()[[dataname()]]))
          input$y
        })

        # Generate plot data
        plot_data <- reactive({
          req(dataname(), x(), y())
          within(data(),
            {
              plot <- ggplot2::ggplot(dataname, ggplot2::aes(x = xvar, y = yvar)) +
                ggplot2::geom_point()
            },
            dataname = as.name(dataname()),
            xvar = as.name(x()),
            yvar = as.name(y())
          )
        })

        # Generate table data
        table_data <- reactive({
          req(dataname())
          within(data(),
            {
              table_data <- data.frame(Filter(Negate(is.na), lapply(dataname, mean, na.rm = TRUE)))
            },
            dataname = as.name(dataname())
          )
        })

        # Apply decorators to plot
        decorated_plot <- srv_transform_teal_data(
          "decorate_plot",
          data = plot_data,
          transformators = decorators$plot
        )

        # Apply decorators to table
        decorated_table <- srv_transform_teal_data(
          "decorate_table",
          data = table_data,
          transformators = decorators$table
        )

        output$plot <- renderPlot(decorated_plot()[["plot"]])

        output$table <- renderTable(decorated_table()[["table_data"]])

        output$text <- renderText({
          plot_code <- teal.code::get_code(req(decorated_plot()))
          table_code <- teal.code::get_code(req(decorated_table()))
          paste("# Plot Code:", plot_code, "\n\n# Table Code:", table_code)
        })
      })
    },
    ui_args = list(decorators = decorators),
    server_args = list(decorators = decorators)
  )
}

Example Decorators

Plot Decorator: Adds a title to the plot.

plot_decorator <- teal_transform_module(
  label = "Decorate plot",
  ui = function(id) {
    ns <- NS(id)
    textInput(ns("plot_title"), "Plot Title", value = "Decorated Title (editable)")
  },
  server = function(id, data) {
    moduleServer(id, function(input, output, session) {
      reactive({
        req(data())
        within(data(),
          {
            plot <- plot + ggplot2::ggtitle(ptitle) +
              ggplot2::theme_minimal() +
              ggplot2::theme(
                plot.title = element_text(face = "bold", size = 30, color = "blue")
              )
          },
          ptitle = input$plot_title
        )
      })
    })
  }
)

Table Decorator: Adds row names to the summary table.

table_decorator <- teal_transform_module(
  label = "Decorate table",
  ui = function(id) shiny::tags$p("No UI needed for table decorator and could be ommited."),
  server = make_teal_transform_server(
    expression({
      table_data[["Added by decorator"]] <- paste0("Row ", seq_len(nrow(table_data)))
    })
  )
)

Application

app <- init(
  data = teal_data(iris = iris, mtcars = mtcars),
  modules = modules(
    tm_decorated_plot_table(
      "plot_and_table",
      decorators = list(
        plot = plot_decorator,
        table = table_decorator
      )
    )
  )
)

if (interactive()) {
  shinyApp(app$ui, app$server)
}

NEST CoreDev

Introduction

Decorators

Requirements and Limitations

Decorators in teal

Server

UI

Handling Various Object Names

Decorating Plots

Example Module

Application

Multiple Decorators

Example Module

Application

Modules with Multiple Outputs

Example Module with Two Outputs

Example Decorators

Application

Decorators in `teal`