Package 'shinyfilters'

Description

Applies a list of filters to an object, returning the filtered object.

Usage

apply_filters(
  x,
  filter_list,
  filter_combine_method = "and",
  expanded = FALSE,
  cols = NULL,
  ...
)

Arguments

Value

A filtered object, or a named list of filtered objects if expanded = TRUE. In a filtered data frame, a column whose class has no [ method keeps its attributes, such as a class that chooses its input.

Examples

library(S7)
df <- data.frame(
 stringsAsFactors = FALSE,
 category = rep(letters[1:3], each = 4),
 value = 1:12,
 date = as.Date('2024-01-01') + 0:11
)
filters <- list(
  category = c("a", "b"),
  value = c(3, 11)
)

# Apply filters with logical AND
apply_filters(df, filters, filter_combine_method = "and")

# Apply filters with logical OR
apply_filters(df, filters, filter_combine_method = "or")

# Get expanded filters
apply_filters(df, filters, expanded = TRUE)

Description

Provides the appropriate function arguments for the input function selected by filterInput() or updateFilterInput().

Usage

args_filter_input(x, ...)

args_update_filter_input(x, ...)

Arguments

Details

The following arguments are supported in ...:

range

(Date, POSIXt). Logical. If TRUE, args_filter_input() will provide the arguments for range date inputs. Only applies when x is of class Date or POSIXt.

textbox

(character). Logical. If FALSE (the default), args_filter_input() will provide the arguments for select inputs.

choices_asis

(character, factor, logical). Logical. If TRUE, the choices provided to select inputs will not be modified. If FALSE (the default), duplicate values will be removed and the choices will be sorted. Only applies when x is of class character, factor, or logical.

server

If TRUE, indicates that the choices will be provided server-side. In this case, arguments are not computed for args_filter_input(). Ignored in args_update_filter_input().

args_unique

An optional named list of arguments passed to unique(), called when x is a character, factor, or logical, textbox = FALSE, and choices_asis = FALSE.

args_sort

An optional named list of arguments passed to sort(), which is called after unique().

Value

A named list of arguments for a shiny input function

Examples

args_filter_input(iris$Petal.Length)

Description

Internal function used to prepare input arguments using args_filter_input(), and gracefully pass them to provided input function.

Usage

call_filter_input(x, .f, ...)

call_update_filter_input(x, .f, ...)

Arguments

Details

call_filter_input() and call_update_filter_input() are used when customizing shinyfilters. For more, see vignette("customizing-shinyfilters").

Value

The result of calling the provided input function.

Examples

library(S7)
library(shiny)
# call_filter_input() is used inside filterInput() methods
method(filterInput, class_numeric) <- function(x, ...) {
  call_filter_input(x, sliderInput, ...)
}

# call_update_filter_input() is used inside updateFilterInput() methods
method(updateFilterInput, class_numeric) <- function(x, ...) {
  call_update_filter_input(x, updateSliderInput, ...)
}

Description

Selects and creates a shiny input based the type of object x and other arguments.

Usage

filterInput(x, ...)

Arguments

Details

The following arguments passed to ... are supported:

area

(character). Logical. Controls whether to use textAreaInput (TRUE) or textInput (FALSE, default). Only applies when textbox is TRUE.

radio

(character, factor, list, logical). Logical. Controls whether to use radioButtons (TRUE) or a dropdown input (FALSE, default). For character vectors, radio only applies if textbox is FALSE, the default.

range

(Date, POSIXt). Logical. Controls whether to use dateRangeInput (TRUE) or dateInput (FALSE, default).

selectize

(character, factor, list, logical). Logical. Controls whether to use selectizeInput (TRUE) or selectInput (FALSE, default). For character vectors, selectize only applies if textbox is FALSE, the default.

slider

(numeric). Logical. Controls whether to use sliderInput (TRUE) or numericInput (FALSE, default).

textbox

(character). Logical. Controls whether to use a text input (TRUE) or a dropdown input (FALSE, default).

ns

An optional namespace created by NS(). Useful when using filterInput() on a data.frame inside a shiny module.

Remaining arguments passed to ... are passed to the args_filter_input() or the selected input function.

Value

One of the following shiny inputs is returned, based on the type of object passed to x, and other specified arguments. See vignette("filter-input-catalog") for the full list of examples.

Examples

library(shiny)

ui <- fluidPage(
	 sidebarLayout(
		 sidebarPanel(
			 # Create a filterInput() inside a shiny app:
			 filterInput(
				x = letters,
				inputId = "letter",
				label = "Pick a letter:"
			 )
		 ),
		 mainPanel(
			 textOutput("selected_letter")
		 )
	 )
)

server <- function(input, output, session) {
	 output$selected_letter <- renderText({
		 paste("You selected:", input$letter)
	 })
}

shinyApp(ui, server)

Description

Computes a logical vector indicating which elements of x match the filter criteria specified by val.

Usage

get_filter_logical(x, val, ...)

Arguments

Details

The following arguments are supported in ...:

column

When x is a data.frame, column is the name of the column intended to be filtered.

comparison

When x is a numeric or Date and val is a length-one numeric or Date, comparison is the function used to compare x with val. The default is <=.

gte

When x is a numeric or Date and val is a length-two numeric or Date, gte controls whether to use >= (TRUE, default) or > (FALSE) on val[[1]].

lte

When x is a numeric or Date and val is a length-two numeric or Date, lte controls whether to use <= (TRUE, default) or < (FALSE) on val[[2]].

When val is a character vector, such as the value of a select or radio input, and x isn't a character, factor, or logical vector, the elements of x are compared as text. A datetime x is compared as its date, whether val is text, a Date, or a datetime.

Value

A logical vector indicating which elements of x match the filter criteria specified by val.

Examples

df <- data.frame(
  stringsAsFactors = FALSE,
  category = rep(letters[1:3], each = 4),
  value = 1:12,
  date = Sys.Date() + 0:11
)

# Filter character column
get_filter_logical(df, c("a", "b"), column = "category")

# Filter numeric column with single value
get_filter_logical(df, 5, column = "value", comparison = `<=`)

# Filter numeric column with range
get_filter_logical(df, c(3, 8), column = "value", gte = TRUE, lte = FALSE)

Description

Returns the (unnamespaced) ids of the inputs for the provided object.

Usage

get_input_ids(x, ...)

Arguments

Value

A character vector of input ids.

Examples

df <- data.frame(
  stringsAsFactors = FALSE,
  name = c("Alice", "Bob"),
  age = c(25, 30),
  completed = c(TRUE, FALSE)
)

get_input_ids(df)

Description

Returns the labels of the shiny inputs for the provided object.

Usage

get_input_labels(x, ...)

Arguments

Value

A character vector of input labels

Examples

df <- data.frame(
  stringsAsFactors = FALSE,
  name = c("Alice", "Bob"),
  age = c(25, 30),
  completed = c(TRUE, FALSE)
)

get_input_labels(df)

Description

Retrieves multiple input values from a shiny input object based on the names provided in x.

Usage

get_input_values(input, x, ...)

Arguments

Value

A named list of input values corresponding to the names in x.

Examples

library(shiny)
df <- data.frame(
	 stringsAsFactors = FALSE,
  name = c("Alice", "Bob"),
  age = c(25, 30),
  completed = c(TRUE, FALSE)
)
ui <- fluidPage(
  sidebarLayout(
    sidebarPanel(
      filterInput(df)
    ),
    mainPanel(
      verbatimTextOutput("output_all"),
      verbatimTextOutput("output_subset")
    )
  )
)
server <- function(input, output, session) {
  output$output_all <- renderPrint({
    get_input_values(input, df)
  })
  output$output_subset <- renderPrint({
    get_input_values(input, c("name", "completed"))
  })
}
shinyApp(ui, server)

Description

A sample of 200 flights that departed New York City airports in 2013, with a mix of column types for trying out filterInput().

Usage

nyc_flights

Format

A data frame with 200 rows and 7 columns:

date

Date of departure.

carrier

Two-letter airline code.

origin

Origin airport: "EWR", "JFK", or "LGA".

dest

Destination airport code.

dep_delay

Departure delay in minutes. Negative values are early departures; NA for cancelled flights.

distance

Distance between airports in miles.

delayed

Whether the flight arrived more than 15 minutes late; NA when the arrival time is unknown.

Source

A random sample of flights from the nycflights13 package (https://github.com/tidyverse/nycflights13), released under CC0.

Description

shinyfilters_server() runs an observer that makes your filters interdependent. serverFilterInput() is an alias of shinyfilters_server().

Usage

serverFilterInput(
  x,
  session = getDefaultReactiveDomain(),
  filter_combine_method = "and",
  args_apply_filters = NULL,
  ...,
  input = deprecated()
)

shinyfilters_server(
  x,
  session = getDefaultReactiveDomain(),
  filter_combine_method = "and",
  args_apply_filters = NULL,
  ...,
  input = deprecated()
)

Arguments

Details

Only the inputs that have no value are updated, and they are left without one. The inputs of a configuration made by shinyfilters() are updated as updateFilterInput() updates them, from the filtered data.

An input set by a function other than the shiny inputs filterInput() creates needs the function that updates it. A shinyWidgets input, such as shinyWidgets::pickerInput, comes with its own. A column's update sends its choices, or its minimum and maximum, and that function is given the ones it has an argument for. One with an argument for none of them, such as shinyWidgets::updateNumericRangeInput, leaves its input as it is, so the input doesn't follow the other filters. Any other input, such as shiny::checkboxGroupInput, needs its update function named with .update_fn = fn among the input's arguments, as described in with_filters(). Without it, shinyfilters_server() errors.

Value

A reactiveValues list with two elements: filtered, the filtered data, and input_values, the current filter input values as a named list.

Examples

library(bslib)
library(DT)
library(S7)
library(shiny)

must_use_radio <- new_S3_class(
	 class = "must_use_radio",
	 constructor = function(.data) .data
)
method(filterInput, must_use_radio) <- function(x, ...) {
	 call_filter_input(x, radioButtons, ...)
}
method(updateFilterInput, must_use_radio) <- function(x, ...) {
	 call_update_filter_input(x, updateRadioButtons, ...)
}

use_radio <- function(x) {
	 structure(x, class = unique(c("must_use_radio", class(x))))
}

df_shared <- data.frame(
  stringsAsFactors = FALSE,
	 x = letters,
	 y = use_radio(sample(c("red", "green", "blue"), 26, replace = TRUE)),
	 z = round(runif(26, 0, 3.5), 2),
	 q = sample(Sys.Date() - 0:7, 26, replace = TRUE)
)

filters_ui <- function(id) {
	 ns <- NS(id)
	 filterInput(
		 x = df_shared,
		 range = TRUE,
		 selectize = TRUE,
		 slider = TRUE,
		 multiple = TRUE,
		 ns = ns
	 )
}

filters_server <- function(id) {
	 moduleServer(id, function(input, output, session) {
 		# serverFilterInput() returns a reactiveValues list
 		serverFilterInput(df_shared, range = TRUE)
 	})
}

ui <- page_sidebar(
 	sidebar = sidebar(filters_ui("demo")),
 	DTOutput("df_full"),
 	verbatimTextOutput("input_values"),
 	DTOutput("df_filt")
)

server <- function(input, output, session) {
	 res <- filters_server("demo")
	 output$df_full <- renderDT(datatable(df_shared))
	 output$input_values <- renderPrint(res$input_values)
	 output$df_filt <- renderDT(datatable(apply_filters(
 		df_shared,
 		res$input_values
 	)))
}

shinyApp(ui, server)

Description

shinyfilters() creates a set of interdependent filters for use in a shiny application.

Usage

shinyfilters(
  .data,
  ...,
  area = FALSE,
  radio = FALSE,
  range = TRUE,
  selectize = TRUE,
  multiple = TRUE,
  slider = TRUE,
  textbox = FALSE,
  ns = NULL
)

Arguments

Value

A shinyfilters object.

See Also

with_filters(), with_ns()

Examples

shinyfilters(nyc_flights)

# Use sliders for every numeric column
shinyfilters(nyc_flights, slider = TRUE)

Description

A shinyfilters object holds a data.frame and all of the data's filters. To create one, call shinyfilters(). To modify one, use with_filters(), with_args(), with_defaults(), or with_ns().

Usage

The object behaves much like the data.frame that defines it.

filters <- shinyfilters(nyc_flights)

Names

names(filters) provides the names of the data.frame.

Shape

dim(filters), length(filters), nrow(filters), and ncol(filters) describe the dimensions.

Data

as.data.frame(filters) returns the data.

Subsetting

filters$col creates the input for one filter.

filters[[cols]] creates the inputs for all filters in cols.

filters[cols] subsets a shinyfilters object, returning only filters matching cols.

shinyfilters supports <tidy-select> in both [ and [[

In a shiny app

A shinyfilters object can be placed directly inside a ui:

library(bslib)
ui <- page_sidebar(sidebar = filters)

To enable interdependent filters, call shinyfilters_server in the server:

server <- function(...) {
  shinyfilters_server(filters)
}

Creating filters

Sometimes, you will have to create the filters as you place them in the app. The following functions create filters:

• [[

• $

• filterInput

ui <- function(request) {
  page_sidebar(sidebar = filterInput(filters))
}

Currently, it is known that you must create the filters when:

1. Bookmarking is enabled

2. When placing a filter inside bslib::accordion.

In these instances, shinyfilters will intercept and throw an informative error.

accordion(shinyfilters(nyc_flights))
#> Error:
#> ! <shinyfilters> objects cannot be used in some shiny functions.
#> i Use `[[`, `$`, or `filterInput()` to render the filters directly.

Printing

Printing shows the input each column receives:

shinyfilters(nyc_flights)
#> <shinyfilters> * 7 filters
#>
#> Filters
#>   date       <date>  dateRangeInput
#>   carrier    <chr>   selectizeInput
#>   origin     <fct>   selectizeInput
#>   dest       <chr>   selectizeInput
#>   dep_delay  <dbl>   sliderInput
#>   distance   <dbl>   sliderInput
#>   delayed    <lgl>   selectizeInput

Any modifications will be displayed:

shinyfilters(nyc_flights, range = FALSE) |>
  with_filters(date = "date") |>
  with_ns("sidebar-mod")
#> <shinyfilters> * 7 filters * namespace "sidebar-mod"
#>
#> Filters
#>   *  date       <date>  dateInput
#>      carrier    <chr>   selectizeInput
#>      origin     <fct>   selectizeInput
#>      dest       <chr>   selectizeInput
#>      dep_delay  <dbl>   sliderInput
#>      distance   <dbl>   sliderInput
#>      delayed    <lgl>   selectizeInput
#>
#> Default Overrides
#>   range = FALSE
#>
#> * Filter set by `with_filters()`

Description

Updates a shiny input based the type of object x and other arguments.

Usage

updateFilterInput(x, ...)

Arguments

Details

The following arguments passed to ... are supported:

area

(character). Logical. Controls whether to use updateTextAreaInput (TRUE) or updateTextInput (FALSE, default). Only applies when textbox is TRUE.

radio

(character, factor, list, logical). Logical. Controls whether to use updateRadioButtons (TRUE) or a dropdown input update function (FALSE, default). For character vectors, radio only applies if textbox is FALSE, the default.

range

(Date, POSIXt). Logical. Controls whether to use updateDateRangeInput (TRUE) or updateDateInput (FALSE, default).

selectize

(character, factor, list, logical). Logical. Controls whether to use updateSelectizeInput (TRUE) or updateSelectInput (FALSE, default). For character vectors, selectize only applies if textbox is FALSE, the default.

slider

(numeric). Logical. Controls whether to use updateSliderInput (TRUE) or updateNumericInput (FALSE, default).

textbox

(character). Logical. Controls whether to update a text input (TRUE) or a dropdown input (FALSE, default).

Remaining arguments passed to ... are passed to args_update_filter_input() or the selected input update function.

When x is a configuration made by shinyfilters(), each column's input is updated with the function and arguments that match the input filterInput() creates for it: the configuration's defaults, and the input and arguments with_filters() or with_args() set for it, computed from the configuration's data. The arguments that set an input's value (value, selected, start, and end) are left out, and the configuration's namespace is applied to the ids.

Value

The result of the following shiny input updates is returned, based on the type of object passed to x, and other specified arguments.

Examples

library(shiny)

fruits <- list(
	"a" = c("apples", "avocados"),
	"b" = c("bananas", "blueberries"),
	"c" = c("cherries", "cantaloupe")
)

ui <- fluidPage(
	sidebarLayout(
		sidebarPanel(
			filterInput(
				x = letters[1:3],
				inputId = "letter",
				label = "Pick a letter:",
       multiple = TRUE
			),
			filterInput(
				x = fruits,
				inputId = "fruits",
				label = "Pick a fruit:"
			)
		),
   mainPanel()
	)
)

server <- function(input, output, session) {
	observe({
		fruits_filtered <- fruits
		if (!is.null(input$letter) && length(input$letter) != 0L) {
			fruits_filtered <- fruits[input$letter]
		}
		# Call updateFilterInput() inside the shiny server:
		updateFilterInput(x = fruits_filtered, inputId = "fruits")
	})
}
shinyApp(ui, server)

Description

with_args() sets arguments for the inputs filterInput() creates for one or more columns of a configuration made by shinyfilters(), such as a slider's value, step, or width. Each column keeps its input.

Usage

with_args(.filters, ...)

Arguments

Details

Arguments stay with their column. Setting one again replaces it, and an input chosen later with with_filters() keeps the ones it has an argument for.

Value

The updated configuration.

See Also

with_filters() to choose inputs, with_defaults() to set arguments for every column.

Examples

filters <- shinyfilters(nyc_flights)

# A range slider: `.x` is the column the input is created for
filters <- with_args(filters, dep_delay ~ list(value = range(.x)))
filters$dep_delay

# Several arguments
with_args(filters, dep_delay ~ list(value = range(.x), step = 5))

# The same arguments for several columns
with_args(filters, c(dep_delay, distance) ~ list(value = range(.x)))

# Different arguments for different columns
with_args(
  filters,
  dep_delay ~ list(step = 5),
  origin ~ list(label = "Airport")
)

Description

with_defaults() adds, replaces, or removes the arguments that a configuration made by shinyfilters() passes to filterInput() for every column.

Usage

with_defaults(.filters, ...)

Arguments

Value

The updated configuration.

See Also

shinyfilters(), with_filters(), with_ns()

Examples

filters <- shinyfilters(nyc_flights, width = "200px")

# Turn a default off
filters <- with_defaults(filters, slider = FALSE)
filters

# Replace one
with_defaults(filters, width = "100%")

# Remove one
with_defaults(filters, width = NULL)

Description

with_filters() sets the input that filterInput() creates for one or more columns of a configuration made by shinyfilters(), and adds or replaces columns computed from the others. When a column's input is chosen more than once, the last one wins; the arguments set for the column stay with it.

Usage

with_filters(.filters, ...)

Arguments

Value

The updated configuration.

See Also

shinyfilters(), with_args()

Examples

filters <- shinyfilters(nyc_flights)
filters <- with_filters(filters, origin = "radio", carrier = "selectize")
filters

# Choose one input for several columns
filters <- with_filters(filters, dep_delay:origin ~ "slider")

# Give an input its arguments. `.x` is the column the input is for.
with_filters(filters, dep_delay ~ "slider" ~ list(value = range(.x)))

# Add a column computed from the others
with_filters(filters, delay_sq = dep_delay^2)

# Give one column the input it has by default
filters <- shinyfilters(nyc_flights, slider = TRUE)
with_filters(filters, distance = "numeric")

Description

with_ns() adds, replaces, or removes the namespace that a configuration made by shinyfilters() applies to its input ids. Use it to build a configuration once and reuse it in several modules.

Usage

with_ns(.filters, ns)

Arguments

Value

The updated configuration.

See Also

shinyfilters()

Examples

filters <- shinyfilters(nyc_flights)

# Add a namespace
filters <- with_ns(filters, "flights")
filters

# Replace it, here with a namespace created by NS()
with_ns(filters, shiny::NS("departures"))

# Remove it
with_ns(filters, NULL)