Skip to contents

Interactively Create a plot2

Source: R/shiny.R
create_interactively.Rd

This Shiny app allows for interactive creation of a plot2.

Usage

create_interactively(
  data = NULL,
  css_code = NULL,
  logo_path = system.file("logo.svg", package = "plot2"),
  pretty_labels = FALSE,
  hide_generated_code = FALSE,
  hide_export_buttons = NULL,
  upload_tab = FALSE,
  run = interactive()
)

Arguments

data

A data set to load, offered as 'Provided' in the data set selector. The generated code refers to it by its name, e.g. df |> plot2() for create_interactively(data = df). Not strictly required: in an interactive session, all data frames in the global environment are offered as well, together with the data sets of base R, plot2, ggplot2 and dplyr. Outside an interactive session (such as on a server), data frames in the global environment are not offered, so that the app does not expose what the hosting script has loaded.

css_code

Additional CSS code to load in the app.

logo_path

Path to the logo shown in the app. Default to the plot2 logo. Use NULL to not use a logo.

pretty_labels

A logical to switch to pretty, readable labels, instead of argument names in code style.

hide_generated_code

A logical to hide generated code.

hide_export_buttons

A logical to hide export buttons and functionality. TRUE will hide the elements completely, NULL will show a clickable text to expand buttons (default), FALSE will show the expanded buttons.

upload_tab

A logical to allow uploading data files. When TRUE, an Upload tab is added to the sidebar, with format-specific options for text (CSV, TSV, TXT), spreadsheet (Excel, OpenDocument), statistical software (SPSS, Stata, SAS, Minitab), Parquet, Feather, JSON, YAML, XML and HTML files; the data set selector then also offers "Upload data set...". R serialisation files (.rds, .rda) are not accepted, since reading them can run code: pass such data with the data argument instead. When FALSE (default), the app offers no way to upload files. Uploading requires the readxl package, and per format the package that reads it; administrators can install all of them with plot2:::install_shiny_deps().

run

A logical to run the app (default in an interactive session). Use run = FALSE to get the app object instead, for example to deploy it on a server, see Deployment.

Value

With run = TRUE, nothing (the app runs until it is closed). With run = FALSE, an app object of class shiny.appobj that can be run with shiny::runApp() or deployed.

Details

The app builds a plot2() call from the chosen settings, shows the plot and the code, and explains in a panel below the code what plot2() decided by itself.

Deployment

The app can run on a server for several users (e.g. Shiny Server or Posit Connect), with an app.R such as:

library(plot2)
create_interactively(run = FALSE)

On a server, the app treats everything a visitor types or sends as untrusted:

  • Settings are converted to values or checked expressions; nothing typed is run as R code. Calculations in the column selectors may only use a fixed list of computing functions (such as log(), mean() and n_distinct()), and the call is evaluated apart from the global environment, so it cannot reach other objects of the R process.

  • Data frames in the global environment are offered only in an interactive session; pass data with data to offer it on a server.

  • Uploads are only possible with upload_tab = TRUE; R serialisation files are never accepted. The maximum upload size is Shiny's own (option shiny.maxRequestSize, 5 MB by default).

  • Each session keeps its own plot, and raster exports are limited to 40 megapixels.

Each plot is computed in the R process of the app, so a visitor can still ask for plots that take long to compute, e.g. with very large data or a very small binwidth. Use the limits of the hosting platform (e.g. timeouts and the number of processes) as for any Shiny app.

Examples

if (FALSE) { # \dontrun{

create_interactively()

iris |> create_interactively()

# With the upload tab for importing external data files:
create_interactively(upload_tab = TRUE)

# The app object, e.g. for app.R on a server:
app <- create_interactively(run = FALSE)
} # }