Add reportTheme argument to plotlyOutput() for reporting CSS styles

[cpsievert]

Adds the ability to access the new CSS styles reported in getCurrentOutputInfo() inside of a renderPlotly() context. For example:

library(shiny)
library(plotly)

ui <- fluidPage(
  tags$style(HTML("body {background-color: black; color: white}")),
  tags$style(HTML("body a {color: orange}")),
  plotlyOutput("p")
)

server <- function(input, output, session) {
  
  output$p <- renderPlotly({
    info <- getCurrentOutputInfo()
    plot_ly(x = 0:10, y = 0:10, color = I(info$accent())) %>%
      layout(
        font = list(color = info$fg()),
        paper_bgcolor = info$bg(),
        plot_bgcolor = info$bg(),
        xaxis = list(
          gridcolor = info$fg(),
          zerolinecolor = info$fg()
        ),
        yaxis = list(
          gridcolor = info$fg(),
          zerolinecolor = info$fg()
        )
      )
  })
  
}

shinyApp(ui, server)

This'll provide the foundation thematic and ggplotly() need to make this possible

library(shiny)
library(plotly)
library(thematic)

thematic_shiny()

ui <- fluidPage(
  titlePanel("Hello auto theming with ggplotly()"),
  tags$head(
    tags$style(HTML("body{background-color:#444444; color:#e4e4e4}")),
    tags$style(HTML("a{color:#e39777}"))
  ),
  plotlyOutput("p")
)

server <- function(input, output, session) {
  output$p <- renderPlotly({
    p <- ggplot(faithfuld, aes(waiting, eruptions, z = density)) + 
      geom_raster(aes(fill = density)) + 
      geom_contour()
    ggplotly(p)
  })
}

shinyApp(ui, server)