---
title: "Shiny Customized Widgets"
output: 
  rmarkdown::html_vignette
author: Zhengjia Wang
vignette: >
  %\VignetteIndexEntry{Shiny Customized Widgets}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE, echo=FALSE}

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

```


```{r setup, echo=FALSE}
library(shiny)
library(dipsaus)
```


### 1. Styled Action Button

The default `shiny` `actionButton` cannot fully use `Bootstrap` theme or the 
full features defined by `HTML`. For example, the button class is always 
`btn btn-default` and we can't disable/enable the button in an easy way.

`actionButtonStyled` solves these two problems:

* `type` allows you to add `Bootstrap` styles to the button. The supported types are: `default`, `primary`, `info`, `success`, `warning`, `danger`
* if you have customized CSS styles, `class` provides more flexible way to add classes
* when updating the button, `disabled` allows to disable/enable the button

The usage is listed as follows:

> Note the website must include `Bootstrap` library. If you are viewing this from default `CRAN` vignette, you might want to [check this link](https://dipterix.org/dipsaus/articles/shiny_customized_widgets.html) to view the demo.

```{r, eval=FALSE}

# UI function
actionButtonStyled(inputId, label, icon = NULL, width = NULL, 
                   btn_type = "button", type = "primary", class = "", ...)

# Update function
updateActionButtonStyled(session, inputId, label = NULL, icon = NULL,
                         type = NULL, disabled = NULL, ...)
```

Show-case:

```{r, results='asis', echo=FALSE}

btypes <- c('default', 'primary', 'info', 'success', 'warning', 'danger')
btypes <- rbind(btypes, stringr::str_to_title(btypes))

tags$table(
  class = "table table-bordered text-center",
  tags$tbody(
    tags$tr(tagList(
      apply(btypes, 2, function(x){
        tags$th(x[2], br(), tags$small(paste0('type="', x[1], '"')))
      }),
      tags$th('')
    )),
    tags$tr(
      apply(btypes, 2, function(x){
        tags$td(actionButtonStyled('btn', x[2], type = x[1], btn_type = 'a', width = '100%'))
      }),
      tags$td('[default]')
    ),
    tags$tr(
      apply(btypes, 2, function(x){
        tags$td(actionButtonStyled('btn', x[2], type = x[1], btn_type = 'a', 
                                   width = '100%', class = 'btn-lg'))
      }),
      tags$td('class="btn-lg"')
    ),
    tags$tr(
      apply(btypes, 2, function(x){
        tags$td(actionButtonStyled('btn', x[2], type = x[1], btn_type = 'a', 
                                   width = '100%', class = 'btn-sm'))
      }),
      tags$td('class="btn-sm"')
    ),
    tags$tr(
      apply(btypes, 2, function(x){
        tags$td(actionButtonStyled('btn', x[2], type = x[1], btn_type = 'a', 
                                   width = '100%', class = 'btn-xs'))
      }),
      tags$td('class="btn-xs"')
    ),
    tags$tr(
      apply(btypes, 2, function(x){
        tags$td(actionButtonStyled('btn', x[2], type = x[1], btn_type = 'a', 
                                   width = '100%', disabled=TRUE))
      }),
      tags$td('disabled=TRUE')
    )
  )
)


```

### 2. Compound Inputs (grouped Inputs)

`compoundInput2` provides group inputs where each group contains multiple shiny 
inputs. For examples

```{r, eval=FALSE}
compoundInput2(
  'compound', 'Group Label', label_color = 1:10,
  components = div(
    textInput('txt', 'Text'),
    selectInput('sel', 'Select', choices = 1:10, multiple = TRUE),
    sliderInput('sli', 'Slider', max=1, min=0, val=0.5)
  ), max_ncomp = 10, min_ncomp = 0, initial_ncomp = 1
)
```

will create a list of input groups with minimum of 0 but maximum of 10. User can control the size of groups by pressing `+` and `-` buttons. The value `input$components` looks like this:

```
#> [[1]]
#> [[1]]$txt
#> [1] ""
#> 
#> [[1]]$sel
#> [1] "1" "3" 
#> 
#> [[1]]$sli
#> [1] 0.5
#> 
#> [[2]]
...
```

I found this input extremely useful in clinic trial when the experiment condition is grouped and developers don't know ahead the number of condition groups.

The further details can be found with `demo('example-compountInput2', package='dipsaus')`. The source file can be found using `system.file('demo/example-compountInput2.R', package='dipsaus')`.


### 3. Synchronize Multiple Inputs

`sync_shiny_inputs` provides a way to update among multiple inputs with no dead-locks. For example, input A (`textInput`) shares the same information as input B (`sliderInput`). Updating A would trigger B to update. What if you want A to be updated when B is changed? The following code might cause recursive updates:

```{r, eval = FALSE}
# Bad example
observeEvent(input$A, {
  updateSliderInput(session, 'B', value = input$A)
})
observeEvent(input$B, {
  updateTextInput(session, 'A', value = input$B)
})
```

In this case, you can use `sync_shiny_inputs`:

```{r, eval = FALSE}
sync_shiny_inputs(input, session, inputIds = c('A', 'B'), uniform = list(
  function(a){as.numeric(a)},
  function(b){ b }
), updates = list(
  function(a){updateTextInput(session, 'A', value = a)},
  function(b){updateSliderInput(session, 'B', value = b)}
))
```

`inputIds` refers to the input ID to be synchronized. When one or more of the inputs are changed, the value will be passed to the corresponding `uniform` functions and stored, then `updates` will notify each inputs to update the UI values. The input of `updates` is the stored value.

For example, if input `A` is changed from `"0"` to `"1"`, then the first function of `uniform` is triggered, `function(a){as.numeric(a)}` will be evaluated with `a="1"`. The result, which is numeric `1` will be stored. Next, each functions in `updates` will be called, with `1` (stored in the previous step) as input, results in changing the slider input `B` to `1`. This whole process will not trigger `A` to re-update.