Table of Contents

Embedding APS datasets in other documents

AnalysisPageServer datasets can be embedded within other (HTML) documents.

Embedding within a knitr document

AnalysisPageServer datasets can be embedded within knitr documents such as the one you are reading.

To do so, first call setup.APS.knitr() to set things up. The main purpose of this call is to include some javascript and CSS that makes the interactivity work. See the documentation of that function for more details:


A few notes about setup.APS.knitr():

  1. By default this function makes a table of contents based on your header elements. Turn this off with include.toc=FALSE.
  2. It also includes some custom styling. This puts the table of contents in the margin on the left and the body of your document to the right of that. For example, this document was created with this default styling. You can control this behavior with the include.css argument.
  3. It is recommended to use message=FALSE, echo=FALSE for this setup block.

Whether you use default or custom styling, after your call to setup.APS.knitr() at the top of your document you can proceed as you normally would. At the point you want to insert an AnalysisPageServer dataset, make a call to embed.APS.dataset(). In typical usage the first argument is a code block wrapped in {curly brackets} that contains your plotting code. You also have to pass the data. This can either be the return value of the code block, or you can provide it explicity as the second argument:

x <- seq(0, 2*pi, length = 100)
y <- sin(x) + rnorm(100)/10
col <- adjustcolor(heat.colors(100), alpha.f = 0.6)
                    plot(x, y, col = col, pch = 19)
                  df = data.frame(x=x, y=y, X=x, Y=y),
                  title = "A shaky sine curve",
                  data.subdir = "embed-example")

You can have two datasets on the same page. In the next example we also hide the sidebar and data table with show.sidebar = FALSE and show.table = FALSE, and pass in a little extra styling to get some margins on the sides.

x <- rep(1:nrow(volcano), each = ncol(volcano))
y <- rep(1:ncol(volcano), nrow(volcano))
volcano.cells <- data.frame(x = x, y = y, Height = as.vector(t(volcano)))
                    image(volcano, xaxt = "n", yaxt = "n", main = "Maunga Whau Volcano")
                  df = volcano.cells,
                  title = "", data.subdir = "embed-example",
                  show.table = FALSE, show.sidebar = FALSE,
                  show.xy = TRUE,     
                  extra.div.attr = c(style = "width:60%; margin:0 auto"))

The <div> element to which the extra attributes, such as styling, are applied is the outer container that holds the plot, the sidebar, and the table. The current version does not support custom styling of the sub-elements.

Omit the first argument to just show the table:

embed.APS.dataset(df = iris,
                  title = "Iris data with no plot", data.subdir = "embed-example")

Embedding within any HTML document

In order to embed an AnalysisPageServer dataset into an HTML document you first have to copy the "front end" files, including CSS, Javascript other resources, into the directory where your HTML document will live. If you are going to write more than one HTML document to the same directory then you only have to do this once. The copy.front.end() function does this for you. For example,

basedir <- dirname(html.file)

Next you must include header lines in your HTML file to bring in the CSS, Javascript and other resources required by the front end. These are returned by this function call:

## <meta name="viewport" content="width=device-width, initial-scale=1.0">
## <link href="bundle-756ba12627.css" rel="stylesheet" type="text/css" />
## <script id="ep-entry-script" src="config-937f59c213.js"></script>
## <script id="ep-entry-script" src="bundle-ee64cfcdb9.js"></script>

Normally they would appear within the <head> section of your document.

After that, use the embed.APS.dataset() function as described above to make plots or datasets. Be careful to include outdir = basedir so it knows where to write the SVG and JSON files. Otherwise they would end up in the current directory, which might not work. Aside from writing those files, embed.APS.dataset returns a <div> each time you call it, and you can then lay these elements out as you like within your HTML document.

Here is an example of what that looks like:

## <div class="ep-analysis-page-data-set container-fluid"
##   data-sidebar-visible="yes"
##   data-table-visible="yes"
##   data-table-rows="10"
##   data-svg="embed-example/dataset-7-BzEF1v.svg"
##   data-set="embed-example/dataset-7-BzEF1v.json"></div>

The embed.APS.dataset() accepts a few arguments that allow you to control the inclusion of the table, of the sidebar, or to add extra HTML classes or attributes to the <div>.


Walk through interactive functionality on basic example