HTML code inside of a R-markdown block for a single line

Okay, it is a year and a half later and I still needed this from time to time, but I learned enough since then to know how to do it properly. You need to write a knitr "output hook", and modify the output so the html can "escape" through.

The following accomplishes this:

• Added a knitr output hook.
• Defined a syntax to specify the needed tag and content
  • for example to get <h1>some_text</h1> use htmlesc<<(h1,some_text)>>
• Figured out a regexp that extracts the h1 and some_text and reformats it as properly tagged html, removing the ## that knitr inserted as well.
• Added some more test cases to make sure it did some additional things (like h4, p, proper placement of plots and tables, etc.)
• Added another regexp to remove double escaped lines which was adding some unwanted whitespace paneled constructs.

So here is the code:

---
title: "Output Hook for HTML Escape"
author: "Someone"
date: "2017 M04 25"
output: 
  html_document:
    keep_md: true
---

```{r setup, include=T,echo=TRUE}
knitr::opts_chunk$set(echo = TRUE)
hook_output <- knitr::knit_hooks$get("output")


knitr::knit_hooks$set(output=function(x,options){
  xn <- hook_output(x,options)
  # interestingly xn is a big character string. 
  # I had expected a list or vector, but the length of x is 1 and the class is character.

  # The following regexp extracts the parameters from a statement of the form
  # htmlesc<<(pat1,pat2)>> and converts it to <pat1>pat2</pat1>
  xn <- gsub("## htmlesc<<([^,]*),{1}([^>>]*)>>","\n```\n<\\1>\\2</\\1>\n```\n",xn)
  # now remove double escaped lines that occur when we do these right after each other
  gsub(">\n```\n\n\n```\n<",">\n<",xn)
}
)
```

## An analysis loop in a single R chunk with R Markdown

In the following, we do a loop and generate 3 sets of data:

(@) We have some explanitory text
(@) then we do a bar plot with ggplot
(@) then we print out a table
(@) then we do a base plot - just for fun

```{r, echo=T, fig.height=3,fig.width=5}
library(knitr)
library(tidyr)
library(ggplot2)
set.seed(123)

for (i in 1:3){
  mdf <- data.frame(matrix(runif(25),5,5))
  cat(sprintf("htmlesc<<h1,Title %d>>\n",i))
  cat(sprintf("htmlesc<<h4,Smaller Title - also for %d>>\n",i))
  cat(sprintf("htmlesc<<p,and some text talking about this %d example>>\n",i))
  print(sapply(mdf,mean))
  gdf <- gather(mdf,series,val)
  gp <- ggplot(gdf)+geom_bar(aes(series,val,fill=series,color=I("black")),stat="identity")
  print(gp)
  print(mdf)
  plot(mdf)
}
```

And this is the output (shrunk a bit as you don't need the details).

The only real docs for this by the way are Yihui's excellent knitr book, a search finds it easily.