{mojspeakr} is a package designed to address multiple aims:

• To allow easy conversion of an Rmarkdown file into a formatted govspeak file suitable for publishing through the Whitehall publisher on (GOV.UK).

• To allow output of statistical releases in both govspeak and another (word, PDF, HTML) format from one RMarkdown file.

• To allow replication of many govspeak tags and formatting appearances in the HTML output.

This package is available on github. To install, run the following code:

library(remotes)
remotes::install_github("moj-analytical-services/mojspeakr")

{mojspeakr} is designed to be used alongside an RMarkdown file, and cannot be called inside the RMarkdown file itself. The conversion acts on a markdown (*.md) file only, so R markdown (*.Rmd) should first be converted/knitted to *.md. This can be achieved through setting the argument keep_md to true in the YAML header of the RMarkdown document.

---
title: "My Rmarkdown File"
output: 
  html_document:
    keep_md: true
---

Dual output of govspeak and word/PDF can be achieved as follows:

---
title: "My Rmarkdown File"
output: 
  html_document:
    keep_md: true
  word_document:
    #any other parameters relevant to the word document here
---

It is also possible to keep the md document associated with the word output, and output a HTML format alongside this:

---
title: "My Rmarkdown File"
output: 
  word_document:
    keep_md: true
  html_document:
    #any other parameters relevant to the html document here
---

To render both outputs, make use of the rmarkdown::render() function with the argument outputs = "all".

Following rendering of the output, the convert_rmd() function modifies formatting of the .md file to produce a govspeak output which can be read by Whitehall Publishing:

• Image tags are converted from standard markdown to govspeak i.e.

Figure 1
![](images/image1.png)<!-- -->!

Figure 1
!!1

• Page breaks are replaced. This feature can be controlled using the "page_break" argument. page_break = "line" replaces them with a horizontal ruling, while page_break = "none" just moves the subsequent text onto a new line, and page_break = "unchanged" makes no replacement.

• YAML header is removed

• As GOV.UK cannot accept first-level headers (single #), all hashed headers are increased by one level i.e. # Title becomes ## Title. This functionality can be controlled using the sub_pattern argument. Setting it to TRUE increases all headers by one level. FALSE results in no changes to the headers as written. Passing the argument a vector allows selective changing of specific headers, e.g. sub_pattern = c("#", "##") would change only first and second level headers.

• Image names can be converted to a tidy standardised form of "image-X.png/svg". Passing a folder path for the image folder working directory (usually the same as the Rmarkdown working directory) converts the image names in that directory.

This function should be run in a separate utility script distinct from the Rmarkdown document. It produces two outputs; a govspeak format md file called x_converted.md, where x is the name of the original .md file, and a csv called img_lookup.csv, which provides a mapping of the expected image names to markdown image tags for QA purposes.

The following functions are all designed to be used inside Rmarkdown documents, and aim to make it simple to include images, tables, and other content permitted on gov.uk. All of these features will appear in the final govspeak output, and in some cases an approximation of their gov.uk appearance will be returned in other formats.

Allows simple addition of figures which are not generated within the Rmarkdown document. These must adhere to the Whitehall publishing size of 960 by 640 pixels if in .png format. Takes two arguments, the file name and its folder location (defaults to "graphs").

Formats a dataframe of data into a table. This will display correctly in govspeak and other (word and HTML) formats, and allows for text to go over multiple lines. Takes a single argument data of the table contents.

Aims to make publishing two formats from one RMarkdown document simple. This wrapper allows any output to be conditional and only appear in one of the specified formats e.g. to output table_one in the HTML output only you can use:

conditional_publishing_output("html", table_one)

This output will not appear at all in any non-HTML outputs such as word or pdf.

The full list of options that can be used in the first argument are:

• "html" for HTML/govspeak output
• "docx" for Word output
• "latex" for PDF output
• "markdown_strict" for markdown output
• "pptx" for powerpoint presentation
• "beamer" for beamer presentation
• "odt" for odt output

MoJ stats summary tables consist of three columns, one of which contains directional arrow images which are not compatible with publication through the Whitehall platform. Instead a 2-column table which omits these images and highlights key directional phrases and words is used. This function takes three vectors and conditionally produces either the two- or three- column table depending on the output type. A two-column table is produced in the govspeak output, and a three-column table is produced in all other output types.

Some publications use a callout box to add emphasis to text in an online publication via $CTA tags. This function automatically adds these $CTA tags to text in the govspeakr output only. If the HTML output is not used as the govspeakr output, CTA tags will appear in a grey box in HTML outputs.

Email addresses should appear as links in the govspeakr output only, and as plain text in all other formats. This function conditionally adds "<>" bracketing email addresses in the govspeakr output only. It can be used in inline code in Rmarkdown e.g.

Email: `r add_email([email protected])`

Postal addresses should appear as formatted text in govspeakr output only, and as plain text in all other formats. This function conditionally adds "$A" bracketing addresses in the govspeakr output only, and as a default also ensures that each part of the address is displayed on a new line in both outputs.

If you would like addresses to appear unformatted in all other formats, please add a call of "unmodified = T" to this function.

It accepts addresses as a single string, with each line separated by a comma e.g.

add_address("123 Fake Street, Leeds, LS1 2DA")

The following functions are all designed to be used inside Rmarkdown documents, and aim to make it simple to add aesthetic features such as logos and headers which are present by default on all gov.uk publications. These features will appear in the HTML output only, and should not be used in the govspeak output.

Adds a logo (e.g. MoJ logo) to the top of the HTML release, as per gov.uk. Takes a single argument, the path of the logo image.

Includes details of the statistical publication in a blue header, mimicking the appearance of gov.uk. Takes three arguments, the title of the release, the stats badging status (defaults to National Statistics), and the pre-release sub-header (defaults to "pre-release").