Using Highlight With Pandoc
by Tristano Ajmone, MIT License (MIT). Last edited: Oct 26, 2017.
"Highlight.pp" v2.1.1 (2017-10-26)
“Highlight PP-Macros” is a set of macros to add cross-platform in-document access to Highlight within Markdown and reStructuredText source files via PP (pandoc preprocessor) and pandoc. It will allow the following operations from within source documents:
- Invoke Highlight on an external source code file, and inject the HTML output into the preprocessed document,
- Highlight an in-document code block via Highlight, instead of using pandoc’s built-in syntax highlighter,
- Import a Highlight theme as an inline CSS stylesheet.
CONTENTS
Introduction
The solution proposed here relies on three tools to achieve integration of Highlight in pandoc source files:
- Pandoc
- PP (a generic pandoc preprocessor)
- The
Highlight.pp
macros file found in this folder
Pandoc
Pandoc — aka “the universal markup converter” — is a well known command line tool for converting documents from one format to another, and especially from/to various markdown flavors, including “pandoc markdown” (an extended markdown variant introduced by pandoc itself). This very document was written in markdown and converted to HTML5 using pandoc and a custom template.
Pandoc is widely used for academic papers and technical documentation, and it ships with a built-in syntax highlighter (skylighting). By default, pandoc doesn’t support third party syntax highlighters, and the supported languages are hardcoded into pandoc and not extensible by end users without recompiling pandoc.1
Furthermore, pandoc doesn’t provide any means to import the source code from an external file — every time the source file changes, you need to manually copy-&-paste its contents again.
PP
Highlight PP-Macros leverage PP, a generic preprocessor “written with pandoc in mind.” PP adds to the pandoc workflow the power of user definable macros and a set of built-in macros for cross-platform file I/O operations and conditional branching — and many other great features not covered here.
Since one of PP’s goals was to introduce literate programming functionality in pandoc workflows, it provides a solid base for handling external files, invoking third party tools, and performing operations on source code blocks.
Highlight Macros
The file “Highlight.pp
” contains the definitions of the macros for integrating Highlight in pandoc workflow. It was taken from the “Pandoc Goodies” project, this file being one of the various macros modules that make up the PP-Macros Library subproject hosted therein:
How It Works
Currently, the Highlight PP-Macros module works only with HTML output. It works on all OS that support PP.
There are three macros available:
!HighlightFile(FILE)(LANG)[(OPTIONS)]
— imports and syntax-highlights an external file into the document as a raw HTML<pre><code>
block.!Highlight(LANG)([OPTIONS])(CODE BLOCK)
— syntax-highlights the block of source code passed asCODE BLOCK
parameter (using tilda fences instead of brackets; see Note below).!HighlightInlineTheme(THEME)
— retrieves a Highlight theme and injects its CSS version into the document.
The !HighlightFile
Macro
The !HighlightFile
macro takes two mandatory parameters and an optional third one:
FILE
— the name of the external source code file.LANG
— the language of the source code.OPTIONS
— additional command line options to be passed to Highlight.
The LANG
parameter is mandatory even if the source code file has an extension known to Highlight: the macro needs LANG
in order to write its value as the class of the enclosing <pre><code>
tags (these tags are written by the macro, not by Highlight). For example, !HighlightFile(somefile.py)(python)
will produce:
<pre class="hl python"><code class="python">
By default, all Highlight macros invokes Highlight with the following basic options:
--fragment
— omit document header and footer, get just the highlighted code.--syntax=<LANG>
— set language of source code toLANG
parameter.--no-trailing-nl
— omit trailing newline.--validate-input
— strip BOM (if present).
… followed by the literal string value of the OPTIONS
parameter (if any), which must consist of options compatible with the above invocation settings, and keeping in mind that Highlight’s output will be enclosed in <pre><code>
tags.
Since this macro outputs raw HTML blocks, pandoc’s built-it highlighter will ignore them. Furthermore, pandoc will still highlight any fenced (or backticked) source code block within the markdown document (ie: unless it’s explicitly invoked with the --no-highlight
options). This allows using Highlight side by side with pandoc’s highlighter, and even benefit from dual syntax color themes (this might require some CSS tweaking, or custom stylesheets).
The !Highlight
Macro
The !Highlight
macro takes two mandatory parameters and an optional third one:
LANG
— the language of the source code.OPTIONS
— command line options to be passed to Highlight.CODE BLOCK
— the source code to highlight.
The CODE BLOCK
parameter is passed as fenced code — ie: using two equal-lenght lines of tildas (or backticks) as delimiters, instead of brackets:
!Highlight( LANG )( OPTIONS )
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
CODE BLOCK
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
From PP’s documentation:
The last argument can be enclosed between lines of tildas or backquotes (of the same length) instead of parenthesis, brackets or braces. This is useful for literate programming, diagrams or scripts (see examples). Code block arguments are not stripped: spaces and blank lines are preserved.
The !Highlight
macro has been rewritten to work both under *nix shells and Window’s CMD — the macro detects the OS and PP-invocation context, and based on the findings it invokes the appropriate internal macro variant. (NOTE: Git Bash for Windows will trigger the shell variant.)
When invoked from CMD, each time the !Highlight
macro is called it will create a temporary file (named “_pp-tempfileX.tmp
”, where X
is a numeric counter) in the folder of the invoking script, to temporarily store the code to highlight. At each PP invocation the X
counter is reset, and the previous temp files are written over.
When run inside Shell/Bash (including Git Bash for Windows), !Highlight
doesn’t write any temporary files to disk.
If you wish to change the location where temporary files are written, set the PP_MACROS_PATH
environment variable to the path of the desired destination folder. Otherwise, you can add to your batch script a delete command to remove all temporary files created:
DEL _pp-tempfile*
The !HighlightInlineTheme
Macro
The macro !HighlightInlineTheme
offers a quick solution for theming Highlight code without having to first export a theme to CSS (via Highlight) and then import it back into the final document as an external CSS file (via pandoc): the macro imports a Highlight theme directly into the document, as an inline CSS stylesheet.
It takes a single mandatory parameter:
THEME
— the name of the Hihglight theme to import (without the “.theme
” extension).
It invokes Highlight with the following options:
--print-style
— print stylesheet only.--style=<THEME>
— select theTHEME
colour style (theme).--stdout
— print output result to STDOUT.
The macros encloses Highlight’s CSS output within <style type="text/css">
tags.
This macro can be invoked anywhere in the document and will apply to all Highlight code blocks (one theme per document). The imported theme will not affect code blocks highlighted by pandoc.
Example Files
build-example.bat
— The batch script to build the example file.build-example.sh
— The shell script to build the example file.example.md
— The example source file (markdown + PP macros).example-preprocessed.md
— The PP-preprocessed versionexample.md
.example.html
— The HTML conversion ofexample-preprocessed.md
via pandoc.example.pb
— The external source code file example (PureBASIC).Highlight.pp
— The required PP macros file.
Building The Examples
This files contains prebuilt version of the example files, but if you want to experiment with them here are the instructions.
Ensure that pandoc, PP and Highlight are available on the system PATH
, then execute the build-example
script (build-example.bat
in CMD, build-example.sh
in Shell/Bash).
The script issues two commands:
pp -import Highlight.pp example.md > example-preprocessed.md
pandoc --self-contained -S --normalize example-preprocessed.md -o example.html
The first line invokes pp
and tells it to -import
the contents of Highlight.pp
(so that the macros herein defined are loaded in the current environment) and then process the file example.md
. Since PP outputs to STDOUT
, we redirect (>
) its output to example-preprocessed.md
.
If you open example-preprocessed.md
in a text editor you’ll see how the macros inside example.md
were expanded.
The second line invokes pandoc
and tells it to convert example-preprocessed.md
and write the output (-o
) to the example.html
file (pandoc is smart enough to deduce the desired input and output formats from the files’ extensions). The --self-contained
option is used to produce a fully standalone HTML file with no external dependencies. The -S
(short for --smart
) and --normalize
options are to produce a nicer output text.
Enjoy…
Tristano Ajmone(Italy)
External Links
- http://pandoc.org
- http://cdsoft.fr/pp/
- https://github.com/tajmone/pandoc-goodies/tree/master/pp#highlight
Changelog
2017/10/26
- “
Highlight.pp
” v2.1.1 (PP v2.0):- Macros updated to work with PP v2.0 (released 2017/10/22).
- Added
!Highlight
macro to invoke Highlight on a code block inside the document.
2017/04/16
- “
Highlight.pp
” v1.2 (PP v1.7-2):- All macros updated to use the new cross-platfrom
!exec
built-in macro instead of!cmd
(requires PP 1.7-2, released the same day).
- All macros updated to use the new cross-platfrom
2017/04/16
( first release )
- “
Highlight.pp
” v1.1 (PP v1.5):!HighlightFile
macro (Win CMD only).!HighlightInlineTheme
macro (Win CMD only).
This is the state of things with pandoc v1; with the upcoming release of pandoc version 2 this will change (see #3334) as dynamic loading of syntax definitions will be supported via the new built-in highlighter. Even so, using Highlight with pandoc will still be possible and desirable because of the features offered by Highlight plugins, or because of syntax definitions not available for pandoc.↩