1   Introduction

Diagrams can be used to easily include pictures in your \(\LaTeX\) documents, using the diagrams-cairo, diagrams-postscript, or diagrams-pgf backends. You directly embed diagrams code in your \(\LaTeX\) source, within special \begin{diagram} ... \end{diagram} blocks; the diagrams-latex package then takes care of automatically running your diagrams code and including the generated images in the output, so you don't need to worry about maintaining a bunch of source files or running special preprocessing tools.

2   Installation

First, you will need the diagrams-builder package, which diagrams-latex depends on. diagrams-builder does not come automatically as part of a basic diagrams installation, so you will need to install it separately. You can install it with a command like

cabal install -fps diagrams-builder

Instead of -fps you may specify whichever backend you plan to use (-fps, -fcairo, or -fpgf). If you wish, you may specify more than one.

These flags (-fps, -fcairo, or -fpgf) cause executables to be installed (diagrams-builder-ps, diagrams-builder-cairo, and diagrams-builder-pgf respectively); make sure the appropriate executable(s) are installed and accessible via your PATH.

Finally, download diagrams-latex.sty and put it somewhere \(\LaTeX\) can find it. One such place is in the same directory as your .tex file. If you plan to use diagrams-latex often, you may prefer putting it in a more global location; if you don't know how to do that, a good place to start is this StackExchange question.

3   diagrams-latex package options

The first step is to include the diagrams-latex package. In the preamble of your \(\LaTeX\) file, add


Note inclusion of the graphicx package is currently required if you are using the PostScript or Cairo backends. You can also pass some options to the diagrams-latex package:

  • outputdir specifies the name of directory where intermediate and final diagram files will be written. The directory will be created if it does not already exist. By default, intermediate files will be written in the same directory as the \(\LaTeX\) document, which may not be desirable.

  • backend specifies the backend to be used. Possible values include cairo, ps, and pgf. The default is cairo.

  • extension specifies the file extension which should be used for compiled diagrams. The default is pdf. Be sure to use an extension which is compatible with the chosen backend; forexample, if you specify backend=ps you likely want to also specify extension=eps; for backend=pgf you likely want extension=pgf.

  • The input flag causes the generated diagrams to be included in the document via \input (instead of the default, which is \includegraphics). This option should be used with the PGF backend.

As a complete example, to specify diagrams compiled with the postscript backend, using the diagrams directory, one would write

\usepackage[backend=ps, extension=eps, outputdir=diagrams]{diagrams-latex}

To use the pgf backend, one might write

\usepackage[backend=pgf, extension=pgf, outputdir=diagrams, input]{diagrams-latex}

4   Diagram blocks

To include a diagram in your document, create a diagram block, like so:

  {-# LANGUAGE FlexibleContexts #-}
  import Data.List

  foo = circle 1 # fc green
  dia = foo ||| foo

The width and height arguments work as you would expect; they are both optional. Within the block, you can place an arbitrary Haskell module (including imports and LANGUAGE pragmas, as illustrated above). Note that the indentation does not matter, as long as it is consistent. Imports of Diagrams.Prelude and whichever backend you are using will be added automatically. The module must contain a definition of a diagram named dia, which will be rendered. It may also contain any other arbitrary helper definitions.

5   Running

To compile your en-diagrammed \(\LaTeX\) document, simply run latex, pdflatex, etc. as usual. However, you must also pass the --enable-write18 command-line flag. For example,

pdflatex --enable-write18 mydocument.tex

This is because in order to compile the embedded diagrams, \(\LaTeX\) must call out to the diagrams-builder executable via the shell, and this requires special permission.

6   Tips and tricks

Each of the diagram blocks is a completely independent scope. If you have definitions that should be shared among multiple diagrams (as is common), simply put the definitions in a separate Haskell module in the same directory, and import this module in each diagram block. diagrams-builder is smart enough to recompile a diagram when its imports have changed. In general, diagrams-builder tries hard to avoid recompiling diagrams when recompilation is not necessary, in order to speed build times.

If you use the PGF backend, any text diagrams will be typeset by \(\LaTeX\). In particular, if your text contains any sections surrounded by dollar signs, these will be typeset in math mode. For example,

  dia = text "hi $\\sqrt{\\pi}$" # scale 0.5 <> circle 1 # fc yellow

will produce a yellow circle with "hi \(\sqrt{\pi}\)" in the middle.

7   Using diagrams-latex with Beamer

diagrams-latex works well in conjunction with beamer for making slide decks with embedded images. However:

Every frame containing a diagram must be marked as [fragile]!

Non-[fragile] frames with diagrams in them cause \(\TeX\) to choke with horrendous, inscrutable errors. If you are using pandoc to generate slides, you can "trick" pandoc into emitting [fragile] annotations by inserting an empty code block into each slide with a diagram.

See this StackExchange answer explaining how to define an alternate frame environment which is fragile by default. It boils down to something like


and then using xframe instead of frame (it's probably possible to replace frame entirely with some clever renaming, but I haven't figured out how to do it).