[Tex/LaTex] Best practice to include (standalone precompiled) graphics

best practicesgraphicspstricksstandalonetikz-pgf

General understanding

I know that several people have asked about how to include plots and graphics (created by tikz and pstricks) into a latex document.

Now it seems evident to ask how to handle a big number of figures.
At first it seems crucial to pre-compile graphics that should enter the document to reduce the final document's compilation time.

But there are many different ways to perform that issue and actually the package standalone seems the most auspicious.
By trying it, I recognized that all standalone tex-files have to be in the same folder to be compiled if changes were performed. Otherwise it would result in errors (Including pictures in subdirectories with standalone package).

Question

How do I now include graphics like the following MWE (adapted from matlab2tikz for including my macros), in order to have the same common formatting (e.g. font size and shape) for all text in my document including the figure labels?

MWE

\documentclass{article}
\input{Style/macros.tex}
\usepackage{standalone}
\standaloneconfig{mode=buildnew}

\usepackage{filecontents}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% File Style/macros.tex
\begin{filecontents*}{Style/macros.tex}
  \fontsize{10}{12}
\end{filecontents*}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% File Documents/test_standalone_slave.tex
\begin{filecontents*}{Documents/test_standalone_slave.tex}
  \documentclass{standalone}
  \input{../Style/macros.tex}
  \usepackage{tikz,pgfplots}
  \pgfplotsset{compat=newest}

  \begin{document}
  \begin{tikzpicture}
  \draw (0,0) circle (2);
  \end{tikzpicture}
  \end{document}
  \endinput
\end{filecontents*}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\usepackage{tikz,pgfplots}
\pgfplotsset{compat=newest}

\begin{document}
This is my MWE!
\includestandalone{Documents/test_standalone_slave}
\end{document}

Best Answer

If I were you, I would do the following:

In order to make your common settings accessible for all figures, diagrams, documents, etc, put the settings into a single package, namely, mycommon.sty. Register this package globally so you can use it throughout your projects. If you don't know how to do this, see my answer here (click).
If you want your diagram or figures to be accessible for other projects, put them in a separate directory that is higher than the project directories. For example,
```
<any directory>\Diagrams\
<any directory>\Diagrams\Projects\Project-01\
<any directory>\Diagrams\Projects\Project-02\
```

To get a tight drawing for each diagram, use standalone document class as follows

\documentclass[pstricks,border=12pt,12pt]{standalone}
\usepackage{mycommon}
\begin{document}
% your drawing code goes here...
\end{document}

Also load your package for common settings.

Compile each diagram with the appropriate compilers to get a PDF version. It will save a lot of your time when compiling the main input file of your projects.
Import the PDF diagrams from within the main input file of your projects. And compile the main input file with pdflatex (recommended because it is fully supported by microtype package). You may also load the mycommon package if necessary.

\includegraphics from graphicx will do the job of importing diagrams. You can set \graphicspath{{../../Diagrams/}} to shorten your path when invoking \includegraphics.

Related Solutions

[Tex/LaTex] How to organize large documents in small nested folders

I am not sure if i can answer the question as posted in a precise way, but here we go.

From the LaTeX point of view, the current working directory is ./ as a unix guy would say. In order to find the image, you have to go down into the contents directory (./contents/), then further down into chapter01 (./contents/chap01/) and finally change into the image folder (./contents/chap01/img/). This is a bit tediuos to write everytime, but there are thing like $TEXINPUTS or \graphicspath. This would be the pure LaTeX side.

Look at it from a different angle, changing directories in a shell can be a lot of fun, or pure hell. If you are working with a file manager, you are clicking your way through.

In my point of view (see, this is opinion based) the best practice is to have just unique filenames, not only for figures, but for all files; tex-files, image-files, pdf-files.

I was once collaborating with a human being, who saved all journal articles as is, meaning with a strange number. The person had quite a system, you had to go at least 4 levels down to see a pdf file (remember, all file names were random numbers). If you were looking for the article written by Jones in 2011, it could take you 15 minutes to find it, as you cannot know where exactly the person had put it.

What do i want to say with that?
If you take care in naming the filenames in a precise and unique way, you will not loose oversight of your files and your supervisor/studentes/colleuges won't go crazy.

You can have a database of a thousand pdf files, looking for jones2011 is not difficult. coffeemakerInit.jpg is different from coffeemakerExploded.jpg but without opening the image file, you know what it contains. Do the same with tex-files and especially Labels. The minute you may decide what a good label is is a minute saved later in searching for the label (or the image), or saving 15 minutes in searching for a journal article ;-)

Keep it simple ;-)

[Tex/LaTex] Drawing rectilinear curves in Tikz, aka an Etch-a-Sketch drawing

\draw (20,12) -- ++(2,0) -- ++(0,2) -- ++(-3,0) -- ++(45:3);

Use ++ before each new incremental coordinate to make it relative to the last one and put the pencil there.

Here's a complete example:

\documentclass{article}
\usepackage{tikz}
\begin{document}
\tikz\draw (20,12) -- ++(2,0) -- ++(0,2) -- ++(-3,0) -- ++(30:3) {[rounded corners=10pt]-- ++(5,0) -- ++(0,-6)} -- ++(-7,0) -- cycle;
\end{document}

Of course, combining this with the -| or |- path operators can simplify the code even further; the following two pieces of code produce the same result:

\tikz\draw (20,12) -- ++(2,0) -- ++(0,2) -- ++(3,0) -- ++(0,1) -- ++(1,0) -- ++(0,-3) -- ++(2,0);\par\bigskip

and

\tikz\draw (20,12) -| ++(2,2) -| ++(3,1) -- ++(1,0) |- ++(2,-3);

I don't think that defining commands in this case adds anything; in fact, I think it reduces the functionality of the existing syntax (which is already simple). The example demonstrates that you can use, for example, polar coordinates and modify (up to TikZ limitations) the path attributes midways; even if the current question doesn't require this, it's a good thing to have the possibility to do those kind of modification if they are required.