| 1 |
% Copyright 2019 by Till Tantau |
| 2 |
% |
| 3 |
% This file may be distributed and/or modified |
| 4 |
% |
| 5 |
% 1. under the LaTeX Project Public License and/or |
| 6 |
% 2. under the GNU Free Documentation License. |
| 7 |
% |
| 8 |
% See the file doc/generic/pgf/licenses/LICENSE for more details. |
| 9 |
|
| 10 |
|
| 11 |
\section{Tutorial: A Picture for Karl's Students} |
| 12 |
|
| 13 |
This tutorial is intended for new users of \tikzname. It does not give an |
| 14 |
exhaustive account of all the features of \tikzname, just of those that you are |
| 15 |
likely to use right away. |
| 16 |
|
| 17 |
Karl is a math and chemistry high-school teacher. He used to create the |
| 18 |
graphics in his worksheets and exams using \LaTeX's |{picture}| environment. |
| 19 |
While the results were acceptable, creating the graphics often turned out to be |
| 20 |
a lengthy process. Also, there tended to be problems with lines having slightly |
| 21 |
wrong angles and circles also seemed to be hard to get right. Naturally, his |
| 22 |
students could not care less whether the lines had the exact right angles and |
| 23 |
they find Karl's exams too difficult no matter how nicely they were drawn. But |
| 24 |
Karl was never entirely satisfied with the result. |
| 25 |
|
| 26 |
Karl's son, who was even less satisfied with the results (he did not have to |
| 27 |
take the exams, after all), told Karl that he might wish to try out a new |
| 28 |
package for creating graphics. A bit confusingly, this package seems to have |
| 29 |
two names: First, Karl had to download and install a package called \pgfname. |
| 30 |
Then it turns out that inside this package there is another package called |
| 31 |
\tikzname, which is supposed to stand for ``\tikzname\ ist \emph{kein} |
| 32 |
Zeichenprogramm''. Karl finds this all a bit strange and \tikzname\ seems to |
| 33 |
indicate that the package does not do what he needs. However, having used |
| 34 |
\textsc{gnu} software for quite some time and ``\textsc{gnu} not being Unix'', |
| 35 |
there seems to be hope yet. His son assures him that \tikzname's name is |
| 36 |
intended to warn people that \tikzname\ is not a program that you can use to |
| 37 |
draw graphics with your mouse or tablet. Rather, it is more like a ``graphics |
| 38 |
language''. |
| 39 |
|
| 40 |
|
| 41 |
\subsection{Problem Statement} |
| 42 |
|
| 43 |
Karl wants to put a graphic on the next worksheet for his students. He is |
| 44 |
currently teaching his students about sine and cosine. What he would like to |
| 45 |
have is something that looks like this (ideally): |
| 46 |
% |
| 47 |
\noindent |
| 48 |
\begin{tikzpicture} |
| 49 |
[scale=3,line cap=round, |
| 50 |
% Styles |
| 51 |
axes/.style=, |
| 52 |
important line/.style={very thick}, |
| 53 |
information text/.style={rounded corners,fill=red!10,inner sep=1ex}] |
| 54 |
|
| 55 |
% Local definitions |
| 56 |
\def\costhirty{0.8660256} |
| 57 |
|
| 58 |
% Colors |
| 59 |
\colorlet{anglecolor}{green!50!black} |
| 60 |
\colorlet{sincolor}{red} |
| 61 |
\colorlet{tancolor}{orange!80!black} |
| 62 |
\colorlet{coscolor}{blue} |
| 63 |
|
| 64 |
% The graphic |
| 65 |
\draw[help lines,step=0.5cm] (-1.4,-1.4) grid (1.4,1.4); |
| 66 |
|
| 67 |
\draw (0,0) circle [radius=1cm]; |
| 68 |
|
| 69 |
\begin{scope}[axes] |
| 70 |
\draw[->] (-1.5,0) -- (1.5,0) node[right] {$x$}; |
| 71 |
\draw[->] (0,-1.5) -- (0,1.5) node[above] {$y$}; |
| 72 |
|
| 73 |
\foreach \x/\xtext in {-1, -.5/-\frac{1}{2}, 1} |
| 74 |
\draw[xshift=\x cm] (0pt,1pt) -- (0pt,-1pt) node[below,fill=white] {$\xtext$}; |
| 75 |
|
| 76 |
\foreach \y/\ytext in {-1, -.5/-\frac{1}{2}, .5/\frac{1}{2}, 1} |
| 77 |
\draw[yshift=\y cm] (1pt,0pt) -- (-1pt,0pt) node[left,fill=white] {$\ytext$}; |
| 78 |
\end{scope} |
| 79 |
|
| 80 |
\filldraw[fill=green!20,draw=anglecolor] (0,0) -- (3mm,0pt) arc(0:30:3mm); |
| 81 |
\draw (15:2mm) node[anglecolor] {$\alpha$}; |
| 82 |
|
| 83 |
\draw[important line,sincolor] |
| 84 |
(30:1cm) -- node[left=1pt,fill=white] {$\sin \alpha$} +(0,-.5); |
| 85 |
|
| 86 |
\draw[important line,coscolor] |
| 87 |
(0,0) -- node[below=2pt,fill=white] {$\cos \alpha$} (\costhirty,0); |
| 88 |
|
| 89 |
\draw[important line,tancolor] (1,0) -- |
| 90 |
node [right=1pt,fill=white] |
| 91 |
{ |
| 92 |
$\displaystyle \tan \alpha \color{black}= |
| 93 |
\frac{{\color{sincolor}\sin \alpha}}{\color{coscolor}\cos \alpha}$ |
| 94 |
} (intersection of 0,0--30:1cm and 1,0--1,1) coordinate (t); |
| 95 |
|
| 96 |
\draw (0,0) -- (t); |
| 97 |
|
| 98 |
\draw[xshift=1.85cm] node [right,text width=6cm,information text] |
| 99 |
{ |
| 100 |
The {\color{anglecolor} angle $\alpha$} is $30^\circ$ in the |
| 101 |
example ($\pi/6$ in radians). The {\color{sincolor}sine of |
| 102 |
$\alpha$}, which is the height of the red line, is |
| 103 |
\[ |
| 104 |
{\color{sincolor} \sin \alpha} = 1/2. |
| 105 |
\] |
| 106 |
By the Theorem of Pythagoras we have ${\color{coscolor}\cos^2 \alpha} + |
| 107 |
{\color{sincolor}\sin^2\alpha} =1$. Thus the length of the blue |
| 108 |
line, which is the {\color{coscolor}cosine of $\alpha$}, must be |
| 109 |
\[ |
| 110 |
{\color{coscolor}\cos\alpha} = \sqrt{1 - 1/4} = \textstyle |
| 111 |
\frac{1}{2} \sqrt 3. |
| 112 |
\]% |
| 113 |
This shows that {\color{tancolor}$\tan \alpha$}, which is the |
| 114 |
height of the orange line, is |
| 115 |
\[ |
| 116 |
{\color{tancolor}\tan\alpha} = \frac{{\color{sincolor}\sin |
| 117 |
\alpha}}{\color{coscolor}\cos \alpha} = 1/\sqrt 3. |
| 118 |
\]% |
| 119 |
}; |
| 120 |
\end{tikzpicture} |
| 121 |
|
| 122 |
|
| 123 |
\subsection{Setting up the Environment} |
| 124 |
|
| 125 |
In \tikzname, to draw a picture, at the start of the picture you need to tell |
| 126 |
\TeX\ or \LaTeX\ that you want to start a picture. In \LaTeX\ this is done |
| 127 |
using the environment |{tikzpicture}|, in plain \TeX\ you just use |
| 128 |
|\tikzpicture| to start the picture and |\endtikzpicture| to end it. |
| 129 |
|
| 130 |
|
| 131 |
\subsubsection{Setting up the Environment in \LaTeX} |
| 132 |
|
| 133 |
Karl, being a \LaTeX\ user, thus sets up his file as follows: |
| 134 |
% |
| 135 |
\begin{codeexample}[code only] |
| 136 |
\documentclass{article} % say |
| 137 |
\usepackage{tikz} |
| 138 |
\begin{document} |
| 139 |
We are working on |
| 140 |
\begin{tikzpicture} |
| 141 |
\draw (-1.5,0) -- (1.5,0); |
| 142 |
\draw (0,-1.5) -- (0,1.5); |
| 143 |
\end{tikzpicture}. |
| 144 |
\end{document} |
| 145 |
\end{codeexample} |
| 146 |
|
| 147 |
When executed, that is, run via |pdflatex| or via |latex| followed by |dvips|, |
| 148 |
the resulting will contain something that looks like this: |
| 149 |
% |
| 150 |
\begin{codeexample}[width=7cm] |
| 151 |
We are working on |
| 152 |
\begin{tikzpicture} |
| 153 |
\draw (-1.5,0) -- (1.5,0); |
| 154 |
\draw (0,-1.5) -- (0,1.5); |
| 155 |
\end{tikzpicture}. |
| 156 |
\end{codeexample} |
| 157 |
|
| 158 |
Admittedly, not quite the whole picture, yet, but we do have the axes |
| 159 |
established. Well, not quite, but we have the lines that make up the axes |
| 160 |
drawn. Karl suddenly has a sinking feeling that the picture is still some way |
| 161 |
off. |
| 162 |
|
| 163 |
Let's have a more detailed look at the code. First, the package |tikz| is |
| 164 |
loaded. This package is a so-called ``frontend'' to the basic \pgfname\ system. |
| 165 |
The basic layer, which is also described in this manual, is somewhat more, |
| 166 |
well, basic and thus harder to use. The frontend makes things easier by |
| 167 |
providing a simpler syntax. |
| 168 |
|
| 169 |
Inside the environment there are two |\draw| commands. They mean: ``The path, |
| 170 |
which is specified following the command up to the semicolon, should be |
| 171 |
drawn.'' The first path is specified as |(-1.5,0) -- (1.5,0)|, which means ``a |
| 172 |
straight line from the point at position $(-1.5,0)$ to the point at position |
| 173 |
$(1.5,0)$''. Here, the positions are specified within a special coordinate |
| 174 |
system in which, initially, one unit is 1cm. |
| 175 |
|
| 176 |
Karl is quite pleased to note that the environment automatically reserves |
| 177 |
enough space to encompass the picture. |
| 178 |
|
| 179 |
|
| 180 |
\subsubsection{Setting up the Environment in Plain \TeX} |
| 181 |
|
| 182 |
Karl's wife Gerda, who also happens to be a math teacher, is not a \LaTeX\ |
| 183 |
user, but uses plain \TeX\ since she prefers to do things ``the old way''. She |
| 184 |
can also use \tikzname. Instead of |\usepackage{tikz}| she has to write |
| 185 |
|\input tikz.tex| and instead of |\begin{tikzpicture}| she writes |
| 186 |
|\tikzpicture| and instead of |\end{tikzpicture}| she writes |\endtikzpicture|. |
| 187 |
|
| 188 |
Thus, she would use: |
| 189 |
% |
| 190 |
\begin{codeexample}[code only] |
| 191 |
%% Plain TeX file |
| 192 |
\input tikz.tex |
| 193 |
\baselineskip=12pt |
| 194 |
\hsize=6.3truein |
| 195 |
\vsize=8.7truein |
| 196 |
We are working on |
| 197 |
\tikzpicture |
| 198 |
\draw (-1.5,0) -- (1.5,0); |
| 199 |
\draw (0,-1.5) -- (0,1.5); |
| 200 |
\endtikzpicture. |
| 201 |
\bye |
| 202 |
\end{codeexample} |
| 203 |
|
| 204 |
Gerda can typeset this file using either |pdftex| or |tex| together with |
| 205 |
|dvips|. \tikzname\ will automatically discern which driver she is using. If |
| 206 |
she wishes to use |dvipdfm| together with |tex|, she either needs to modify the |
| 207 |
file |pgf.cfg| or can write |\def\pgfsysdriver{pgfsys-dvipdfm.def}| somewhere |
| 208 |
\emph{before} she inputs |tikz.tex| or |pgf.tex|. |
| 209 |
|
| 210 |
|
| 211 |
\subsubsection{Setting up the Environment in Con\TeX t} |
| 212 |
|
| 213 |
Karl's uncle Hans uses Con\TeX t. Like Gerda, Hans can also use \tikzname. |
| 214 |
Instead of |\usepackage{tikz}| he says |\usemodule[tikz]|. Instead of |
| 215 |
|\begin{tikzpicture}| he writes |\starttikzpicture| and instead of |
| 216 |
|\end{tikzpicture}| he writes |\stoptikzpicture|. |
| 217 |
|
| 218 |
His version of the example looks like this: |
| 219 |
% |
| 220 |
\begin{codeexample}[code only] |
| 221 |
%% ConTeXt file |
| 222 |
\usemodule[tikz] |
| 223 |
|
| 224 |
\starttext |
| 225 |
We are working on |
| 226 |
\starttikzpicture |
| 227 |
\draw (-1.5,0) -- (1.5,0); |
| 228 |
\draw (0,-1.5) -- (0,1.5); |
| 229 |
\stoptikzpicture. |
| 230 |
\stoptext |
| 231 |
\end{codeexample} |
| 232 |
|
| 233 |
Hans will now typeset this file in the usual way using |texexec| or |context|. |
| 234 |
|
| 235 |
|
| 236 |
\subsection{Straight Path Construction} |
| 237 |
|
| 238 |
The basic building block of all pictures in \tikzname\ is the path. A |
| 239 |
\emph{path} is a series of straight lines and curves that are connected (that |
| 240 |
is not the whole picture, but let us ignore the complications for the moment). |
| 241 |
You start a path by specifying the coordinates of the start position as a point |
| 242 |
in round brackets, as in |(0,0)|. This is followed by a series of ``path |
| 243 |
extension operations''. The simplest is |--|, which we used already. It must be |
| 244 |
followed by another coordinate and it extends the path in a straight line to |
| 245 |
this new position. For example, if we were to turn the two paths of the axes |
| 246 |
into one path, the following would result: |
| 247 |
% |
| 248 |
\begin{codeexample}[] |
| 249 |
\tikz \draw (-1.5,0) -- (1.5,0) -- (0,-1.5) -- (0,1.5); |
| 250 |
\end{codeexample} |
| 251 |
|
| 252 |
Karl is a bit confused by the fact that there is no |{tikzpicture}| |
| 253 |
environment, here. Instead, the little command |\tikz| is used. This command |
| 254 |
either takes one argument (starting with an opening brace as in |
| 255 |
|\tikz{\draw (0,0) -- (1.5,0)}|, which yields \tikz{\draw (0,0) --(1.5,0);}) or |
| 256 |
collects everything up to the next semicolon and puts it inside a |
| 257 |
|{tikzpicture}| environment. As a rule of thumb, all \tikzname\ graphic drawing |
| 258 |
commands must occur as an argument of |\tikz| or inside a |{tikzpicture}| |
| 259 |
environment. Fortunately, the command |\draw| will only be defined inside this |
| 260 |
environment, so there is little chance that you will accidentally do something |
| 261 |
wrong here. |
| 262 |
|
| 263 |
|
| 264 |
\subsection{Curved Path Construction} |
| 265 |
|
| 266 |
The next thing Karl wants to do is to draw the circle. For this, straight lines |
| 267 |
obviously will not do. Instead, we need some way to draw curves. For this, |
| 268 |
\tikzname\ provides a special syntax. One or two ``control points'' are needed. |
| 269 |
The math behind them is not quite trivial, but here is the basic idea: Suppose |
| 270 |
you are at point $x$ and the first control point is $y$. Then the curve will |
| 271 |
start ``going in the direction of~$y$ at~$x$'', that is, the tangent of the |
| 272 |
curve at $x$ will point toward~$y$. Next, suppose the curve should end at $z$ |
| 273 |
and the second support point is $w$. Then the curve will, indeed, end at $z$ |
| 274 |
and the tangent of the curve at point $z$ will go through $w$. |
| 275 |
|
| 276 |
Here is an example (the control points have been added for clarity): |
| 277 |
% |
| 278 |
\begin{codeexample}[] |
| 279 |
\begin{tikzpicture} |
| 280 |
\filldraw [gray] (0,0) circle [radius=2pt] |
| 281 |
(1,1) circle [radius=2pt] |
| 282 |
(2,1) circle [radius=2pt] |
| 283 |
(2,0) circle [radius=2pt]; |
| 284 |
\draw (0,0) .. controls (1,1) and (2,1) .. (2,0); |
| 285 |
\end{tikzpicture} |
| 286 |
\end{codeexample} |
| 287 |
|
| 288 |
The general syntax for extending a path in a ``curved'' way is |.. controls| |
| 289 |
\meta{first control point} |and| \meta{second control point} |..| |
| 290 |
\meta{end point}. You can leave out the |and| \meta{second control point}, |
| 291 |
which causes the first one to be used twice. |
| 292 |
|
| 293 |
So, Karl can now add the first half circle to the picture: |
| 294 |
% |
| 295 |
\begin{codeexample}[] |
| 296 |
\begin{tikzpicture} |
| 297 |
\draw (-1.5,0) -- (1.5,0); |
| 298 |
\draw (0,-1.5) -- (0,1.5); |
| 299 |
\draw (-1,0) .. controls (-1,0.555) and (-0.555,1) .. (0,1) |
| 300 |
.. controls (0.555,1) and (1,0.555) .. (1,0); |
| 301 |
\end{tikzpicture} |
| 302 |
\end{codeexample} |
| 303 |
|
| 304 |
Karl is happy with the result, but finds specifying circles in this way to be |
| 305 |
extremely awkward. Fortunately, there is a much simpler way. |
| 306 |
|
| 307 |
|
| 308 |
\subsection{Circle Path Construction} |
| 309 |
|
| 310 |
In order to draw a circle, the path construction operation |circle| can be |
| 311 |
used. This operation is followed by a radius in brackets as in the following |
| 312 |
example: (Note that the previous position is used as the \emph{center} of the |
| 313 |
circle.) |
| 314 |
% |
| 315 |
\begin{codeexample}[] |
| 316 |
\tikz \draw (0,0) circle [radius=10pt]; |
| 317 |
\end{codeexample} |
| 318 |
|
| 319 |
You can also append an ellipse to the path using the |ellipse| operation. |
| 320 |
Instead of a single radius you can specify two of them: |
| 321 |
% |
| 322 |
\begin{codeexample}[] |
| 323 |
\tikz \draw (0,0) ellipse [x radius=20pt, y radius=10pt]; |
| 324 |
\end{codeexample} |
| 325 |
|
| 326 |
To draw an ellipse whose axes are not horizontal and vertical, but point in an |
| 327 |
arbitrary direction (a ``turned ellipse'' like \tikz \draw[rotate=30] (0,0) |
| 328 |
ellipse [x radius=6pt, y radius=3pt];) you can use transformations, which are |
| 329 |
explained later. The code for the little ellipse is |
| 330 |
|\tikz \draw[rotate=30] (0,0) ellipse [x radius=6pt, y radius=3pt];|, by the |
| 331 |
way. |
| 332 |
|
| 333 |
So, returning to Karl's problem, he can write |
| 334 |
|\draw (0,0) circle [radius=1cm];| to draw the circle: |
| 335 |
% |
| 336 |
\begin{codeexample}[] |
| 337 |
\begin{tikzpicture} |
| 338 |
\draw (-1.5,0) -- (1.5,0); |
| 339 |
\draw (0,-1.5) -- (0,1.5); |
| 340 |
\draw (0,0) circle [radius=1cm]; |
| 341 |
\end{tikzpicture} |
| 342 |
\end{codeexample} |
| 343 |
|
| 344 |
At this point, Karl is a bit alarmed that the circle is so small when he wants |
| 345 |
the final picture to be much bigger. He is pleased to learn that \tikzname\ has |
| 346 |
powerful transformation options and scaling everything by a factor of three is |
| 347 |
very easy. But let us leave the size as it is for the moment to save some |
| 348 |
space. |
| 349 |
|
| 350 |
|
| 351 |
\subsection{Rectangle Path Construction} |
| 352 |
|
| 353 |
The next things we would like to have is the grid in the background. There are |
| 354 |
several ways to produce it. For example, one might draw lots of rectangles. |
| 355 |
Since rectangles are so common, there is a special syntax for them: To add a |
| 356 |
rectangle to the current path, use the |rectangle| path construction operation. |
| 357 |
This operation should be followed by another coordinate and will append a |
| 358 |
rectangle to the path such that the previous coordinate and the next |
| 359 |
coordinates are corners of the rectangle. So, let us add two rectangles to the |
| 360 |
picture: |
| 361 |
% |
| 362 |
\begin{codeexample}[] |
| 363 |
\begin{tikzpicture} |
| 364 |
\draw (-1.5,0) -- (1.5,0); |
| 365 |
\draw (0,-1.5) -- (0,1.5); |
| 366 |
\draw (0,0) circle [radius=1cm]; |
| 367 |
\draw (0,0) rectangle (0.5,0.5); |
| 368 |
\draw (-0.5,-0.5) rectangle (-1,-1); |
| 369 |
\end{tikzpicture} |
| 370 |
\end{codeexample} |
| 371 |
|
| 372 |
While this may be nice in other situations, this is not really leading anywhere |
| 373 |
with Karl's problem: First, we would need an awful lot of these rectangles and |
| 374 |
then there is the border that is not ``closed''. |
| 375 |
|
| 376 |
So, Karl is about to resort to simply drawing four vertical and four horizontal |
| 377 |
lines using the nice |\draw| command, when he learns that there is a |grid| |
| 378 |
path construction operation. |
| 379 |
|
| 380 |
|
| 381 |
\subsection{Grid Path Construction} |
| 382 |
|
| 383 |
The |grid| path operation adds a grid to the current path. It will add lines |
| 384 |
making up a grid that fills the rectangle whose one corner is the current point |
| 385 |
and whose other corner is the point following the |grid| operation. For |
| 386 |
example, the code |\tikz \draw[step=2pt] (0,0) grid (10pt,10pt);| produces |
| 387 |
\tikz \draw[step=2pt] (0,0) grid (10pt,10pt);. Note how the optional argument |
| 388 |
for |\draw| can be used to specify a grid width (there are also |xstep| and |
| 389 |
|ystep| to define the steppings independently). As Karl will learn soon, there |
| 390 |
are \emph{lots} of things that can be influenced using such options. |
| 391 |
|
| 392 |
For Karl, the following code could be used: |
| 393 |
% |
| 394 |
\begin{codeexample}[] |
| 395 |
\begin{tikzpicture} |
| 396 |
\draw (-1.5,0) -- (1.5,0); |
| 397 |
\draw (0,-1.5) -- (0,1.5); |
| 398 |
\draw (0,0) circle [radius=1cm]; |
| 399 |
\draw[step=.5cm] (-1.4,-1.4) grid (1.4,1.4); |
| 400 |
\end{tikzpicture} |
| 401 |
\end{codeexample} |
| 402 |
|
| 403 |
Having another look at the desired picture, Karl notices that it would be nice |
| 404 |
for the grid to be more subdued. (His son told him that grids tend to be |
| 405 |
distracting if they are not subdued.) To subdue the grid, Karl adds two more |
| 406 |
options to the |\draw| command that draws the grid. First, he uses the color |
| 407 |
|gray| for the grid lines. Second, he reduces the line width to |very thin|. |
| 408 |
Finally, he swaps the ordering of the commands so that the grid is drawn first |
| 409 |
and everything else on top. |
| 410 |
% |
| 411 |
\begin{codeexample}[] |
| 412 |
\begin{tikzpicture} |
| 413 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 414 |
\draw (-1.5,0) -- (1.5,0); |
| 415 |
\draw (0,-1.5) -- (0,1.5); |
| 416 |
\draw (0,0) circle [radius=1cm]; |
| 417 |
\end{tikzpicture} |
| 418 |
\end{codeexample} |
| 419 |
|
| 420 |
|
| 421 |
\subsection{Adding a Touch of Style} |
| 422 |
|
| 423 |
Instead of the options |gray,very thin| Karl could also have said |help lines|. |
| 424 |
\emph{Styles} are predefined sets of options that can be used to organize how a |
| 425 |
graphic is drawn. By saying |help lines| you say ``use the style that I (or |
| 426 |
someone else) has set for drawing help lines''. If Karl decides, at some later |
| 427 |
point, that grids should be drawn, say, using the color |blue!50| instead of |
| 428 |
|gray|, he could provide the following option somewhere: |
| 429 |
% |
| 430 |
\begin{codeexample}[code only] |
| 431 |
help lines/.style={color=blue!50,very thin} |
| 432 |
\end{codeexample} |
| 433 |
% |
| 434 |
The effect of this ``style setter'' is that in the current scope or environment |
| 435 |
the |help lines| option has the same effect as |color=blue!50,very thin|. |
| 436 |
|
| 437 |
Using styles makes your graphics code more flexible. You can change the way |
| 438 |
things look easily in a consistent manner. Normally, styles are defined at the |
| 439 |
beginning of a picture. However, you may sometimes wish to define a style |
| 440 |
globally, so that all pictures of your document can use this style. Then you |
| 441 |
can easily change the way all graphics look by changing this one style. In this |
| 442 |
situation you can use the |\tikzset| command at the beginning of the document |
| 443 |
as in |
| 444 |
% |
| 445 |
\begin{codeexample}[code only] |
| 446 |
\tikzset{help lines/.style=very thin} |
| 447 |
\end{codeexample} |
| 448 |
|
| 449 |
To build a hierarchy of styles you can have one style use another. So in order |
| 450 |
to define a style |Karl's grid| that is based on the |grid| style Karl could |
| 451 |
say |
| 452 |
% |
| 453 |
\begin{codeexample}[code only] |
| 454 |
\tikzset{Karl's grid/.style={help lines,color=blue!50}} |
| 455 |
... |
| 456 |
\draw[Karl's grid] (0,0) grid (5,5); |
| 457 |
\end{codeexample} |
| 458 |
|
| 459 |
Styles are made even more powerful by parametrization. This means that, like |
| 460 |
other options, styles can also be used with a parameter. For instance, Karl |
| 461 |
could parameterize his grid so that, by default, it is blue, but he could also |
| 462 |
use another color. |
| 463 |
% |
| 464 |
\begin{codeexample}[code only] |
| 465 |
\begin{tikzpicture} |
| 466 |
[Karl's grid/.style ={help lines,color=#1!50}, |
| 467 |
Karl's grid/.default=blue] |
| 468 |
|
| 469 |
\draw[Karl's grid] (0,0) grid (1.5,2); |
| 470 |
\draw[Karl's grid=red] (2,0) grid (3.5,2); |
| 471 |
\end{tikzpicture} |
| 472 |
\end{codeexample} |
| 473 |
|
| 474 |
In this example, the definition of the style |Karl's grid| is given as an |
| 475 |
optional argument to the |{tikzpicture}| environment. Additional styles for other |
| 476 |
elements would follow after a comma. With many styles in effect, the optional |
| 477 |
argument of the environment may easily happen to be longer than the actual |
| 478 |
contents. |
| 479 |
|
| 480 |
\subsection{Drawing Options} |
| 481 |
|
| 482 |
Karl wonders what other options there are that influence how a path is drawn. |
| 483 |
He saw already that the |color=|\meta{color} option can be used to set the |
| 484 |
line's color. The option |draw=|\meta{color} does nearly the same, only it sets |
| 485 |
the color for the lines only and a different color can be used for filling |
| 486 |
(Karl will need this when he fills the arc for the angle). |
| 487 |
|
| 488 |
He saw that the style |very thin| yields very thin lines. Karl is not really |
| 489 |
surprised by this and neither is he surprised to learn that |thin| yields thin |
| 490 |
lines, |thick| yields thick lines, |very thick| yields very thick lines, |
| 491 |
|ultra thick| yields really, really thick lines and |ultra thin| yields lines |
| 492 |
that are so thin that low-resolution printers and displays will have trouble |
| 493 |
showing them. He wonders what gives lines of ``normal'' thickness. It turns out |
| 494 |
that |thin| is the correct choice, since it gives the same thickness as \TeX's |
| 495 |
|\hrule| command. Nevertheless, Karl would like to know whether there is |
| 496 |
anything ``in the middle'' between |thin| and |thick|. There is: |semithick|. |
| 497 |
|
| 498 |
Another useful thing one can do with lines is to dash or dot them. For this, |
| 499 |
the two styles |dashed| and |dotted| can be used, yielding \tikz[baseline] |
| 500 |
\draw[dashed] (0,.5ex) -- ++(2em,0pt); and \tikz[baseline] \draw[dotted] |
| 501 |
(0,.5ex) -- ++(2em,0pt);. Both options also exist in a loose and a dense |
| 502 |
version, called |loosely dashed|, |densely dashed|, |loosely dotted|, and |
| 503 |
|densely dotted|. If he really, really needs to, Karl can also define much |
| 504 |
more complex dashing patterns with the |dash pattern| option, but his son |
| 505 |
insists that dashing is to be used with utmost care and mostly distracts. |
| 506 |
Karl's son claims that complicated dashing patterns are evil. Karl's students |
| 507 |
do not care about dashing patterns. |
| 508 |
|
| 509 |
|
| 510 |
\subsection{Arc Path Construction} |
| 511 |
|
| 512 |
Our next obstacle is to draw the arc for the angle. For this, the |arc| path |
| 513 |
construction operation is useful, which draws part of a circle or ellipse. This |
| 514 |
|arc| operation is followed by options in brackets that specify the arc. An |
| 515 |
example would be \texttt{arc[start angle=10, end angle=80, radius=10pt]}, which |
| 516 |
means exactly what it says. Karl obviously needs an arc from $0^\circ$ to |
| 517 |
$30^\circ$. The radius should be something relatively small, perhaps around one |
| 518 |
third of the circle's radius. When one uses the arc path construction |
| 519 |
operation, the specified arc will be added with its starting point at the |
| 520 |
current position. So, we first have to ``get there''. |
| 521 |
% |
| 522 |
\begin{codeexample}[] |
| 523 |
\begin{tikzpicture} |
| 524 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 525 |
\draw (-1.5,0) -- (1.5,0); |
| 526 |
\draw (0,-1.5) -- (0,1.5); |
| 527 |
\draw (0,0) circle [radius=1cm]; |
| 528 |
\draw (3mm,0mm) arc [start angle=0, end angle=30, radius=3mm]; |
| 529 |
\end{tikzpicture} |
| 530 |
\end{codeexample} |
| 531 |
|
| 532 |
Karl thinks this is really a bit small and he cannot continue unless he learns |
| 533 |
how to do scaling. For this, he can add the |[scale=3]| option. He could add |
| 534 |
this option to each |\draw| command, but that would be awkward. Instead, he |
| 535 |
adds it to the whole environment, which causes this option to apply to |
| 536 |
everything within. |
| 537 |
% |
| 538 |
\begin{codeexample}[] |
| 539 |
\begin{tikzpicture}[scale=3] |
| 540 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 541 |
\draw (-1.5,0) -- (1.5,0); |
| 542 |
\draw (0,-1.5) -- (0,1.5); |
| 543 |
\draw (0,0) circle [radius=1cm]; |
| 544 |
\draw (3mm,0mm) arc [start angle=0, end angle=30, radius=3mm]; |
| 545 |
\end{tikzpicture} |
| 546 |
\end{codeexample} |
| 547 |
|
| 548 |
As for circles, you can specify ``two'' radii in order to get an elliptical |
| 549 |
arc. |
| 550 |
% |
| 551 |
\begin{codeexample}[] |
| 552 |
\tikz \draw (0,0) |
| 553 |
arc [start angle=0, end angle=315, |
| 554 |
x radius=1.75cm, y radius=1cm]; |
| 555 |
\end{codeexample} |
| 556 |
|
| 557 |
|
| 558 |
\subsection{Clipping a Path} |
| 559 |
|
| 560 |
In order to save space in this manual, it would be nice to clip Karl's graphics |
| 561 |
a bit so that we can focus on the ``interesting'' parts. Clipping is pretty |
| 562 |
easy in \tikzname. You can use the |\clip| command to clip all subsequent |
| 563 |
drawing. It works like |\draw|, only it does not draw anything, but uses the |
| 564 |
given path to clip everything subsequently. |
| 565 |
% |
| 566 |
\begin{codeexample}[] |
| 567 |
\begin{tikzpicture}[scale=3] |
| 568 |
\clip (-0.1,-0.2) rectangle (1.1,0.75); |
| 569 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 570 |
\draw (-1.5,0) -- (1.5,0); |
| 571 |
\draw (0,-1.5) -- (0,1.5); |
| 572 |
\draw (0,0) circle [radius=1cm]; |
| 573 |
\draw (3mm,0mm) arc [start angle=0, end angle=30, radius=3mm]; |
| 574 |
\end{tikzpicture} |
| 575 |
\end{codeexample} |
| 576 |
|
| 577 |
You can also do both at the same time: Draw \emph{and} clip a path. For this, |
| 578 |
use the |\draw| command and add the |clip| option. (This is not the whole |
| 579 |
picture: You can also use the |\clip| command and add the |draw| option. Well, |
| 580 |
that is also not the whole picture: In reality, |\draw| is just a shorthand for |
| 581 |
|\path[draw]| and |\clip| is a shorthand for |\path[clip]| and you could also |
| 582 |
say |\path[draw,clip]|.) Here is an example: |
| 583 |
% |
| 584 |
\begin{codeexample}[] |
| 585 |
\begin{tikzpicture}[scale=3] |
| 586 |
\clip[draw] (0.5,0.5) circle (.6cm); |
| 587 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 588 |
\draw (-1.5,0) -- (1.5,0); |
| 589 |
\draw (0,-1.5) -- (0,1.5); |
| 590 |
\draw (0,0) circle [radius=1cm]; |
| 591 |
\draw (3mm,0mm) arc [start angle=0, end angle=30, radius=3mm]; |
| 592 |
\end{tikzpicture} |
| 593 |
\end{codeexample} |
| 594 |
|
| 595 |
|
| 596 |
\subsection{Parabola and Sine Path Construction} |
| 597 |
|
| 598 |
Although Karl does not need them for his picture, he is pleased to learn that |
| 599 |
there are |parabola| and |sin| and |cos| path operations for adding parabolas |
| 600 |
and sine and cosine curves to the current path. For the |parabola| operation, |
| 601 |
the current point will lie on the parabola as well as the point given after the |
| 602 |
parabola operation. Consider the following example: |
| 603 |
% |
| 604 |
\begin{codeexample}[] |
| 605 |
\tikz \draw (0,0) rectangle (1,1) (0,0) parabola (1,1); |
| 606 |
\end{codeexample} |
| 607 |
|
| 608 |
It is also possible to place the bend somewhere else: |
| 609 |
% |
| 610 |
\begin{codeexample}[] |
| 611 |
\tikz \draw[x=1pt,y=1pt] (0,0) parabola bend (4,16) (6,12); |
| 612 |
\end{codeexample} |
| 613 |
|
| 614 |
The operations |sin| and |cos| add a sine or cosine curve in the interval |
| 615 |
$[0,\pi/2]$ such that the previous current point is at the start of the curve |
| 616 |
and the curve ends at the given end point. Here are two examples: |
| 617 |
% |
| 618 |
\begin{codeexample}[] |
| 619 |
A sine \tikz \draw[x=1ex,y=1ex] (0,0) sin (1.57,1); curve. |
| 620 |
\end{codeexample} |
| 621 |
|
| 622 |
\begin{codeexample}[] |
| 623 |
\tikz \draw[x=1.57ex,y=1ex] (0,0) sin (1,1) cos (2,0) sin (3,-1) cos (4,0) |
| 624 |
(0,1) cos (1,0) sin (2,-1) cos (3,0) sin (4,1); |
| 625 |
\end{codeexample} |
| 626 |
|
| 627 |
|
| 628 |
\subsection{Filling and Drawing} |
| 629 |
|
| 630 |
Returning to the picture, Karl now wants the angle to be ``filled'' with a very |
| 631 |
light green. For this he uses |\fill| instead of |\draw|. Here is what Karl |
| 632 |
does: |
| 633 |
% |
| 634 |
\begin{codeexample}[] |
| 635 |
\begin{tikzpicture}[scale=3] |
| 636 |
\clip (-0.1,-0.2) rectangle (1.1,0.75); |
| 637 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 638 |
\draw (-1.5,0) -- (1.5,0); |
| 639 |
\draw (0,-1.5) -- (0,1.5); |
| 640 |
\draw (0,0) circle [radius=1cm]; |
| 641 |
\fill[green!20!white] (0,0) -- (3mm,0mm) |
| 642 |
arc [start angle=0, end angle=30, radius=3mm] -- (0,0); |
| 643 |
\end{tikzpicture} |
| 644 |
\end{codeexample} |
| 645 |
|
| 646 |
The color |green!20!white| means 20\% green and 80\% white mixed together. Such |
| 647 |
color expression are possible since \tikzname\ uses Uwe Kern's |xcolor| |
| 648 |
package, see the documentation of that package for details on color |
| 649 |
expressions. |
| 650 |
|
| 651 |
What would have happened, if Karl had not ``closed'' the path using |--(0,0)| |
| 652 |
at the end? In this case, the path is closed automatically, so this could have |
| 653 |
been omitted. Indeed, it would even have been better to write the following, |
| 654 |
instead: |
| 655 |
% |
| 656 |
\begin{codeexample}[code only] |
| 657 |
\fill[green!20!white] (0,0) -- (3mm,0mm) |
| 658 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 659 |
\end{codeexample} |
| 660 |
% |
| 661 |
The |--cycle| causes the current path to be closed (actually the current part |
| 662 |
of the current path) by smoothly joining the first and last point. To |
| 663 |
appreciate the difference, consider the following example: |
| 664 |
% |
| 665 |
\begin{codeexample}[] |
| 666 |
\begin{tikzpicture}[line width=5pt] |
| 667 |
\draw (0,0) -- (1,0) -- (1,1) -- (0,0); |
| 668 |
\draw (2,0) -- (3,0) -- (3,1) -- cycle; |
| 669 |
\useasboundingbox (0,1.5); % make bounding box higher |
| 670 |
\end{tikzpicture} |
| 671 |
\end{codeexample} |
| 672 |
|
| 673 |
You can also fill and draw a path at the same time using the |\filldraw| |
| 674 |
command. This will first draw the path, then fill it. This may not seem too |
| 675 |
useful, but you can specify different colors to be used for filling and for |
| 676 |
stroking. These are specified as optional arguments like this: |
| 677 |
% |
| 678 |
\begin{codeexample}[] |
| 679 |
\begin{tikzpicture}[scale=3] |
| 680 |
\clip (-0.1,-0.2) rectangle (1.1,0.75); |
| 681 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 682 |
\draw (-1.5,0) -- (1.5,0); |
| 683 |
\draw (0,-1.5) -- (0,1.5); |
| 684 |
\draw (0,0) circle [radius=1cm]; |
| 685 |
\filldraw[fill=green!20!white, draw=green!50!black] (0,0) -- (3mm,0mm) |
| 686 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 687 |
\end{tikzpicture} |
| 688 |
\end{codeexample} |
| 689 |
|
| 690 |
|
| 691 |
\subsection{Shading} |
| 692 |
|
| 693 |
Karl briefly considers the possibility of making the angle ``more fancy'' by |
| 694 |
\emph{shading} it. Instead of filling the area with a uniform color, a smooth |
| 695 |
transition between different colors is used. For this, |\shade| and |
| 696 |
|\shadedraw|, for shading and drawing at the same time, can be used: |
| 697 |
% |
| 698 |
\begin{codeexample}[] |
| 699 |
\tikz \shade (0,0) rectangle (2,1) (3,0.5) circle (.5cm); |
| 700 |
\end{codeexample} |
| 701 |
% |
| 702 |
The default shading is a smooth transition from gray to white. To specify |
| 703 |
different colors, you can use options: |
| 704 |
% |
| 705 |
\begin{codeexample}[] |
| 706 |
\begin{tikzpicture}[rounded corners,ultra thick] |
| 707 |
\shade[top color=yellow,bottom color=black] (0,0) rectangle +(2,1); |
| 708 |
\shade[left color=yellow,right color=black] (3,0) rectangle +(2,1); |
| 709 |
\shadedraw[inner color=yellow,outer color=black,draw=yellow] (6,0) rectangle +(2,1); |
| 710 |
\shade[ball color=green] (9,.5) circle (.5cm); |
| 711 |
\end{tikzpicture} |
| 712 |
\end{codeexample} |
| 713 |
|
| 714 |
For Karl, the following might be appropriate: |
| 715 |
% |
| 716 |
\begin{codeexample}[] |
| 717 |
\begin{tikzpicture}[scale=3] |
| 718 |
\clip (-0.1,-0.2) rectangle (1.1,0.75); |
| 719 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 720 |
\draw (-1.5,0) -- (1.5,0); |
| 721 |
\draw (0,-1.5) -- (0,1.5); |
| 722 |
\draw (0,0) circle [radius=1cm]; |
| 723 |
\shadedraw[left color=gray,right color=green, draw=green!50!black] |
| 724 |
(0,0) -- (3mm,0mm) |
| 725 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 726 |
\end{tikzpicture} |
| 727 |
\end{codeexample} |
| 728 |
|
| 729 |
However, he wisely decides that shadings usually only distract without adding |
| 730 |
anything to the picture. |
| 731 |
|
| 732 |
|
| 733 |
\subsection{Specifying Coordinates} |
| 734 |
|
| 735 |
Karl now wants to add the sine and cosine lines. He knows already that he can |
| 736 |
use the |color=| option to set the lines' colors. So, what is the best way to |
| 737 |
specify the coordinates? |
| 738 |
|
| 739 |
There are different ways of specifying coordinates. The easiest way is to say |
| 740 |
something like |(10pt,2cm)|. This means 10pt in $x$-direction and 2cm in |
| 741 |
$y$-directions. Alternatively, you can also leave out the units as in |(1,2)|, |
| 742 |
which means ``one times the current $x$-vector plus twice the current |
| 743 |
$y$-vector''. These vectors default to 1cm in the $x$-direction and 1cm in the |
| 744 |
$y$-direction, respectively. |
| 745 |
|
| 746 |
In order to specify points in polar coordinates, use the notation |(30:1cm)|, |
| 747 |
which means 1cm in direction 30 degree. This is obviously quite useful to ``get |
| 748 |
to the point $(\cos 30^\circ,\sin 30^\circ)$ on the circle''. |
| 749 |
|
| 750 |
You can add a single |+| sign in front of a coordinate or two of them as in |
| 751 |
|+(0cm,1cm)| or |++(2cm,0cm)|. Such coordinates are interpreted differently: |
| 752 |
The first form means ``1cm upwards from the previous specified position'' and |
| 753 |
the second means ``2cm to the right of the previous specified position, making |
| 754 |
this the new specified position''. For example, we can draw the sine line as |
| 755 |
follows: |
| 756 |
% |
| 757 |
\begin{codeexample}[] |
| 758 |
\begin{tikzpicture}[scale=3] |
| 759 |
\clip (-0.1,-0.2) rectangle (1.1,0.75); |
| 760 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 761 |
\draw (-1.5,0) -- (1.5,0); |
| 762 |
\draw (0,-1.5) -- (0,1.5); |
| 763 |
\draw (0,0) circle [radius=1cm]; |
| 764 |
\filldraw[fill=green!20,draw=green!50!black] (0,0) -- (3mm,0mm) |
| 765 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 766 |
\draw[red,very thick] (30:1cm) -- +(0,-0.5); |
| 767 |
\end{tikzpicture} |
| 768 |
\end{codeexample} |
| 769 |
|
| 770 |
Karl used the fact $\sin 30^\circ = 1/2$. However, he very much doubts that his |
| 771 |
students know this, so it would be nice to have a way of specifying ``the point |
| 772 |
straight down from |(30:1cm)| that lies on the $x$-axis''. This is, indeed, |
| 773 |
possible using a special syntax: Karl can write \verb!(30:1cm |- 0,0)!. In |
| 774 |
general, the meaning of |(|\meta{p}\verb! |- !\meta{q}|)| is ``the intersection |
| 775 |
of a vertical line through $p$ and a horizontal line through $q$''. |
| 776 |
|
| 777 |
Next, let us draw the cosine line. One way would be to say |
| 778 |
\verb!(30:1cm |- 0,0) -- (0,0)!. Another way is the following: we ``continue'' |
| 779 |
from where the sine ends: |
| 780 |
% |
| 781 |
\begin{codeexample}[] |
| 782 |
\begin{tikzpicture}[scale=3] |
| 783 |
\clip (-0.1,-0.2) rectangle (1.1,0.75); |
| 784 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 785 |
\draw (-1.5,0) -- (1.5,0); |
| 786 |
\draw (0,-1.5) -- (0,1.5); |
| 787 |
\draw (0,0) circle [radius=1cm]; |
| 788 |
\filldraw[fill=green!20,draw=green!50!black] (0,0) -- (3mm,0mm) |
| 789 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 790 |
\draw[red,very thick] (30:1cm) -- +(0,-0.5); |
| 791 |
\draw[blue,very thick] (30:1cm) ++(0,-0.5) -- (0,0); |
| 792 |
\end{tikzpicture} |
| 793 |
\end{codeexample} |
| 794 |
|
| 795 |
Note that there is no |--| between |(30:1cm)| and |++(0,-0.5)|. In detail, this |
| 796 |
path is interpreted as follows: ``First, the |(30:1cm)| tells me to move my pen |
| 797 |
to $(\cos 30^\circ,1/2)$. Next, there comes another coordinate specification, |
| 798 |
so I move my pen there without drawing anything. This new point is half a unit |
| 799 |
down from the last position, thus it is at $(\cos 30^\circ,0)$. Finally, I move |
| 800 |
the pen to the origin, but this time drawing something (because of the |--|).'' |
| 801 |
|
| 802 |
To appreciate the difference between |+| and |++| consider the following |
| 803 |
example: |
| 804 |
% |
| 805 |
\begin{codeexample}[] |
| 806 |
\begin{tikzpicture} |
| 807 |
\def\rectanglepath{-- ++(1cm,0cm) -- ++(0cm,1cm) -- ++(-1cm,0cm) -- cycle} |
| 808 |
\draw (0,0) \rectanglepath; |
| 809 |
\draw (1.5,0) \rectanglepath; |
| 810 |
\end{tikzpicture} |
| 811 |
\end{codeexample} |
| 812 |
|
| 813 |
By comparison, when using a single |+|, the coordinates are different: |
| 814 |
% |
| 815 |
\begin{codeexample}[] |
| 816 |
\begin{tikzpicture} |
| 817 |
\def\rectanglepath{-- +(1cm,0cm) -- +(1cm,1cm) -- +(0cm,1cm) -- cycle} |
| 818 |
\draw (0,0) \rectanglepath; |
| 819 |
\draw (1.5,0) \rectanglepath; |
| 820 |
\end{tikzpicture} |
| 821 |
\end{codeexample} |
| 822 |
|
| 823 |
|
| 824 |
Naturally, all of this could have been written more clearly and more |
| 825 |
economically like this (either with a single or a double |+|): |
| 826 |
% |
| 827 |
\begin{codeexample}[] |
| 828 |
\tikz \draw (0,0) rectangle +(1,1) (1.5,0) rectangle +(1,1); |
| 829 |
\end{codeexample} |
| 830 |
|
| 831 |
|
| 832 |
\subsection{Intersecting Paths} |
| 833 |
|
| 834 |
Karl is left with the line for $\tan \alpha$, which seems difficult to specify |
| 835 |
using transformations and polar coordinates. The first -- and easiest -- thing |
| 836 |
he can do is so simply use the coordinate |(1,{tan(30)})| since \tikzname's |
| 837 |
math engine knows how to compute things like |tan(30)|. Note the added braces |
| 838 |
since, otherwise, \tikzname's parser would think that the first closing |
| 839 |
parenthesis ends the coordinate (in general, you need to add braces around |
| 840 |
components of coordinates when these components contain parentheses). |
| 841 |
|
| 842 |
Karl can, however, also use a more elaborate, but also more ``geometric'' way |
| 843 |
of computing the length of the orange line: He can specify intersections of |
| 844 |
paths as coordinates. The line for $\tan \alpha$ starts at $(1,0)$ and goes |
| 845 |
upward to a point that is at the intersection of a line going ``up'' and a line |
| 846 |
going from the origin through |(30:1cm)|. Such computations are made available |
| 847 |
by the |intersections| library. |
| 848 |
|
| 849 |
What Karl must do is to create two ``invisible'' paths that intersect at the |
| 850 |
position of interest. Creating paths that are not otherwise seen can be done |
| 851 |
using the |\path| command without any options like |draw| or |fill|. Then, Karl |
| 852 |
can add the |name path| option to the path for later reference. Once the paths |
| 853 |
have been constructed, Karl can use the |name intersections| to assign names to |
| 854 |
the coordinate for later reference. |
| 855 |
% |
| 856 |
\begin{codeexample}[code only] |
| 857 |
\path [name path=upward line] (1,0) -- (1,1); |
| 858 |
\path [name path=sloped line] (0,0) -- (30:1.5cm); % a bit longer, so that there is an intersection |
| 859 |
|
| 860 |
% (add `\usetikzlibrary{intersections}' after loading tikz in the preamble) |
| 861 |
\draw [name intersections={of=upward line and sloped line, by=x}] |
| 862 |
[very thick,orange] (1,0) -- (x); |
| 863 |
\end{codeexample} |
| 864 |
|
| 865 |
|
| 866 |
\subsection{Adding Arrow Tips} |
| 867 |
|
| 868 |
Karl now wants to add the little arrow tips at the end of the axes. He has |
| 869 |
noticed that in many plots, even in scientific journals, these arrow tips seem |
| 870 |
to be missing, presumably because the generating programs cannot produce them. |
| 871 |
Karl thinks arrow tips belong at the end of axes. His son agrees. His students |
| 872 |
do not care about arrow tips. |
| 873 |
|
| 874 |
It turns out that adding arrow tips is pretty easy: Karl adds the option |->| |
| 875 |
to the drawing commands for the axes: |
| 876 |
% |
| 877 |
\begin{codeexample}[preamble={\usetikzlibrary{intersections}}] |
| 878 |
\begin{tikzpicture}[scale=3] |
| 879 |
\clip (-0.1,-0.2) rectangle (1.1,1.51); |
| 880 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 881 |
\draw[->] (-1.5,0) -- (1.5,0); |
| 882 |
\draw[->] (0,-1.5) -- (0,1.5); |
| 883 |
\draw (0,0) circle [radius=1cm]; |
| 884 |
\filldraw[fill=green!20,draw=green!50!black] (0,0) -- (3mm,0mm) |
| 885 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 886 |
\draw[red,very thick] (30:1cm) -- +(0,-0.5); |
| 887 |
\draw[blue,very thick] (30:1cm) ++(0,-0.5) -- (0,0); |
| 888 |
|
| 889 |
\path [name path=upward line] (1,0) -- (1,1); |
| 890 |
\path [name path=sloped line] (0,0) -- (30:1.5cm); |
| 891 |
\draw [name intersections={of=upward line and sloped line, by=x}] |
| 892 |
[very thick,orange] (1,0) -- (x); |
| 893 |
\end{tikzpicture} |
| 894 |
\end{codeexample} |
| 895 |
|
| 896 |
If Karl had used the option |<-| instead of |->|, arrow tips would have been |
| 897 |
put at the beginning of the path. The option |<->| puts arrow tips at both ends |
| 898 |
of the path. |
| 899 |
|
| 900 |
There are certain restrictions to the kind of paths to which arrow tips can be |
| 901 |
added. As a rule of thumb, you can add arrow tips only to a single open |
| 902 |
``line''. For example, you cannot add tips to, say, a rectangle or a circle. |
| 903 |
However, you can add arrow tips to curved paths and to paths that have several |
| 904 |
segments, as in the following examples: |
| 905 |
% |
| 906 |
\begin{codeexample}[] |
| 907 |
\begin{tikzpicture} |
| 908 |
\draw [<->] (0,0) arc [start angle=180, end angle=30, radius=10pt]; |
| 909 |
\draw [<->] (1,0) -- (1.5cm,10pt) -- (2cm,0pt) -- (2.5cm,10pt); |
| 910 |
\end{tikzpicture} |
| 911 |
\end{codeexample} |
| 912 |
|
| 913 |
Karl has a more detailed look at the arrow that \tikzname\ puts at the end. It |
| 914 |
looks like this when he zooms it: \tikz[baseline] \draw[->,line width=1pt] |
| 915 |
(0pt,.5ex) -- ++(10pt,0pt);. The shape seems vaguely familiar and, indeed, this |
| 916 |
is exactly the end of \TeX's standard arrow used in something like $f\colon A |
| 917 |
\to B$. |
| 918 |
|
| 919 |
Karl likes the arrow, especially since it is not ``as thick'' as the arrows |
| 920 |
offered by many other packages. However, he expects that, sometimes, he might |
| 921 |
need to use some other kinds of arrow. To do so, Karl can say |>=|\meta{kind of |
| 922 |
end arrow tip}, where \meta{kind of end arrow tip} is a special arrow tip |
| 923 |
specification. For example, if Karl says |>=Stealth|, then he tells \tikzname\ |
| 924 |
that he would like ``stealth-fighter-like'' arrow tips: |
| 925 |
\todosp{remaining instance of bug \#473} |
| 926 |
% |
| 927 |
\begin{codeexample}[preamble={\usetikzlibrary{arrows.meta}}] |
| 928 |
\begin{tikzpicture}[>=Stealth] |
| 929 |
\draw [->] (0,0) arc [start angle=180, end angle=30, radius=10pt]; |
| 930 |
\draw [<<-,very thick] (1,0) -- (1.5cm,10pt) -- (2cm,0pt) -- (2.5cm,10pt); |
| 931 |
\end{tikzpicture} |
| 932 |
\end{codeexample} |
| 933 |
|
| 934 |
Karl wonders whether such a military name for the arrow type is really |
| 935 |
necessary. He is not really mollified when his son tells him that Microsoft's |
| 936 |
PowerPoint uses the same name. He decides to have his students discuss this at |
| 937 |
some point. |
| 938 |
|
| 939 |
In addition to |Stealth| there are several other predefined kinds of arrow tips |
| 940 |
Karl can choose from, see Section~\ref{section-arrows}. Furthermore, he can |
| 941 |
define arrows types himself, if he needs new ones. |
| 942 |
|
| 943 |
|
| 944 |
\subsection{Scoping} |
| 945 |
|
| 946 |
Karl saw already that there are numerous graphic options that affect how paths |
| 947 |
are rendered. Often, he would like to apply certain options to a whole set of |
| 948 |
graphic commands. For example, Karl might wish to draw three paths using a |
| 949 |
|thick| pen, but would like everything else to be drawn ``normally''. |
| 950 |
|
| 951 |
If Karl wishes to set a certain graphic option for the whole picture, he can |
| 952 |
simply pass this option to the |\tikz| command or to the |{tikzpicture}| |
| 953 |
environment (Gerda would pass the options to |\tikzpicture| and Hans passes |
| 954 |
them to |\starttikzpicture|). However, if Karl wants to apply graphic options |
| 955 |
to a local group, he put these commands inside a |{scope}| environment (Gerda |
| 956 |
uses |\scope| and |\endscope|, Hans uses |\startscope| and |\stopscope|). This |
| 957 |
environment takes graphic options as an optional argument and these options |
| 958 |
apply to everything inside the scope, but not to anything outside. |
| 959 |
|
| 960 |
Here is an example: |
| 961 |
% |
| 962 |
\begin{codeexample}[] |
| 963 |
\begin{tikzpicture}[ultra thick] |
| 964 |
\draw (0,0) -- (0,1); |
| 965 |
\begin{scope}[thin] |
| 966 |
\draw (1,0) -- (1,1); |
| 967 |
\draw (2,0) -- (2,1); |
| 968 |
\end{scope} |
| 969 |
\draw (3,0) -- (3,1); |
| 970 |
\end{tikzpicture} |
| 971 |
\end{codeexample} |
| 972 |
|
| 973 |
Scoping has another interesting effect: Any changes to the clipping area are |
| 974 |
local to the scope. Thus, if you say |\clip| somewhere inside a scope, the |
| 975 |
effect of the |\clip| command ends at the end of the scope. This is useful |
| 976 |
since there is no other way of ``enlarging'' the clipping area. |
| 977 |
|
| 978 |
Karl has also already seen that giving options to commands like |\draw| apply |
| 979 |
only to that command. It turns out that the situation is slightly more complex. |
| 980 |
First, options to a command like |\draw| are not really options to the command, |
| 981 |
but they are ``path options'' and can be given anywhere on the path. So, |
| 982 |
instead of |\draw[thin] (0,0) -- (1,0);| one can also write |
| 983 |
|\draw (0,0) [thin] -- (1,0);| or |\draw (0,0) -- (1,0) [thin];|; all of these |
| 984 |
have the same effect. This might seem strange since in the last case, it would |
| 985 |
appear that the |thin| should take effect only ``after'' the line from $(0,0)$ |
| 986 |
to $(1,0)$ has been drawn. However, most graphic options only apply to the |
| 987 |
whole path. Indeed, if you say both |thin| and |thick| on the same path, the |
| 988 |
last option given will ``win''. |
| 989 |
|
| 990 |
When reading the above, Karl notices that only ``most'' graphic options apply |
| 991 |
to the whole path. Indeed, all transformation options do \emph{not} apply to |
| 992 |
the whole path, but only to ``everything following them on the path''. We will |
| 993 |
have a more detailed look at this in a moment. Nevertheless, all options given |
| 994 |
during a path construction apply only to this path. |
| 995 |
|
| 996 |
|
| 997 |
\subsection{Transformations} |
| 998 |
|
| 999 |
When you specify a coordinate like |(1cm,1cm)|, where is that coordinate |
| 1000 |
placed on the page? To determine the position, \tikzname, \TeX, and |
| 1001 |
\textsc{pdf} or PostScript all apply certain transformations to the given |
| 1002 |
coordinate in order to determine the final position on the page. |
| 1003 |
|
| 1004 |
\tikzname\ provides numerous options that allow you to transform coordinates in |
| 1005 |
\tikzname's private coordinate system. For example, the |xshift| option allows |
| 1006 |
you to shift all subsequent points by a certain amount: |
| 1007 |
|
| 1008 |
\begin{codeexample}[] |
| 1009 |
\tikz \draw (0,0) -- (0,0.5) [xshift=2pt] (0,0) -- (0,0.5); |
| 1010 |
\end{codeexample} |
| 1011 |
|
| 1012 |
It is important to note that you can change transformation ``in the middle of a |
| 1013 |
path'', a feature that is not supported by \pdf\ or PostScript. The reason is |
| 1014 |
that \tikzname\ keeps track of its own transformation matrix. |
| 1015 |
|
| 1016 |
Here is a more complicated example: |
| 1017 |
% |
| 1018 |
\begin{codeexample}[] |
| 1019 |
\begin{tikzpicture}[even odd rule,rounded corners=2pt,x=10pt,y=10pt] |
| 1020 |
\filldraw[fill=yellow!80!black] (0,0) rectangle (1,1) |
| 1021 |
[xshift=5pt,yshift=5pt] (0,0) rectangle (1,1) |
| 1022 |
[rotate=30] (-1,-1) rectangle (2,2); |
| 1023 |
\end{tikzpicture} |
| 1024 |
\end{codeexample} |
| 1025 |
|
| 1026 |
The most useful transformations are |xshift| and |yshift| for shifting, |shift| |
| 1027 |
for shifting to a given point as in |shift={(1,0)}| or |shift={+(0,0)}| (the |
| 1028 |
braces are necessary so that \TeX\ does not mistake the comma for separating |
| 1029 |
options), |rotate| for rotating by a certain angle (there is also a |
| 1030 |
|rotate around| for rotating around a given point), |scale| for scaling by a |
| 1031 |
certain factor, |xscale| and |yscale| for scaling only in the $x$- or |
| 1032 |
$y$-direction (|xscale=-1| is a flip), and |xslant| and |yslant| for slanting. |
| 1033 |
If these transformation and those that I have not mentioned are not sufficient, |
| 1034 |
the |cm| option allows you to apply an arbitrary transformation matrix. Karl's |
| 1035 |
students, by the way, do not know what a transformation matrix is. |
| 1036 |
|
| 1037 |
|
| 1038 |
\subsection{Repeating Things: For-Loops} |
| 1039 |
|
| 1040 |
Karl's next aim is to add little ticks on the axes at positions $-1$, $-1/2$, |
| 1041 |
$1/2$, and $1$. For this, it would be nice to use some kind of ``loop'', |
| 1042 |
especially since he wishes to do the same thing at each of these positions. |
| 1043 |
There are different packages for doing this. \LaTeX\ has its own internal |
| 1044 |
command for this, |pstricks| comes along with the powerful |\multido| command. |
| 1045 |
All of these can be used together with \tikzname, so if you are familiar with |
| 1046 |
them, feel free to use them. \tikzname\ introduces yet another command, called |
| 1047 |
|\foreach|, which I introduced since I could never remember the syntax of the |
| 1048 |
other packages. |\foreach| is defined in the package |pgffor| and can be used |
| 1049 |
independently of \tikzname, but \tikzname\ includes it automatically. |
| 1050 |
|
| 1051 |
In its basic form, the |\foreach| command is easy to use: |
| 1052 |
% |
| 1053 |
\begin{codeexample}[] |
| 1054 |
\foreach \x in {1,2,3} {$x =\x$, } |
| 1055 |
\end{codeexample} |
| 1056 |
|
| 1057 |
The general syntax is |
| 1058 |
|\foreach| \meta{variable}| in {|\meta{list of values}|} |\meta{commands}. |
| 1059 |
Inside the \meta{commands}, the \meta{variable} will be assigned to the |
| 1060 |
different values. If the \meta{commands} do not start with a brace, everything |
| 1061 |
up to the next semicolon is used as \meta{commands}. |
| 1062 |
|
| 1063 |
For Karl and the ticks on the axes, he could use the following code: |
| 1064 |
% |
| 1065 |
\begin{codeexample}[] |
| 1066 |
\begin{tikzpicture}[scale=3] |
| 1067 |
\clip (-0.1,-0.2) rectangle (1.1,1.51); |
| 1068 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 1069 |
\filldraw[fill=green!20,draw=green!50!black] (0,0) -- (3mm,0mm) |
| 1070 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 1071 |
\draw[->] (-1.5,0) -- (1.5,0); |
| 1072 |
\draw[->] (0,-1.5) -- (0,1.5); |
| 1073 |
\draw (0,0) circle [radius=1cm]; |
| 1074 |
|
| 1075 |
\foreach \x in {-1cm,-0.5cm,1cm} |
| 1076 |
\draw (\x,-1pt) -- (\x,1pt); |
| 1077 |
\foreach \y in {-1cm,-0.5cm,0.5cm,1cm} |
| 1078 |
\draw (-1pt,\y) -- (1pt,\y); |
| 1079 |
\end{tikzpicture} |
| 1080 |
\end{codeexample} |
| 1081 |
|
| 1082 |
As a matter of fact, there are many different ways of creating the ticks. For |
| 1083 |
example, Karl could have put the |\draw ...;| inside curly braces. He could |
| 1084 |
also have used, say, |
| 1085 |
% |
| 1086 |
\begin{codeexample}[code only] |
| 1087 |
\foreach \x in {-1,-0.5,1} |
| 1088 |
\draw[xshift=\x cm] (0pt,-1pt) -- (0pt,1pt); |
| 1089 |
\end{codeexample} |
| 1090 |
|
| 1091 |
Karl is curious what would happen in a more complicated situation where there |
| 1092 |
are, say, 20 ticks. It seems bothersome to explicitly mention all these numbers |
| 1093 |
in the set for |\foreach|. Indeed, it is possible to use |...| inside the |
| 1094 |
|\foreach| statement to iterate over a large number of values (which must, |
| 1095 |
however, be dimensionless real numbers) as in the following example: |
| 1096 |
% |
| 1097 |
\begin{codeexample}[] |
| 1098 |
\tikz \foreach \x in {1,...,10} |
| 1099 |
\draw (\x,0) circle (0.4cm); |
| 1100 |
\end{codeexample} |
| 1101 |
|
| 1102 |
If you provide \emph{two} numbers before the |...|, the |\foreach| statement |
| 1103 |
will use their difference for the stepping: |
| 1104 |
% |
| 1105 |
\begin{codeexample}[] |
| 1106 |
\tikz \foreach \x in {-1,-0.5,...,1} |
| 1107 |
\draw (\x cm,-1pt) -- (\x cm,1pt); |
| 1108 |
\end{codeexample} |
| 1109 |
|
| 1110 |
We can also nest loops to create interesting effects: |
| 1111 |
% |
| 1112 |
\begin{codeexample}[] |
| 1113 |
\begin{tikzpicture} |
| 1114 |
\foreach \x in {1,2,...,5,7,8,...,12} |
| 1115 |
\foreach \y in {1,...,5} |
| 1116 |
{ |
| 1117 |
\draw (\x,\y) +(-.5,-.5) rectangle ++(.5,.5); |
| 1118 |
\draw (\x,\y) node{\x,\y}; |
| 1119 |
} |
| 1120 |
\end{tikzpicture} |
| 1121 |
\end{codeexample} |
| 1122 |
|
| 1123 |
The |\foreach| statement can do even trickier stuff, but the above gives the |
| 1124 |
idea. |
| 1125 |
|
| 1126 |
|
| 1127 |
\subsection{Adding Text} |
| 1128 |
|
| 1129 |
Karl is, by now, quite satisfied with the picture. However, the most important |
| 1130 |
parts, namely the labels, are still missing! |
| 1131 |
|
| 1132 |
\tikzname\ offers an easy-to-use and powerful system for adding text and, more |
| 1133 |
generally, complex shapes to a picture at specific positions. The basic idea is |
| 1134 |
the following: When \tikzname\ is constructing a path and encounters the |
| 1135 |
keyword |node| in the middle of a path, it reads a \emph{node specification}. |
| 1136 |
The keyword |node| is typically followed by some options and then some text |
| 1137 |
between curly braces. This text is put inside a normal \TeX\ box (if the node |
| 1138 |
specification directly follows a coordinate, which is usually the case, |
| 1139 |
\tikzname\ is able to perform some magic so that it is even possible to use |
| 1140 |
verbatim text inside the boxes) and then placed at the current position, that |
| 1141 |
is, at the last specified position (possibly shifted a bit, according to the |
| 1142 |
given options). However, all nodes are drawn only after the path has been |
| 1143 |
completely drawn/filled/shaded/clipped/whatever. |
| 1144 |
% |
| 1145 |
\begin{codeexample}[] |
| 1146 |
\begin{tikzpicture} |
| 1147 |
\draw (0,0) rectangle (2,2); |
| 1148 |
\draw (0.5,0.5) node [fill=yellow!80!black] |
| 1149 |
{Text at \verb!node 1!} |
| 1150 |
-- (1.5,1.5) node {Text at \verb!node 2!}; |
| 1151 |
\end{tikzpicture} |
| 1152 |
\end{codeexample} |
| 1153 |
|
| 1154 |
Obviously, Karl would not only like to place nodes \emph{on} the last specified |
| 1155 |
position, but also to the left or the right of these positions. For this, every |
| 1156 |
node object that you put in your picture is equipped with several |
| 1157 |
\emph{anchors}. For example, the |north| anchor is in the middle at the upper |
| 1158 |
end of the shape, the |south| anchor is at the bottom and the |north east| |
| 1159 |
anchor is in the upper right corner. When you give the option |anchor=north|, |
| 1160 |
the text will be placed such that this northern anchor will lie on the current |
| 1161 |
position and the text is, thus, below the current position. Karl uses this to |
| 1162 |
draw the ticks as follows: |
| 1163 |
% |
| 1164 |
\begin{codeexample}[] |
| 1165 |
\begin{tikzpicture}[scale=3] |
| 1166 |
\clip (-0.6,-0.2) rectangle (0.6,1.51); |
| 1167 |
\draw[step=.5cm,help lines] (-1.4,-1.4) grid (1.4,1.4); |
| 1168 |
\filldraw[fill=green!20,draw=green!50!black] (0,0) -- (3mm,0mm) |
| 1169 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 1170 |
\draw[->] (-1.5,0) -- (1.5,0); \draw[->] (0,-1.5) -- (0,1.5); |
| 1171 |
\draw (0,0) circle [radius=1cm]; |
| 1172 |
|
| 1173 |
\foreach \x in {-1,-0.5,1} |
| 1174 |
\draw (\x cm,1pt) -- (\x cm,-1pt) node[anchor=north] {$\x$}; |
| 1175 |
\foreach \y in {-1,-0.5,0.5,1} |
| 1176 |
\draw (1pt,\y cm) -- (-1pt,\y cm) node[anchor=east] {$\y$}; |
| 1177 |
\end{tikzpicture} |
| 1178 |
\end{codeexample} |
| 1179 |
|
| 1180 |
This is quite nice, already. Using these anchors, Karl can now add most of the |
| 1181 |
other text elements. However, Karl thinks that, though ``correct'', it is quite |
| 1182 |
counter-intuitive that in order to place something \emph{below} a given point, |
| 1183 |
he has to use the \emph{north} anchor. For this reason, there is an option |
| 1184 |
called |below|, which does the same as |anchor=north|. Similarly, |above right| |
| 1185 |
does the same as |anchor=south west|. In addition, |below| takes an optional |
| 1186 |
dimension argument. If given, the shape will additionally be shifted downwards |
| 1187 |
by the given amount. So, |below=1pt| can be used to put a text label below some |
| 1188 |
point and, additionally shift it 1pt downwards. |
| 1189 |
|
| 1190 |
Karl is not quite satisfied with the ticks. He would like to have $1/2$ or |
| 1191 |
$\frac{1}{2}$ shown instead of $0.5$, partly to show off the nice capabilities |
| 1192 |
of \TeX\ and \tikzname, partly because for positions like $1/3$ or $\pi$ it is |
| 1193 |
certainly very much preferable to have the ``mathematical'' tick there instead |
| 1194 |
of just the ``numeric'' tick. His students, on the other hand, prefer $0.5$ |
| 1195 |
over $1/2$ since they are not too fond of fractions in general. |
| 1196 |
|
| 1197 |
Karl now faces a problem: For the |\foreach| statement, the position |\x| |
| 1198 |
should still be given as |0.5| since \tikzname\ will not know where |
| 1199 |
|\frac{1}{2}| is supposed to be. On the other hand, the typeset text should |
| 1200 |
really be |\frac{1}{2}|. To solve this problem, |\foreach| offers a special |
| 1201 |
syntax: Instead of having one variable |\x|, Karl can specify two (or even |
| 1202 |
more) variables separated by a slash as in |\x / \xtext|. Then, the elements in |
| 1203 |
the set over which |\foreach| iterates must also be of the form |
| 1204 |
\meta{first}|/|\meta{second}. In each iteration, |\x| will be set to |
| 1205 |
\meta{first} and |\xtext| will be set to \meta{second}. If no \meta{second} is |
| 1206 |
given, the \meta{first} will be used again. So, here is the new code for the |
| 1207 |
ticks: |
| 1208 |
% |
| 1209 |
\begin{codeexample}[] |
| 1210 |
\begin{tikzpicture}[scale=3] |
| 1211 |
\clip (-0.6,-0.2) rectangle (0.6,1.51); |
| 1212 |
\draw[step=.5cm,help lines] (-1.4,-1.4) grid (1.4,1.4); |
| 1213 |
\filldraw[fill=green!20,draw=green!50!black] (0,0) -- (3mm,0mm) |
| 1214 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 1215 |
\draw[->] (-1.5,0) -- (1.5,0); \draw[->] (0,-1.5) -- (0,1.5); |
| 1216 |
\draw (0,0) circle [radius=1cm]; |
| 1217 |
|
| 1218 |
\foreach \x/\xtext in {-1, -0.5/-\frac{1}{2}, 1} |
| 1219 |
\draw (\x cm,1pt) -- (\x cm,-1pt) node[anchor=north] {$\xtext$}; |
| 1220 |
\foreach \y/\ytext in {-1, -0.5/-\frac{1}{2}, 0.5/\frac{1}{2}, 1} |
| 1221 |
\draw (1pt,\y cm) -- (-1pt,\y cm) node[anchor=east] {$\ytext$}; |
| 1222 |
\end{tikzpicture} |
| 1223 |
\end{codeexample} |
| 1224 |
|
| 1225 |
Karl is quite pleased with the result, but his son points out that this is |
| 1226 |
still not perfectly satisfactory: The grid and the circle interfere with the |
| 1227 |
numbers and decrease their legibility. Karl is not very concerned by this (his |
| 1228 |
students do not even notice), but his son insists that there is an easy |
| 1229 |
solution: Karl can add the |[fill=white]| option to fill out the background of |
| 1230 |
the text shape with a white color. |
| 1231 |
|
| 1232 |
The next thing Karl wants to do is to add the labels like $\sin \alpha$. For |
| 1233 |
this, he would like to place a label ``in the middle of the line''. To do so, |
| 1234 |
instead of specifying the label |node {$\sin\alpha$}| directly after one of |
| 1235 |
the endpoints of the line (which would place the label at that endpoint), Karl |
| 1236 |
can give the label directly after the |--|, before the coordinate. By default, |
| 1237 |
this places the label in the middle of the line, but the |pos=| options can be |
| 1238 |
used to modify this. Also, options like |near start| and |near end| can be used |
| 1239 |
to modify this position: |
| 1240 |
% |
| 1241 |
\begin{codeexample}[preamble={\usetikzlibrary{intersections}}] |
| 1242 |
\begin{tikzpicture}[scale=3] |
| 1243 |
\clip (-2,-0.2) rectangle (2,0.8); |
| 1244 |
\draw[step=.5cm,gray,very thin] (-1.4,-1.4) grid (1.4,1.4); |
| 1245 |
\filldraw[fill=green!20,draw=green!50!black] (0,0) -- (3mm,0mm) |
| 1246 |
arc [start angle=0, end angle=30, radius=3mm] -- cycle; |
| 1247 |
\draw[->] (-1.5,0) -- (1.5,0) coordinate (x axis); |
| 1248 |
\draw[->] (0,-1.5) -- (0,1.5) coordinate (y axis); |
| 1249 |
\draw (0,0) circle [radius=1cm]; |
| 1250 |
|
| 1251 |
\draw[very thick,red] |
| 1252 |
(30:1cm) -- node[left=1pt,fill=white] {$\sin \alpha$} (30:1cm |- x axis); |
| 1253 |
\draw[very thick,blue] |
| 1254 |
(30:1cm |- x axis) -- node[below=2pt,fill=white] {$\cos \alpha$} (0,0); |
| 1255 |
\path [name path=upward line] (1,0) -- (1,1); |
| 1256 |
\path [name path=sloped line] (0,0) -- (30:1.5cm); |
| 1257 |
\draw [name intersections={of=upward line and sloped line, by=t}] |
| 1258 |
[very thick,orange] (1,0) -- node [right=1pt,fill=white] |
| 1259 |
{$\displaystyle \tan \alpha \color{black}= |
| 1260 |
\frac{{\color{red}\sin \alpha}}{\color{blue}\cos \alpha}$} (t); |
| 1261 |
|
| 1262 |
\draw (0,0) -- (t); |
| 1263 |
|
| 1264 |
\foreach \x/\xtext in {-1, -0.5/-\frac{1}{2}, 1} |
| 1265 |
\draw (\x cm,1pt) -- (\x cm,-1pt) node[anchor=north,fill=white] {$\xtext$}; |
| 1266 |
\foreach \y/\ytext in {-1, -0.5/-\frac{1}{2}, 0.5/\frac{1}{2}, 1} |
| 1267 |
\draw (1pt,\y cm) -- (-1pt,\y cm) node[anchor=east,fill=white] {$\ytext$}; |
| 1268 |
\end{tikzpicture} |
| 1269 |
\end{codeexample} |
| 1270 |
|
| 1271 |
You can also position labels on curves and, by adding the |sloped| option, have |
| 1272 |
them rotated such that they match the line's slope. Here is an example: |
| 1273 |
% |
| 1274 |
\begin{codeexample}[] |
| 1275 |
\begin{tikzpicture} |
| 1276 |
\draw (0,0) .. controls (6,1) and (9,1) .. |
| 1277 |
node[near start,sloped,above] {near start} |
| 1278 |
node {midway} |
| 1279 |
node[very near end,sloped,below] {very near end} (12,0); |
| 1280 |
\end{tikzpicture} |
| 1281 |
\end{codeexample} |
| 1282 |
|
| 1283 |
It remains to draw the explanatory text at the right of the picture. The main |
| 1284 |
difficulty here lies in limiting the width of the text ``label'', which is |
| 1285 |
quite long, so that line breaking is used. Fortunately, Karl can use the option |
| 1286 |
|text width=6cm| to get the desired effect. So, here is the full code: |
| 1287 |
% |
| 1288 |
\begin{codeexample}[code only] |
| 1289 |
\begin{tikzpicture} |
| 1290 |
[scale=3,line cap=round, |
| 1291 |
% Styles |
| 1292 |
axes/.style=, |
| 1293 |
important line/.style={very thick}, |
| 1294 |
information text/.style={rounded corners,fill=red!10,inner sep=1ex}] |
| 1295 |
|
| 1296 |
% Colors |
| 1297 |
\colorlet{anglecolor}{green!50!black} |
| 1298 |
\colorlet{sincolor}{red} |
| 1299 |
\colorlet{tancolor}{orange!80!black} |
| 1300 |
\colorlet{coscolor}{blue} |
| 1301 |
|
| 1302 |
% The graphic |
| 1303 |
\draw[help lines,step=0.5cm] (-1.4,-1.4) grid (1.4,1.4); |
| 1304 |
|
| 1305 |
\draw (0,0) circle [radius=1cm]; |
| 1306 |
|
| 1307 |
\begin{scope}[axes] |
| 1308 |
\draw[->] (-1.5,0) -- (1.5,0) node[right] {$x$} coordinate(x axis); |
| 1309 |
\draw[->] (0,-1.5) -- (0,1.5) node[above] {$y$} coordinate(y axis); |
| 1310 |
|
| 1311 |
\foreach \x/\xtext in {-1, -.5/-\frac{1}{2}, 1} |
| 1312 |
\draw[xshift=\x cm] (0pt,1pt) -- (0pt,-1pt) node[below,fill=white] {$\xtext$}; |
| 1313 |
|
| 1314 |
\foreach \y/\ytext in {-1, -.5/-\frac{1}{2}, .5/\frac{1}{2}, 1} |
| 1315 |
\draw[yshift=\y cm] (1pt,0pt) -- (-1pt,0pt) node[left,fill=white] {$\ytext$}; |
| 1316 |
\end{scope} |
| 1317 |
|
| 1318 |
\filldraw[fill=green!20,draw=anglecolor] (0,0) -- (3mm,0pt) |
| 1319 |
arc [start angle=0, end angle=30, radius=3mm]; |
| 1320 |
\draw (15:2mm) node[anglecolor] {$\alpha$}; |
| 1321 |
|
| 1322 |
\draw[important line,sincolor] |
| 1323 |
(30:1cm) -- node[left=1pt,fill=white] {$\sin \alpha$} (30:1cm |- x axis); |
| 1324 |
|
| 1325 |
\draw[important line,coscolor] |
| 1326 |
(30:1cm |- x axis) -- node[below=2pt,fill=white] {$\cos \alpha$} (0,0); |
| 1327 |
|
| 1328 |
\path [name path=upward line] (1,0) -- (1,1); |
| 1329 |
\path [name path=sloped line] (0,0) -- (30:1.5cm); |
| 1330 |
\draw [name intersections={of=upward line and sloped line, by=t}] |
| 1331 |
[very thick,orange] (1,0) -- node [right=1pt,fill=white] |
| 1332 |
{$\displaystyle \tan \alpha \color{black}= |
| 1333 |
\frac{{\color{red}\sin \alpha}}{\color{blue}\cos \alpha}$} (t); |
| 1334 |
|
| 1335 |
\draw (0,0) -- (t); |
| 1336 |
|
| 1337 |
\draw[xshift=1.85cm] |
| 1338 |
node[right,text width=6cm,information text] |
| 1339 |
{ |
| 1340 |
The {\color{anglecolor} angle $\alpha$} is $30^\circ$ in the |
| 1341 |
example ($\pi/6$ in radians). The {\color{sincolor}sine of |
| 1342 |
$\alpha$}, which is the height of the red line, is |
| 1343 |
\[ |
| 1344 |
{\color{sincolor} \sin \alpha} = 1/2. |
| 1345 |
\] |
| 1346 |
By the Theorem of Pythagoras ... |
| 1347 |
}; |
| 1348 |
\end{tikzpicture} |
| 1349 |
\end{codeexample} |
| 1350 |
|
| 1351 |
|
| 1352 |
\subsection{Pics: The Angle Revisited} |
| 1353 |
|
| 1354 |
Karl expects that the code of certain parts of the picture he created might be |
| 1355 |
so useful that he might wish to reuse them in the future. A natural thing to do |
| 1356 |
is to create \TeX\ macros that store the code he wishes to reuse. However, |
| 1357 |
\tikzname\ offers another way that is integrated directly into its parser: |
| 1358 |
pics! |
| 1359 |
|
| 1360 |
A ``pic'' is ``not quite a full picture'', hence the short name. The idea is |
| 1361 |
that a pic is simply some code that you can add to a picture at different |
| 1362 |
places using the |pic| command whose syntax is almost identical to the |node| |
| 1363 |
command. The main difference is that instead of specifying some text in curly |
| 1364 |
braces that should be shown, you specify the name of a predefined picture that |
| 1365 |
should be shown. |
| 1366 |
|
| 1367 |
Defining new pics is easy enough, see Section~\ref{section-pics}, but right now |
| 1368 |
we just want to use one such predefined pic: the |angle| pic. As the name |
| 1369 |
suggests, it is a small drawing of an angle consisting of a little wedge and an |
| 1370 |
arc together with some text (Karl needs to load the |angles| library and the |
| 1371 |
|quotes| for the following examples). What makes this pic useful is the fact |
| 1372 |
that the size of the wedge will be computed automatically. |
| 1373 |
|
| 1374 |
The |angle| pic draws an angle between the two lines $BA$ and $BC$, where $A$, |
| 1375 |
$B$, and $C$ are three coordinates. In our case, $B$ is the origin, $A$ is |
| 1376 |
somewhere on the $x$-axis and $C$ is somewhere on a line at $30^\circ$. |
| 1377 |
% |
| 1378 |
\begin{codeexample}[preamble={\usetikzlibrary{angles,quotes}}] |
| 1379 |
\begin{tikzpicture}[scale=3] |
| 1380 |
\coordinate (A) at (1,0); |
| 1381 |
\coordinate (B) at (0,0); |
| 1382 |
\coordinate (C) at (30:1cm); |
| 1383 |
|
| 1384 |
\draw (A) -- (B) -- (C) |
| 1385 |
pic [draw=green!50!black, fill=green!20, angle radius=9mm, |
| 1386 |
"$\alpha$"] {angle = A--B--C}; |
| 1387 |
\end{tikzpicture} |
| 1388 |
\end{codeexample} |
| 1389 |
|
| 1390 |
Let us see, what is happening here. First we have specified three |
| 1391 |
\emph{coordinates} using the |\coordinate| command. It allows us to name a |
| 1392 |
specific coordinate in the picture. Then comes something that starts as a |
| 1393 |
normal |\draw|, but then comes the |pic| command. This command gets lots of |
| 1394 |
options and, in curly braces, comes the most important point: We specify that |
| 1395 |
we want to add an |angle| pic and this angle should be between the points we |
| 1396 |
named |A|, |B|, and |C| (we could use other names). Note that the text that we |
| 1397 |
want to be shown in the pic is specified in quotes inside the options of the |
| 1398 |
|pic|, not inside the curly braces. |
| 1399 |
|
| 1400 |
To learn more about pics, please see Section~\ref{section-pics}. |