| 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{Matrices and Alignment} |
| 12 |
\label{section-matrices} |
| 13 |
|
| 14 |
\subsection{Overview} |
| 15 |
|
| 16 |
When creating pictures, one often faces the problem of correctly aligning parts |
| 17 |
of the picture. For example, you might wish that the |baseline|s of certain |
| 18 |
nodes should be on the same line and some further nodes should be below these |
| 19 |
nodes with, say, their centers on a vertical lines. There are different ways of |
| 20 |
solving such problems. For example, by making clever use of anchors, nearly all |
| 21 |
such alignment problems can be solved. However, this often leads to complicated |
| 22 |
code. An often simpler way is to use \emph{matrices}, the use of which is |
| 23 |
explained in the current section. |
| 24 |
|
| 25 |
A \tikzname\ matrix is similar to \LaTeX's |{tabular}| or |{array}| |
| 26 |
environment, only instead of text each cell contains a little picture or a |
| 27 |
node. The sizes of the cells are automatically adjusted such that they are |
| 28 |
large enough to contain all the cell contents. |
| 29 |
|
| 30 |
Matrices are a powerful tool and they need to be handled with some care. For |
| 31 |
impatient readers who skip the rest of this section: you \emph{must} end |
| 32 |
\emph{every} row with |\\|. In particular, the last row \emph{must} be ended |
| 33 |
with |\\|. |
| 34 |
|
| 35 |
Many of the ideas implemented in \tikzname's matrix support are due to Mark |
| 36 |
Wibrow -- many thanks to Mark at this point! |
| 37 |
|
| 38 |
|
| 39 |
\subsection{Matrices are Nodes} |
| 40 |
|
| 41 |
Matrices are special in many ways, but for most purposes matrices are treated |
| 42 |
like nodes. This means, that you use the |node| path command to create a matrix |
| 43 |
and you only use a special option, namely the |matrix| option, to signal that |
| 44 |
the node will contain a matrix. Instead of the usual \TeX-box that makes up the |
| 45 |
|text| part of the node's shape, the matrix is used. Thus, in particular, a |
| 46 |
matrix can have a shape, this shape can be drawn or filled, it can be used in a |
| 47 |
tree, and so on. Also, you can refer to the different anchors of a matrix. |
| 48 |
|
| 49 |
\begin{key}{/tikz/matrix=\meta{true or false} (default true)} |
| 50 |
This option can be passed to a |node| path command. It signals that the |
| 51 |
node will contain a matrix. |
| 52 |
% |
| 53 |
\begin{codeexample}[] |
| 54 |
\begin{tikzpicture} |
| 55 |
\draw[help lines] (0,0) grid (4,2); |
| 56 |
\node [matrix,fill=red!20,draw=blue,very thick] (my matrix) at (2,1) |
| 57 |
{ |
| 58 |
\draw (0,0) circle (4mm); & \node[rotate=10] {Hello}; \\ |
| 59 |
\draw (0.2,0) circle (2mm); & \fill[red] (0,0) circle (3mm); \\ |
| 60 |
}; |
| 61 |
|
| 62 |
\draw [very thick,->] (0,0) |- (my matrix.west); |
| 63 |
\end{tikzpicture} |
| 64 |
\end{codeexample} |
| 65 |
% |
| 66 |
The exact syntax of the matrix is explained in the course of this section. |
| 67 |
% |
| 68 |
\begin{stylekey}{/tikz/every matrix (initially \normalfont empty)} |
| 69 |
This style is used in every matrix. |
| 70 |
\end{stylekey} |
| 71 |
% |
| 72 |
\begin{stylekey}{/tikz/every outer matrix (initially \normalfont empty)} |
| 73 |
While the |every matrix| key also applies to the matrix contents, this |
| 74 |
only applies to the outer node which holds the matrix. |
| 75 |
\end{stylekey} |
| 76 |
\end{key} |
| 77 |
|
| 78 |
Even more so than nodes, matrices will often be the only object on a path. |
| 79 |
Because of this, there is a special abbreviation for creating matrices: |
| 80 |
|
| 81 |
\begin{command}{\matrix} |
| 82 |
Inside |{tikzpicture}| this is an abbreviation for |\path node[matrix]|. |
| 83 |
\end{command} |
| 84 |
|
| 85 |
Even though matrices are nodes, some options do not have the same effect as for |
| 86 |
normal nodes: |
| 87 |
% |
| 88 |
\begin{enumerate} |
| 89 |
\item Rotations and scaling have no effect on a matrix as a whole (however, |
| 90 |
you can still transform the contents of the cells normally). Before the |
| 91 |
matrix is typeset, the rotational and scaling part of the |
| 92 |
transformation matrix is reset. |
| 93 |
\item For multi-part shapes you can only set the |text| part of the node. |
| 94 |
\item All options starting with |text| such as |text width| have no effect. |
| 95 |
\item If you place a matrix on a path, the matrix contents will be |
| 96 |
collected into a macro, which tokenizes them. This means that |&| will |
| 97 |
lose its meaning as an alignment character, resulting in an error. If |
| 98 |
you need to place a matrix on a path, use |ampersand replacement| to |
| 99 |
work around that problem. |
| 100 |
\end{enumerate} |
| 101 |
|
| 102 |
|
| 103 |
\subsection{Cell Pictures} |
| 104 |
\label{section-tikz-cell-pictures} |
| 105 |
|
| 106 |
A matrix consists of rows of \emph{cells}. Each row (including the last one!) |
| 107 |
is ended by the command |\\|. The character |&| is used to separate cells. |
| 108 |
Inside each cell, you must place commands for drawing a picture, called the |
| 109 |
\emph{cell picture} in the following. (However, cell pictures are not enclosed |
| 110 |
in a complete |{pgfpicture}| environment, they are a bit more light-weight. The |
| 111 |
main difference is that cell pictures cannot have layers.) It is not necessary |
| 112 |
to specify beforehand how many rows or columns there are going to be and if a |
| 113 |
row contains less cell pictures than another line, empty cells are |
| 114 |
automatically added as needed. |
| 115 |
|
| 116 |
|
| 117 |
\subsubsection{Alignment of Cell Pictures} |
| 118 |
|
| 119 |
For each cell picture a bounding box is computed. These bounding boxes and the |
| 120 |
origins of the cell pictures determine how the cells are aligned. Let us start |
| 121 |
with the rows: Consider the cell pictures on the first row. Each has a bounding |
| 122 |
box and somewhere inside this bounding box the origin of the cell picture can |
| 123 |
be found (the origin might even lie outside the bounding box, but let us ignore |
| 124 |
this problem for the moment). The cell pictures are then shifted around such |
| 125 |
that all origins lie on the same horizontal line. This may make it necessary to |
| 126 |
shift some cell pictures upwards and others downwards, but it can be done and |
| 127 |
this yields the vertical alignment of the cell pictures this row. The top of |
| 128 |
the row is then given by the top of the ``highest'' cell picture in the row, |
| 129 |
the bottom of the row is given by the bottom of the lowest cell picture. (To be |
| 130 |
more precise, the height of the row is the maximum $y$-value of any of the |
| 131 |
bounding boxes and the depth of the row is the negated minimum $y$-value of the |
| 132 |
bounding boxes). |
| 133 |
% |
| 134 |
\begin{codeexample}[] |
| 135 |
\begin{tikzpicture} |
| 136 |
[every node/.style={draw=black,anchor=base,font=\huge}] |
| 137 |
|
| 138 |
\matrix [draw=red] |
| 139 |
{ |
| 140 |
\node {a}; \fill[blue] (0,0) circle (2pt); & |
| 141 |
\node {X}; \fill[blue] (0,0) circle (2pt); & |
| 142 |
\node {g}; \fill[blue] (0,0) circle (2pt); \\ |
| 143 |
}; |
| 144 |
\end{tikzpicture} |
| 145 |
\end{codeexample} |
| 146 |
|
| 147 |
Each row is aligned in this fashion: For each row the cell pictures |
| 148 |
are vertically aligned such that the origins lie on the same |
| 149 |
line. Then the second row is placed below the first row such that the |
| 150 |
bottom of the first row touches the top of the second row (unless a |
| 151 |
|row sep| is used to add a bit of space). Then the bottom of the |
| 152 |
second row touches the top of the third row, and so on. Typically, |
| 153 |
each row will have an individual height and depth. |
| 154 |
% |
| 155 |
\begin{codeexample}[] |
| 156 |
\begin{tikzpicture} |
| 157 |
[every node/.style={draw=black,anchor=base}] |
| 158 |
|
| 159 |
\matrix [draw=red] |
| 160 |
{ |
| 161 |
\node {a}; & \node {X}; & \node {g}; \\ |
| 162 |
\node {a}; & \node {X}; & \node {g}; \\ |
| 163 |
}; |
| 164 |
|
| 165 |
\matrix [row sep=3mm,draw=red] at (0,-2) |
| 166 |
{ |
| 167 |
\node {a}; & \node {X}; & \node {g}; \\ |
| 168 |
\node {a}; & \node {X}; & \node {g}; \\ |
| 169 |
}; |
| 170 |
\end{tikzpicture} |
| 171 |
\end{codeexample} |
| 172 |
|
| 173 |
Let us now have a look at the columns. The rules for how the pictures on any |
| 174 |
given column are aligned are very similar to the row alignment: Consider all |
| 175 |
cell pictures in the first column. Each is shifted horizontally such that the |
| 176 |
origins lie on the same vertical line. Then, the left end of the column is at |
| 177 |
the left end of the bounding box that protrudes furthest to the left. The right |
| 178 |
end of the column is at the right end of the bounding box that protrudes |
| 179 |
furthest to the right. This fixes the horizontal alignment of the cell pictures |
| 180 |
in the first column and the same happens the cell pictures in the other |
| 181 |
columns. Then, the right end of the first column touches the left end of the |
| 182 |
second column (unless |column sep| is used). The right end of the second column |
| 183 |
touches the left end of the third column, and so on. (Internally, two columns |
| 184 |
are actually used to achieve the desired horizontal alignment, but that is only |
| 185 |
an implementation detail.) |
| 186 |
% |
| 187 |
\begin{codeexample}[] |
| 188 |
\begin{tikzpicture}[every node/.style={draw}] |
| 189 |
\matrix [draw=red] |
| 190 |
{ |
| 191 |
\node[left] {Hallo}; \fill[blue] (0,0) circle (2pt); \\ |
| 192 |
\node {X}; \fill[blue] (0,0) circle (2pt); \\ |
| 193 |
\node[right] {g}; \fill[blue] (0,0) circle (2pt); \\ |
| 194 |
}; |
| 195 |
\end{tikzpicture} |
| 196 |
\end{codeexample} |
| 197 |
|
| 198 |
\begin{codeexample}[] |
| 199 |
\begin{tikzpicture}[every node/.style={draw}] |
| 200 |
\matrix [draw=red,column sep=1cm] |
| 201 |
{ |
| 202 |
\node {8}; & \node{1}; & \node {6}; \\ |
| 203 |
\node {3}; & \node{5}; & \node {7}; \\ |
| 204 |
\node {4}; & \node{9}; & \node {2}; \\ |
| 205 |
}; |
| 206 |
\end{tikzpicture} |
| 207 |
\end{codeexample} |
| 208 |
|
| 209 |
|
| 210 |
\subsubsection{Setting and Adjusting Column and Row Spacing} |
| 211 |
|
| 212 |
There are different ways of setting and adjusting the spacing between columns |
| 213 |
and rows. First, you can use the options |column sep| and |row sep| to set a |
| 214 |
default spacing for all rows and all columns. Second, you can add options to |
| 215 |
the |&| character and the |\\| command to adjust the spacing between two |
| 216 |
specific columns or rows. Additionally, you can specify whether the space |
| 217 |
between two columns or rows should be considered between the origins of cells |
| 218 |
in the column or row or between their borders. |
| 219 |
|
| 220 |
\begin{key}{/tikz/column sep=\meta{spacing list}} |
| 221 |
This option sets a default space that is added between every two columns. |
| 222 |
This space can be positive or negative and is zero by default. The |
| 223 |
\meta{spacing list} normally contains a single dimension like |2pt|. |
| 224 |
% |
| 225 |
\begin{codeexample}[] |
| 226 |
\begin{tikzpicture} |
| 227 |
\matrix [draw,column sep=1cm,nodes=draw] |
| 228 |
{ |
| 229 |
\node(a) {123}; & \node (b) {1}; & \node {1}; \\ |
| 230 |
\node {12}; & \node {12}; & \node {1}; \\ |
| 231 |
\node(c) {1}; & \node (d) {123}; & \node {1}; \\ |
| 232 |
}; |
| 233 |
\draw [red,thick] (a.east) -- (a.east |- c) |
| 234 |
(d.west) -- (d.west |- b); |
| 235 |
\draw [<->,red,thick] (a.east) -- (d.west |- b) |
| 236 |
node [above,midway] {1cm}; |
| 237 |
\end{tikzpicture} |
| 238 |
\end{codeexample} |
| 239 |
% |
| 240 |
More generally, the \meta{spacing list} may contain a whole list of |
| 241 |
numbers, separated by commas, and occurrences of the two key words |
| 242 |
|between origins| and |between borders|. The effect of specifying such a |
| 243 |
list is the following: First, all numbers occurring in the list are simply |
| 244 |
added to compute the final spacing. Second, concerning the two keywords, |
| 245 |
the last occurrence of one of the keywords is important. If the last |
| 246 |
occurrence is |between borders| or if neither occurs, then the space is |
| 247 |
inserted between the two columns normally. However, if the last occurs is |
| 248 |
|between origins|, then the following happens: The distance between the |
| 249 |
columns is adjusted such that the difference between the origins of all the |
| 250 |
cells in the first column (remember that they all lie on straight line) and |
| 251 |
the origins of all the cells in the second column is exactly the given |
| 252 |
distance. |
| 253 |
|
| 254 |
\emph{The }|between origins|\emph{ option can only be used for columns |
| 255 |
mentioned in the first row, that is, you cannot specify this option for |
| 256 |
columns introduced only in later rows.} |
| 257 |
% |
| 258 |
\begin{codeexample}[] |
| 259 |
\begin{tikzpicture} |
| 260 |
\matrix [draw,column sep={1cm,between origins},nodes=draw] |
| 261 |
{ |
| 262 |
\node(a) {123}; & \node (b) {1}; & \node {1}; \\ |
| 263 |
\node {12}; & \node {12}; & \node {1}; \\ |
| 264 |
\node {1}; & \node {123}; & \node {1}; \\ |
| 265 |
}; |
| 266 |
\draw [<->,red,thick] (a.center) -- (b.center) node [above,midway] {1cm}; |
| 267 |
\end{tikzpicture} |
| 268 |
\end{codeexample} |
| 269 |
% |
| 270 |
\end{key} |
| 271 |
|
| 272 |
\begin{key}{/tikz/row sep=\meta{spacing list}} |
| 273 |
This option works like |column sep|, only for rows. Here, too, you can |
| 274 |
specify whether the space is added between the lower end of the first row |
| 275 |
and the upper end of the second row, or whether the space is computed |
| 276 |
between the origins of the two rows. |
| 277 |
% |
| 278 |
\begin{codeexample}[] |
| 279 |
\begin{tikzpicture} |
| 280 |
\matrix [draw,row sep=1cm,nodes=draw] |
| 281 |
{ |
| 282 |
\node (a) {123}; & \node {1}; & \node {1}; \\ |
| 283 |
\node (b) {12}; & \node {12}; & \node {1}; \\ |
| 284 |
\node {1}; & \node {123}; & \node {1}; \\ |
| 285 |
}; |
| 286 |
\draw [<->,red,thick] (a.south) -- (b.north) node [right,midway] {1cm}; |
| 287 |
\end{tikzpicture} |
| 288 |
\end{codeexample} |
| 289 |
% |
| 290 |
\begin{codeexample}[] |
| 291 |
\begin{tikzpicture} |
| 292 |
\matrix [draw,row sep={1cm,between origins},nodes=draw] |
| 293 |
{ |
| 294 |
\node (a) {123}; & \node {1}; & \node {1}; \\ |
| 295 |
\node (b) {12}; & \node {12}; & \node {1}; \\ |
| 296 |
\node {1}; & \node {123}; & \node {1}; \\ |
| 297 |
}; |
| 298 |
\draw [<->,red,thick] (a.center) -- (b.center) node [right,midway] {1cm}; |
| 299 |
\end{tikzpicture} |
| 300 |
\end{codeexample} |
| 301 |
% |
| 302 |
\end{key} |
| 303 |
|
| 304 |
The row-end command |\\| allows you to provide an optional argument, which must |
| 305 |
be a dimension. This dimension will be added to the list in |row sep|. This |
| 306 |
means that, firstly, any numbers you list in this argument will be added as an |
| 307 |
extra row separation between the line being ended and the next line and, |
| 308 |
secondly, you can use the keywords |between origins| and |between borders| to |
| 309 |
locally overrule the standard setting for this line pair. |
| 310 |
% |
| 311 |
\begin{codeexample}[] |
| 312 |
\begin{tikzpicture} |
| 313 |
\matrix [row sep=1mm] |
| 314 |
{ |
| 315 |
\draw (0,0) circle (2mm); & \draw (0,0) circle (2mm); \\ |
| 316 |
\draw (0,0) circle (2mm); & \draw (0,0) circle (2mm); \\[-1mm] |
| 317 |
\draw (0,0) coordinate (a) circle (2mm); & |
| 318 |
\draw (0,0) circle (2mm); \\[1cm,between origins] |
| 319 |
\draw (0,0) coordinate (b) circle (2mm); & |
| 320 |
\draw (0,0) circle (2mm); \\ |
| 321 |
}; |
| 322 |
\draw [<->,red,thick] (a.center) -- (b.center) node [right,midway] {1cm}; |
| 323 |
\end{tikzpicture} |
| 324 |
\end{codeexample} |
| 325 |
|
| 326 |
The cell separation character |&| also takes an optional argument, which must |
| 327 |
also be a spacing list. This spacing list is added to the |column sep| having a |
| 328 |
similar effect as the option for the |\\| command for rows. |
| 329 |
|
| 330 |
This optional spacing list can only be given the first time a new column is |
| 331 |
started (usually in the first row), subsequent usages of this option in later |
| 332 |
rows have no effect. |
| 333 |
% |
| 334 |
\begin{codeexample}[] |
| 335 |
\begin{tikzpicture} |
| 336 |
\matrix [draw,nodes=draw,column sep=1mm] |
| 337 |
{ |
| 338 |
\node {8}; &[2mm] \node{1}; &[-1mm] \node {6}; \\ |
| 339 |
\node {3}; & \node{5}; & \node {7}; \\ |
| 340 |
\node {4}; & \node{9}; & \node {2}; \\ |
| 341 |
}; |
| 342 |
\end{tikzpicture} |
| 343 |
\end{codeexample} |
| 344 |
% |
| 345 |
\begin{codeexample}[] |
| 346 |
\begin{tikzpicture} |
| 347 |
\matrix [draw,nodes=draw,column sep=1mm] |
| 348 |
{ |
| 349 |
\node {8}; &[2mm] \node(a){1}; &[1cm,between origins] \node(b){6}; \\ |
| 350 |
\node {3}; & \node {5}; & \node {7}; \\ |
| 351 |
\node {4}; & \node {9}; & \node {2}; \\ |
| 352 |
}; |
| 353 |
\draw [<->,red,thick] (a.center) -- (b.center) node [above,midway] {11mm}; |
| 354 |
\end{tikzpicture} |
| 355 |
\end{codeexample} |
| 356 |
% |
| 357 |
\begin{codeexample}[] |
| 358 |
\begin{tikzpicture} |
| 359 |
\matrix [draw,nodes=draw,column sep={1cm,between origins}] |
| 360 |
{ |
| 361 |
\node (a) {8}; & \node (b) {1}; &[between borders] \node (c) {6}; \\ |
| 362 |
\node {3}; & \node {5}; & \node {7}; \\ |
| 363 |
\node {4}; & \node {9}; & \node {2}; \\ |
| 364 |
}; |
| 365 |
\draw [<->,red,thick] (a.center) -- (b.center) node [above,midway] {10mm}; |
| 366 |
\draw [<->,red,thick] (b.east) -- (c.west) node [above,midway] {10mm}; |
| 367 |
\end{tikzpicture} |
| 368 |
\end{codeexample} |
| 369 |
|
| 370 |
|
| 371 |
\subsubsection{Cell Styles and Options} |
| 372 |
|
| 373 |
The following styles and options are useful for changing the appearance of all |
| 374 |
cell pictures: |
| 375 |
|
| 376 |
\begin{stylekey}{/tikz/every cell=\marg{row}\marg{column} (initially \normalfont empty)} |
| 377 |
This style is installed at the beginning of each cell picture with the two |
| 378 |
parameters being the current \meta{row} and \meta{column} of the cell. Note |
| 379 |
that setting this style to |draw| will \emph{not} cause all nodes to be |
| 380 |
drawn since the |draw| option has to be passed to each node individually. |
| 381 |
|
| 382 |
Inside this style (and inside all cells), the current \meta{row} and |
| 383 |
\meta{column} number are also accessible via the counters |
| 384 |
|\pgfmatrixcurrentrow| and |\pgfmatrixcurrentcolumn|. |
| 385 |
\end{stylekey} |
| 386 |
|
| 387 |
\begin{key}{/tikz/cells=\meta{options}} |
| 388 |
This key adds the \meta{options} to the style |every cell|. It is mainly |
| 389 |
just a shorthand for the code |every cell/.append style=|\meta{options}. |
| 390 |
\end{key} |
| 391 |
|
| 392 |
\begin{key}{/tikz/nodes=\meta{options}} |
| 393 |
This key adds the \meta{options} to the style |every node|. It is mainly |
| 394 |
just a shorthand for the code |every node/.append style=|\meta{options}. |
| 395 |
|
| 396 |
The main use of this option is the install some options for the nodes |
| 397 |
\emph{inside} the matrix that should not apply to the matrix \emph{itself}. |
| 398 |
% |
| 399 |
\begin{codeexample}[] |
| 400 |
\begin{tikzpicture} |
| 401 |
\matrix [nodes={fill=blue!20,minimum size=5mm}] |
| 402 |
{ |
| 403 |
\node {8}; & \node{1}; & \node {6}; \\ |
| 404 |
\node {3}; & \node{5}; & \node {7}; \\ |
| 405 |
\node {4}; & \node{9}; & \node {2}; \\ |
| 406 |
}; |
| 407 |
\end{tikzpicture} |
| 408 |
\end{codeexample} |
| 409 |
% |
| 410 |
\end{key} |
| 411 |
|
| 412 |
The next set of styles can be used to change the appearance of certain rows, |
| 413 |
columns, or cells. If more than one of these styles is defined, they are |
| 414 |
executed in the below order (the |every cell| style is executed before all of |
| 415 |
the below). |
| 416 |
% |
| 417 |
\begin{stylekey}{/tikz/column \meta{number}} |
| 418 |
This style is used for every cell in column \meta{number}. |
| 419 |
\end{stylekey} |
| 420 |
|
| 421 |
\begin{stylekey}{/tikz/every odd column} |
| 422 |
This style is used for every cell in an odd column. |
| 423 |
\end{stylekey} |
| 424 |
|
| 425 |
\begin{stylekey}{/tikz/every even column} |
| 426 |
This style is used for every cell in an even column. |
| 427 |
\end{stylekey} |
| 428 |
|
| 429 |
\begin{stylekey}{/tikz/row \meta{number}} |
| 430 |
This style is used for every cell in row \meta{number}. |
| 431 |
\end{stylekey} |
| 432 |
|
| 433 |
\begin{stylekey}{/tikz/every odd row} |
| 434 |
This style is used for every cell in an odd row. |
| 435 |
\end{stylekey} |
| 436 |
|
| 437 |
\begin{stylekey}{/tikz/every even row} |
| 438 |
This style is used for every cell in an even row. |
| 439 |
\end{stylekey} |
| 440 |
|
| 441 |
\begin{stylekey}{/tikz/row \meta{row number} column \meta{column number}} |
| 442 |
This style is used for the cell in row \meta{row number} and column |
| 443 |
\meta{column number}. |
| 444 |
\end{stylekey} |
| 445 |
% |
| 446 |
\begin{codeexample}[] |
| 447 |
\begin{tikzpicture} |
| 448 |
[row 1/.style={red}, |
| 449 |
column 2/.style={green!50!black}, |
| 450 |
row 3 column 3/.style={blue}] |
| 451 |
|
| 452 |
\matrix |
| 453 |
{ |
| 454 |
\node {8}; & \node{1}; & \node {6}; \\ |
| 455 |
\node {3}; & \node{5}; & \node {7}; \\ |
| 456 |
\node {4}; & \node{9}; & \node {2}; \\ |
| 457 |
}; |
| 458 |
\end{tikzpicture} |
| 459 |
\end{codeexample} |
| 460 |
|
| 461 |
You can use the |column |\meta{number} option to change the alignment for |
| 462 |
different columns. |
| 463 |
% |
| 464 |
\begin{codeexample}[] |
| 465 |
\begin{tikzpicture} |
| 466 |
[column 1/.style={anchor=base west}, |
| 467 |
column 2/.style={anchor=base east}, |
| 468 |
column 3/.style={anchor=base}] |
| 469 |
\matrix |
| 470 |
{ |
| 471 |
\node {123}; & \node{456}; & \node {789}; \\ |
| 472 |
\node {12}; & \node{45}; & \node {78}; \\ |
| 473 |
\node {1}; & \node{4}; & \node {7}; \\ |
| 474 |
}; |
| 475 |
\end{tikzpicture} |
| 476 |
\end{codeexample} |
| 477 |
|
| 478 |
In some cases, it is desirable to include some automation in each column/row |
| 479 |
separately. A typical example is to apply stripe-pattern to almost all columns |
| 480 |
with exceptions. For these type of use-cases, nesting these keys can open up |
| 481 |
a lot of possibilities; in the following example a ``feature comparison'' table |
| 482 |
is demonstrated. It is intentionally made rather verbose and a bit redundant |
| 483 |
to show how the column and row settings can be progressively overwritten to |
| 484 |
create certain effects. |
| 485 |
|
| 486 |
\begin{codeexample}[preamble={\usetikzlibrary{matrix,fit}}] |
| 487 |
\begin{tikzpicture}[ |
| 488 |
font=\sffamily, |
| 489 |
head color/.style args={#1/#2}{ |
| 490 |
row 1 column #1/.append style={nodes={fill=#2}}}, |
| 491 |
% swap order of row and column styles |
| 492 |
matrix/inner style order={ |
| 493 |
every cell, |
| 494 |
row, even odd row, |
| 495 |
column, even odd column, |
| 496 |
cell |
| 497 |
} |
| 498 |
] |
| 499 |
|
| 500 |
\matrix [ |
| 501 |
matrix of nodes, nodes in empty cells, |
| 502 |
nodes={text width=2cm, align=center, |
| 503 |
minimum height=1.5em, anchor=center}, |
| 504 |
% add striped row style |
| 505 |
every even row/.style={nodes={fill=olive!50}}, |
| 506 |
% modify the feature column and header row |
| 507 |
column 1/.style= {nodes={fill=olive, inner ysep=0}}, |
| 508 |
row 1/.style= {nodes={text depth=0.2ex, text=white}}, |
| 509 |
row 1 column 1/.style={nodes={fill=none, draw=none}}, |
| 510 |
head color/.list={2/orange,3/teal,4/cyan,5/magenta} % specify header colors |
| 511 |
] (m) |
| 512 |
{ |
| 513 |
& Basic & Standard & Professional & Enterprise \\ |
| 514 |
Feature A & $\bullet$ & $\bullet$ & $\bullet$ & $\bullet$ \\ |
| 515 |
Feature B & $\bullet$ & $\bullet$ & $\bullet$ & $\bullet$ \\ |
| 516 |
Feature C & & & & $\bullet$ \\ |
| 517 |
Feature D & & $\bullet$ & $\bullet$ & $\bullet$ \\ |
| 518 |
Feature E & & & $\bullet$ & $\bullet$ \\ |
| 519 |
}; |
| 520 |
% Add emphasis on selection by the use of "fit" library |
| 521 |
\node[fit={(m-1-4.north west) (m-6-4.south east)}, |
| 522 |
ultra thick, inner sep=0pt, rounded corners=1mm, |
| 523 |
draw=cyan, label={[cyan,align=center]270:Popular\\Choice!}]{}; |
| 524 |
\end{tikzpicture} |
| 525 |
\end{codeexample} |
| 526 |
|
| 527 |
The order in which these styles are applied is configurable. You can also |
| 528 |
install your own styles. The following styles (in fact, internally they are |
| 529 |
|/.code| keys) wrap the styles introduced in the previous paragraph passing the |
| 530 |
correct argument and ensuring that they are only called for even or odd rows. |
| 531 |
However, it is not recommended to override these. |
| 532 |
|
| 533 |
\begin{stylekey}{/tikz/matrix/inner style/every cell} |
| 534 |
Wraps |/tikz/every cell|. |
| 535 |
\end{stylekey} |
| 536 |
\begin{stylekey}{/tikz/matrix/inner style/column} |
| 537 |
Wraps |/tikz/column |\meta{number}. |
| 538 |
\end{stylekey} |
| 539 |
\begin{stylekey}{/tikz/matrix/inner style/even odd column} |
| 540 |
Wraps |/tikz/every even column| and |/tikz/every odd column|. |
| 541 |
\end{stylekey} |
| 542 |
\begin{stylekey}{/tikz/matrix/inner style/row} |
| 543 |
Wraps |/tikz/row |\meta{number}. |
| 544 |
\end{stylekey} |
| 545 |
\begin{stylekey}{/tikz/matrix/inner style/even odd row} |
| 546 |
Wraps |/tikz/every even row| and |/tikz/every odd row|. |
| 547 |
\end{stylekey} |
| 548 |
\begin{stylekey}{/tikz/matrix/inner style/cell} |
| 549 |
Wraps |/tikz/row |\meta{number}| column |\meta{number}. |
| 550 |
\end{stylekey} |
| 551 |
|
| 552 |
\begin{stylekey}{/tikz/matrix/inner style order} |
| 553 |
The order in which these styles are applied to the matrix cells is |
| 554 |
specified by this key. By default it is |
| 555 |
% |
| 556 |
\begin{codeexample}[code only] |
| 557 |
\tikzset{ |
| 558 |
matrix/inner style order={ |
| 559 |
every cell, |
| 560 |
column, |
| 561 |
even odd column, |
| 562 |
row, |
| 563 |
even odd row, |
| 564 |
cell, |
| 565 |
}, |
| 566 |
} |
| 567 |
\end{codeexample} |
| 568 |
% |
| 569 |
You can use this to install your own styles here, but only \emph{names} of |
| 570 |
styles are permitted here. The style specification has to be placed |
| 571 |
outside of |matrix/inner style order| and unless it is installed inside |
| 572 |
|/tikz/matrix/inner style/|, it has to be fully qualified. |
| 573 |
% |
| 574 |
\begin{codeexample}[code only] |
| 575 |
\tikzset{ |
| 576 |
my style/.code={% |
| 577 |
\ifnum\pgfmatrixcurrentcolumn=2 |
| 578 |
\tikzset{font=\itshape}% |
| 579 |
\fi |
| 580 |
}, |
| 581 |
matrix/inner style order={ |
| 582 |
every cell, |
| 583 |
even odd column, |
| 584 |
even odd row, |
| 585 |
column, |
| 586 |
row, |
| 587 |
cell, |
| 588 |
/tikz/my style, |
| 589 |
}, |
| 590 |
} |
| 591 |
\end{codeexample} |
| 592 |
% |
| 593 |
\end{stylekey} |
| 594 |
|
| 595 |
In many matrices all cell pictures have nearly the same code. For example, |
| 596 |
cells typically start with |\node{| and end |};|. The following options allow |
| 597 |
you to execute such code in all cells: |
| 598 |
|
| 599 |
\begin{key}{/tikz/execute at begin cell=\meta{code}} |
| 600 |
The code will be executed at the beginning of each nonempty cell. |
| 601 |
\end{key} |
| 602 |
% |
| 603 |
\begin{key}{/tikz/execute at end cell=\meta{code}} |
| 604 |
The code will be executed at the end of each nonempty cell. |
| 605 |
\end{key} |
| 606 |
% |
| 607 |
\begin{key}{/tikz/execute at empty cell=\meta{code}} |
| 608 |
The code will be executed inside each empty cell. |
| 609 |
\end{key} |
| 610 |
% |
| 611 |
\begin{codeexample}[] |
| 612 |
\begin{tikzpicture} |
| 613 |
[matrix of nodes/.style={ |
| 614 |
execute at begin cell=\node\bgroup, |
| 615 |
execute at end cell=\egroup;% |
| 616 |
}] |
| 617 |
\matrix [matrix of nodes] |
| 618 |
{ |
| 619 |
8 & 1 & 6 \\ |
| 620 |
3 & 5 & 7 \\ |
| 621 |
4 & 9 & 2 \\ |
| 622 |
}; |
| 623 |
\end{tikzpicture} |
| 624 |
\end{codeexample} |
| 625 |
% |
| 626 |
\begin{codeexample}[] |
| 627 |
\begin{tikzpicture} |
| 628 |
[matrix of nodes/.style={ |
| 629 |
execute at begin cell=\node\bgroup, |
| 630 |
execute at end cell=\egroup;,% |
| 631 |
execute at empty cell=\node{--};% |
| 632 |
}] |
| 633 |
\matrix [matrix of nodes] |
| 634 |
{ |
| 635 |
8 & 1 & \\ |
| 636 |
3 & & 7 \\ |
| 637 |
& & 2 \\ |
| 638 |
}; |
| 639 |
\end{tikzpicture} |
| 640 |
\end{codeexample} |
| 641 |
|
| 642 |
The |matrix| library defines a number of styles that make use of the above |
| 643 |
options. |
| 644 |
|
| 645 |
|
| 646 |
\subsection{Anchoring a Matrix} |
| 647 |
|
| 648 |
Since matrices are nodes, they can be anchored in the usual fashion using the |
| 649 |
|anchor| option. However, there are two ways to influence this placement |
| 650 |
further. First, the following option is often useful: |
| 651 |
|
| 652 |
\begin{key}{/tikz/matrix anchor=\meta{anchor}} |
| 653 |
This option has the same effect as |anchor|, but the option applies only to |
| 654 |
the matrix itself, not to the cells inside. If you just say |anchor=north| |
| 655 |
as an option to the matrix node, all nodes inside matrix will also have |
| 656 |
this anchor, unless it is explicitly set differently for each node. By |
| 657 |
comparison, |matrix anchor| sets the anchor for the matrix, but for the |
| 658 |
nodes inside the value of |anchor| remain unchanged. |
| 659 |
% |
| 660 |
\begin{codeexample}[] |
| 661 |
\begin{tikzpicture} |
| 662 |
\matrix [matrix anchor=west] at (0,0) |
| 663 |
{ |
| 664 |
\node {123}; \\ % still center anchor |
| 665 |
\node {12}; \\ |
| 666 |
\node {1}; \\ |
| 667 |
}; |
| 668 |
\matrix [anchor=west] at (0,-2) |
| 669 |
{ |
| 670 |
\node {123}; \\ % inherited west anchor |
| 671 |
\node {12}; \\ |
| 672 |
\node {1}; \\ |
| 673 |
}; |
| 674 |
\end{tikzpicture} |
| 675 |
\end{codeexample} |
| 676 |
% |
| 677 |
\end{key} |
| 678 |
|
| 679 |
The second way to anchor a matrix is to use \emph{an anchor of a node inside |
| 680 |
the matrix}. For this, the |anchor| option has a special effect when given as |
| 681 |
an argument to a matrix: |
| 682 |
|
| 683 |
\begin{key}{/tikz/anchor=\meta{anchor or node.anchor}} |
| 684 |
Normally, the argument of this option refers to anchor of the matrix node, |
| 685 |
which is the node that includes all of the stuff of the matrix. However, |
| 686 |
you can also provide an argument of the form \meta{node}|.|\meta{anchor} |
| 687 |
where \meta{node} must be node defined inside the matrix and \meta{anchor} |
| 688 |
is an anchor of this node. In this case, the whole matrix is shifted around |
| 689 |
in such a way that this particular anchor of this particular node lies at |
| 690 |
the |at| position of the matrix. The same is true for |matrix anchor|. |
| 691 |
% |
| 692 |
\begin{codeexample}[] |
| 693 |
\begin{tikzpicture} |
| 694 |
\draw[help lines] (0,0) grid (3,2); |
| 695 |
\matrix[matrix anchor=inner node.south,anchor=base,row sep=3mm] at (1,1) |
| 696 |
{ |
| 697 |
\node {a}; & \node {b}; & \node {c}; & \node {d}; \\ |
| 698 |
\node {a}; & \node(inner node) {b}; & \node {c}; & \node {d}; \\ |
| 699 |
\node {a}; & \node {b}; & \node {c}; & \node {d}; \\ |
| 700 |
}; |
| 701 |
\draw (inner node.south) circle (1pt); |
| 702 |
\end{tikzpicture} |
| 703 |
\end{codeexample} |
| 704 |
% |
| 705 |
\end{key} |
| 706 |
|
| 707 |
|
| 708 |
\subsection{Considerations Concerning Active Characters} |
| 709 |
|
| 710 |
Even though \tikzname\ seems to use |&| to separate cells, \pgfname\ actually |
| 711 |
uses a different command to separate cells, namely the command |
| 712 |
|\pgfmatrixnextcell| and using a normal |&| character will normally fail. What |
| 713 |
happens is that, \tikzname\ makes |&| an active character and then defines this |
| 714 |
character to be equal to |\pgfmatrixnextcell|. In most situations this will |
| 715 |
work nicely, but sometimes |&| cannot be made active; for instance because the |
| 716 |
matrix is used in an argument of some macro or the matrix contains nodes that |
| 717 |
contain normal |{tabular}| environments. In this case you can use the following |
| 718 |
option to avoid having to type |\pgfmatrixnextcell| each time: |
| 719 |
|
| 720 |
\begin{key}{/tikz/ampersand replacement=\meta{macro name or empty}} |
| 721 |
If a macro name is provided, this macro will be defined to be equal to |
| 722 |
|\pgfmatrixnextcell| inside matrices and |&| will not be made active. For |
| 723 |
instance, you could say |ampersand replacement=\&| and then use |\&| to |
| 724 |
separate columns as in the following example: |
| 725 |
% |
| 726 |
\begin{codeexample}[] |
| 727 |
\tikz |
| 728 |
\matrix [ampersand replacement=\&] |
| 729 |
{ |
| 730 |
\draw (0,0) circle (4mm); \& \node[rotate=10] {Hello}; \\ |
| 731 |
\draw (0.2,0) circle (2mm); \& \fill[red] (0,0) circle (3mm); \\ |
| 732 |
}; |
| 733 |
\end{codeexample} |
| 734 |
% |
| 735 |
\end{key} |
| 736 |
|
| 737 |
|
| 738 |
\subsection{Examples} |
| 739 |
|
| 740 |
The following examples are adapted from code by Mark Wibrow. The first two |
| 741 |
redraw pictures from Timothy van Zandt's PStricks documentation: |
| 742 |
% |
| 743 |
{\catcode`\|=12 |
| 744 |
\begin{codeexample}[preamble={\usetikzlibrary{matrix}}] |
| 745 |
\begin{tikzpicture} |
| 746 |
\matrix [matrix of math nodes,row sep=1cm] |
| 747 |
{ |
| 748 |
|(U)| U &[2mm] &[8mm] \\ |
| 749 |
& |(XZY)| X \times_Z Y & |(X)| X \\ |
| 750 |
& |(Y)| Y & |(Z)| Z \\ |
| 751 |
}; |
| 752 |
\begin{scope}[every node/.style={midway,auto,font=\scriptsize}] |
| 753 |
\draw [double, dashed] (U) -- node {$x$} (X); |
| 754 |
\draw (X) -- node {$p$} (X -| XZY.east) |
| 755 |
(X) -- node {$f$} (Z) |
| 756 |
-- node {$g$} (Y) |
| 757 |
-- node {$q$} (XZY) |
| 758 |
-- node {$y$} (U); |
| 759 |
\end{scope} |
| 760 |
\end{tikzpicture} |
| 761 |
\end{codeexample} |
| 762 |
|
| 763 |
\begin{codeexample}[ |
| 764 |
preamble={\usetikzlibrary{matrix}}, |
| 765 |
pre={\definecolor{graphicbackground}{rgb}{0.96,0.96,0.8}}, |
| 766 |
] |
| 767 |
\begin{tikzpicture}[>=stealth,->,shorten >=2pt,looseness=.5,auto] |
| 768 |
\matrix [matrix of math nodes, |
| 769 |
column sep={2cm,between origins}, |
| 770 |
row sep={3cm,between origins}, |
| 771 |
nodes={circle, draw, minimum size=7.5mm}] |
| 772 |
{ |
| 773 |
& |(A)| A & \\ |
| 774 |
|(B)| B & |(E)| E & |(C)| C \\ |
| 775 |
& |(D)| D \\ |
| 776 |
}; |
| 777 |
\begin{scope}[every node/.style={font=\small\itshape}] |
| 778 |
\draw (A) to [bend left] node [midway] {g} (B); |
| 779 |
\draw (B) to [bend left] node [midway] {f} (A); |
| 780 |
\draw (D) -- node [midway] {c} (B); |
| 781 |
\draw (E) -- node [midway] {b} (B); |
| 782 |
\draw (E) -- node [near end] {a} (C); |
| 783 |
\draw [-,line width=8pt,draw=graphicbackground] |
| 784 |
(D) to [bend right, looseness=1] (A); |
| 785 |
\draw (D) to [bend right, looseness=1] |
| 786 |
node [near start] {b} node [near end] {e} (A); |
| 787 |
\end{scope} |
| 788 |
\end{tikzpicture} |
| 789 |
\end{codeexample} |
| 790 |
|
| 791 |
\begin{codeexample}[preamble={\usetikzlibrary{matrix}}] |
| 792 |
\begin{tikzpicture} |
| 793 |
\matrix (network) |
| 794 |
[matrix of nodes,% |
| 795 |
nodes in empty cells, |
| 796 |
nodes={outer sep=0pt,circle,minimum size=4pt,draw}, |
| 797 |
column sep={1cm,between origins}, |
| 798 |
row sep={1cm,between origins}] |
| 799 |
{ |
| 800 |
& & & \\ |
| 801 |
& & & \\ |
| 802 |
|[draw=none]| & |[xshift=1mm]| & |[xshift=-1mm]| \\ |
| 803 |
}; |
| 804 |
\foreach \a in {1,...,4}{ |
| 805 |
\draw (network-3-2) -- (network-2-\a); |
| 806 |
\draw (network-3-3) -- (network-2-\a); |
| 807 |
\draw [-stealth] ([yshift=5mm]network-1-\a.north) -- (network-1-\a); |
| 808 |
\foreach \b in {1,...,4} |
| 809 |
\draw (network-1-\a) -- (network-2-\b); |
| 810 |
} |
| 811 |
\draw [stealth-] ([yshift=-5mm]network-3-2.south) -- (network-3-2); |
| 812 |
\draw [stealth-] ([yshift=-5mm]network-3-3.south) -- (network-3-3); |
| 813 |
\end{tikzpicture} |
| 814 |
\end{codeexample} |
| 815 |
|
| 816 |
The following example is adapted from code written by Kjell Magne Fauske, which |
| 817 |
is based on the following paper: K.~Bossley, M.~Brown, and C.~Harris, |
| 818 |
Neurofuzzy identification of an autonomous underwater vehicle, |
| 819 |
\emph{International Journal of Systems Science}, 1999, 30, 901--913. |
| 820 |
% |
| 821 |
\begin{codeexample}[preamble={\usetikzlibrary{arrows,shapes.geometric}}] |
| 822 |
\begin{tikzpicture} |
| 823 |
[auto, |
| 824 |
decision/.style={diamond, draw=blue, thick, fill=blue!20, |
| 825 |
text width=4.5em,align=flush center, |
| 826 |
inner sep=1pt}, |
| 827 |
block/.style ={rectangle, draw=blue, thick, fill=blue!20, |
| 828 |
text width=5em,align=center, rounded corners, |
| 829 |
minimum height=4em}, |
| 830 |
line/.style ={draw, thick, -latex',shorten >=2pt}, |
| 831 |
cloud/.style ={draw=red, thick, ellipse,fill=red!20, |
| 832 |
minimum height=2em}] |
| 833 |
|
| 834 |
\matrix [column sep=5mm,row sep=7mm] |
| 835 |
{ |
| 836 |
% row 1 |
| 837 |
\node [cloud] (expert) {expert}; & |
| 838 |
\node [block] (init) {initialize model}; & |
| 839 |
\node [cloud] (system) {system}; \\ |
| 840 |
% row 2 |
| 841 |
& \node [block] (identify) {identify candidate model}; & \\ |
| 842 |
% row 3 |
| 843 |
\node [block] (update) {update model}; & |
| 844 |
\node [block] (evaluate) {evaluate candidate models}; & \\ |
| 845 |
% row 4 |
| 846 |
& \node [decision] (decide) {is best candidate}; & \\ |
| 847 |
% row 5 |
| 848 |
& \node [block] (stop) {stop}; & \\ |
| 849 |
}; |
| 850 |
\begin{scope}[every path/.style=line] |
| 851 |
\path (init) -- (identify); |
| 852 |
\path (identify) -- (evaluate); |
| 853 |
\path (evaluate) -- (decide); |
| 854 |
\path (update) |- (identify); |
| 855 |
\path (decide) -| node [near start] {yes} (update); |
| 856 |
\path (decide) -- node [midway] {no} (stop); |
| 857 |
\path [dashed] (expert) -- (init); |
| 858 |
\path [dashed] (system) -- (init); |
| 859 |
\path [dashed] (system) |- (evaluate); |
| 860 |
\end{scope} |
| 861 |
\end{tikzpicture} |
| 862 |
\end{codeexample} |
| 863 |
} |
| 864 |
|
| 865 |
|
| 866 |
%%% Local Variables: |
| 867 |
%%% mode: latex |
| 868 |
%%% TeX-master: "pgfmanual" |
| 869 |
%%% End: |