1 | /* |
2 | * $Id: TraverseFunctions.java 594 2010-11-19 20:41:04Z PSpeed $ |
3 | * |
4 | * The Filament BSD license. |
5 | * |
6 | * Copyright (c) 2009-2010, the original author or authors |
7 | * |
8 | * Redistribution and use in source and binary forms, with or without |
9 | * modification, are permitted provided that the following conditions |
10 | * are met: |
11 | * |
12 | * 1) Redistributions of source code must retain the above copyright notice, |
13 | * this list of conditions and the following disclaimer. |
14 | * 2) Redistributions in binary form must reproduce the above copyright |
15 | * notice, this list of conditions and the following disclaimer in the |
16 | * documentation and/or other materials provided with the distribution. |
17 | * 3) Neither the names "Filament", "fgraph.org", "filamentgraph.org", nor the |
18 | * names of its contributors may be used to endorse or promote products |
19 | * derived from this software without specific prior written permission. |
20 | * |
21 | * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
22 | * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
23 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
24 | * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE |
25 | * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR |
26 | * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF |
27 | * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS |
28 | * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN |
29 | * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) |
30 | * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE |
31 | * POSSIBILITY OF SUCH DAMAGE. |
32 | */ |
33 | |
34 | package org.fgraph.steps; |
35 | |
36 | import java.util.*; |
37 | |
38 | import com.google.common.base.Function; |
39 | import com.google.common.base.Predicate; |
40 | import com.google.common.collect.Iterables; |
41 | import com.google.common.collect.Iterators; |
42 | |
43 | |
44 | /** |
45 | * |
46 | * @version $Revision: 594 $ |
47 | * @author Paul Speed |
48 | */ |
49 | public class TraverseFunctions |
50 | { |
51 | /** |
52 | * Creates a function that performs a pre-order depth-first |
53 | * traversal starting at an applied node and using the specified |
54 | * adjacency function for traversal. |
55 | * |
56 | * <p>If an optional visitMarker function is provided then the |
57 | * traversal is acyclic based on the 'marker' returned from that |
58 | * function. For example, a Functions.identity() visit marker |
59 | * will make sure a node is only visited once based on its own |
60 | * .equals() implementation. A null indicates that no cycle |
61 | * detection should be performed.</p> |
62 | */ |
63 | public static <T> StepFunction<T,T> preOrder( Function<T,? extends Iterable<T>> fstep, |
64 | Function<? super T,? extends Object> visitMarker ) |
65 | { |
66 | return StepFunctions.unwrapSteps( preOrderSteps(fstep, visitMarker) ); |
67 | } |
68 | |
69 | /** |
70 | * Creates a function that performs a pre-order depth-first |
71 | * traversal starting at an applied node and using the specified |
72 | * adjacency function for traversal, also providing additional |
73 | * control over and information about the ongoing traversal. |
74 | * |
75 | * <p>This is similar to the non-step based version except that |
76 | * the traversed values are returned wrapped in DepthFirstStep |
77 | * objects that provide additional information about the state |
78 | * of the traversal such as depth. These steps also allow themselves |
79 | * to be 'pruned', resulting in any potential descendants being ignored |
80 | * during the subsequent traverses.</p> |
81 | * |
82 | * <p>If an optional visitMarker function is provided then the |
83 | * traversal is acyclic based on the 'marker' returned from that |
84 | * function. For example, a Functions.identity() visit marker |
85 | * will make sure a node is only visited once based on its own |
86 | * .equals() implementation. A null indicates that no cycle |
87 | * detection should be performed.</p> |
88 | */ |
89 | public static <T> StepFunction<T,DepthFirstStep<T>> preOrderSteps( |
90 | Function<T,? extends Iterable<T>> fstep, |
91 | Function<? super T,? extends Object> visitMarker ) |
92 | { |
93 | StepFunction<T,DepthFirstStep<T>> fTraverse = depthFirstSteps(fstep, visitMarker); |
94 | fTraverse = StepFunctions.filter( fTraverse, StepPredicates.<T>descending() ); |
95 | return fTraverse; |
96 | } |
97 | |
98 | /** |
99 | * Creates a function that performs a post-order depth-first |
100 | * traversal starting at an applied node and using the specified |
101 | * adjacency function for traversal. |
102 | * |
103 | * <p>If an optional visitMarker function is provided then the |
104 | * traversal is acyclic based on the 'marker' returned from that |
105 | * function. For example, a Functions.identity() visit marker |
106 | * will make sure a node is only visited once based on its own |
107 | * .equals() implementation. A null indicates that no cycle |
108 | * detection should be performed.</p> |
109 | */ |
110 | public static <T> StepFunction<T,T> postOrder( Function<T,? extends Iterable<T>> fstep, |
111 | Function<? super T,? extends Object> visitMarker ) |
112 | { |
113 | return StepFunctions.unwrapSteps( postOrderSteps(fstep, visitMarker) ); |
114 | } |
115 | |
116 | /** |
117 | * Creates a function that performs a post-order depth-first |
118 | * traversal starting at an applied node and using the specified |
119 | * adjacency function for traversal, also providing additional |
120 | * control over and information about the ongoing traversal. |
121 | * |
122 | * <p>This is similar to the non-step based version except that |
123 | * the traversed values are returned wrapped in DepthFirstStep |
124 | * objects that provide additional information about the state |
125 | * of the traversal such as depth. These steps also allow themselves |
126 | * to be 'pruned', resulting in any potential descendants being ignored |
127 | * during the subsequent traverses.</p> |
128 | * |
129 | * <p>If an optional visitMarker function is provided then the |
130 | * traversal is acyclic based on the 'marker' returned from that |
131 | * function. For example, a Functions.identity() visit marker |
132 | * will make sure a node is only visited once based on its own |
133 | * .equals() implementation. A null indicates that no cycle |
134 | * detection should be performed.</p> |
135 | */ |
136 | public static <T> StepFunction<T,DepthFirstStep<T>> postOrderSteps( |
137 | Function<T,? extends Iterable<T>> fstep, |
138 | Function<? super T,? extends Object> visitMarker ) |
139 | { |
140 | StepFunction<T,DepthFirstStep<T>> fTraverse = depthFirstSteps(fstep, visitMarker); |
141 | fTraverse = StepFunctions.filter( fTraverse, StepPredicates.<T>ascending() ); |
142 | return fTraverse; |
143 | } |
144 | |
145 | /** |
146 | * Creates a function that performs a full depth-first |
147 | * traversal (both pre- and post-) starting at an applied node |
148 | * and using the specified adjacency function for traversal. |
149 | * Every 'node' will therefore be visited twice, once descending |
150 | * and once ascending. Use the step-based depthFirstSteps() |
151 | * method to distinguish descending from ascending. |
152 | * |
153 | * <p>If an optional visitMarker function is provided then the |
154 | * traversal is acyclic based on the 'marker' returned from that |
155 | * function. For example, a Functions.identity() visit marker |
156 | * will make sure a node is only visited once based on its own |
157 | * .equals() implementation. A null indicates that no cycle |
158 | * detection should be performed.</p> |
159 | */ |
160 | public static <T> StepFunction<T,T> depthFirst( Function<T,? extends Iterable<T>> fstep, |
161 | Function<? super T,? extends Object> visitMarker ) |
162 | { |
163 | return StepFunctions.unwrapSteps( depthFirstSteps(fstep, visitMarker) ); |
164 | } |
165 | |
166 | /** |
167 | * Creates a function that performs a full depth-first |
168 | * traversal (both pre- and post-) starting at an applied node |
169 | * and using the specified adjacency function for traversal, |
170 | * also providing additional control over and information about |
171 | * the ongoing traversal. |
172 | * |
173 | * <p>This is similar to the non-step based version except that |
174 | * the traversed values are returned wrapped in DepthFirstStep |
175 | * objects that provide additional information about the state |
176 | * of the traversal such as depth, descending/ascending, etc. |
177 | * These steps also allow themselves to be 'pruned', resulting |
178 | * in any potential descendants being ignored during the subsequent |
179 | * traverses.</p> |
180 | * |
181 | * <p>If an optional visitMarker function is provided then the |
182 | * traversal is acyclic based on the 'marker' returned from that |
183 | * function. For example, a Functions.identity() visit marker |
184 | * will make sure a node is only visited once based on its own |
185 | * .equals() implementation. A null indicates that no cycle |
186 | * detection should be performed.</p> |
187 | */ |
188 | public static <T> StepFunction<T,DepthFirstStep<T>> depthFirstSteps( |
189 | Function<T,? extends Iterable<T>> fstep, |
190 | Function<? super T,? extends Object> visitMarker ) |
191 | { |
192 | return new DepthFirstFunction<T>( fstep, visitMarker ); |
193 | } |
194 | |
195 | /** |
196 | * Creates a function that performs a breadth-first |
197 | * traversal starting at an applied node and using the specified |
198 | * adjacency function for traversal. |
199 | * |
200 | * <p>If an optional visitMarker function is provided then the |
201 | * traversal is acyclic based on the 'marker' returned from that |
202 | * function. For example, a Functions.identity() visit marker |
203 | * will make sure a node is only visited once based on its own |
204 | * .equals() implementation. A null indicates that no cycle |
205 | * detection should be performed.</p> |
206 | */ |
207 | public static <T> StepFunction<T,T> breadthFirst( Function<T,? extends Iterable<T>> fstep, |
208 | Function<? super T,? extends Object> visitMarker ) |
209 | { |
210 | return StepFunctions.unwrapSteps( breadthFirstSteps(fstep, visitMarker) ); |
211 | } |
212 | |
213 | /** |
214 | * Creates a function that performs a breadth-first |
215 | * traversal starting at an applied node and using the specified |
216 | * adjacency function for traversal, also providing additional |
217 | * control over and information about the ongoing traversal. |
218 | * |
219 | * <p>This is similar to the non-step based version except that |
220 | * the traversed values are returned wrapped in PrunableStep |
221 | * objects that provide additional information about the state |
222 | * of the traversal such as depth. These steps also allow themselves |
223 | * to be 'pruned', resulting in any potential descendants being ignored |
224 | * during the subsequent traverses.</p> |
225 | * |
226 | * <p>If an optional visitMarker function is provided then the |
227 | * traversal is acyclic based on the 'marker' returned from that |
228 | * function. For example, a Functions.identity() visit marker |
229 | * will make sure a node is only visited once based on its own |
230 | * .equals() implementation. A null indicates that no cycle |
231 | * detection should be performed.</p> |
232 | */ |
233 | public static <T> StepFunction<T,PrunableStep<T>> breadthFirstSteps( |
234 | Function<T,? extends Iterable<T>> fstep, |
235 | Function<? super T,? extends Object> visitMarker ) |
236 | { |
237 | return new BreadthFirstFunction<T>( fstep, visitMarker ); |
238 | } |
239 | |
240 | /** |
241 | * Creates a function that performs a priority-ordered |
242 | * traversal starting at an applied node and using the specified |
243 | * adjacency function for traversal. |
244 | * |
245 | * <p>The priority comparator is used to select the next step/path |
246 | * from the current pending set of descendants. For example, a |
247 | * comparator could be based to select the smallest/shortest route |
248 | * based on the sum of the accumulated weights/distances up to that step.</p> |
249 | * |
250 | * <p>If an optional visitMarker function is provided then the |
251 | * traversal is acyclic based on the 'marker' returned from that |
252 | * function. For example, a Functions.identity() visit marker |
253 | * will make sure a node is only visited once based on its own |
254 | * .equals() implementation. A null indicates that no cycle |
255 | * detection should be performed.</p> |
256 | */ |
257 | public static <T> StepFunction<T,T> priority( Function<T,? extends Iterable<T>> fstep, |
258 | Comparator<? super T> priority, |
259 | Function<? super T,? extends Object> visitMarker ) |
260 | { |
261 | return StepFunctions.unwrapSteps( prioritySteps( fstep, priority, visitMarker ) ); |
262 | } |
263 | |
264 | /** |
265 | * Creates a function that performs a priority-ordered |
266 | * traversal starting at an applied node and using the specified |
267 | * adjacency function for traversal, also providing additional |
268 | * control over and information about the ongoing traversal. |
269 | * |
270 | * <p>The priority comparator is used to select the next step/path |
271 | * from the current pending set of descendants. For example, a |
272 | * comparator could be based to select the smallest/shortest route |
273 | * based on the sum of the accumulated weights/distances up to that step.</p> |
274 | * |
275 | * <p>This is similar to the non-step based version except that |
276 | * the traversed values are returned wrapped in PrunableStep |
277 | * objects that provide additional information about the state |
278 | * of the traversal such as depth. These steps also allow themselves |
279 | * to be 'pruned', resulting in any potential descendants being ignored |
280 | * during the subsequent traverses. Note: for traversals that branch |
281 | * quickly, pruning a step may be a non-trivial amount of work since |
282 | * it must exhaustively search the current sorted pending steps to |
283 | * remove any queued descendants.</p> |
284 | * |
285 | * <p>If an optional visitMarker function is provided then the |
286 | * traversal is acyclic based on the 'marker' returned from that |
287 | * function. For example, a Functions.identity() visit marker |
288 | * will make sure a node is only visited once based on its own |
289 | * .equals() implementation. A null indicates that no cycle |
290 | * detection should be performed.</p> |
291 | */ |
292 | public static <T> StepFunction<T,PrunableStep<T>> prioritySteps( |
293 | Function<T,? extends Iterable<T>> fstep, |
294 | Comparator<? super T> priority, |
295 | Function<? super T,? extends Object> visitMarker ) |
296 | { |
297 | return new PriorityTraverseFunction<T>( fstep, priority, visitMarker ); |
298 | } |
299 | |
300 | /** |
301 | * Function implementation that creates depth first iterators on demand. |
302 | */ |
303 | private static class DepthFirstFunction<T> implements StepFunction<T,DepthFirstStep<T>> |
304 | { |
305 | private Function<T,? extends Iterable<T>> fstep; |
306 | private Function<? super T,? extends Object> visitMarker; |
307 | |
308 | public DepthFirstFunction( Function<T,? extends Iterable<T>> fstep, |
309 | Function<? super T,? extends Object> visitMarker ) |
310 | { |
311 | this.fstep = fstep; |
312 | this.visitMarker = visitMarker; |
313 | } |
314 | |
315 | public Iterable<DepthFirstStep<T>> apply( T start ) |
316 | { |
317 | return Traversers.depthFirstSteps( start, fstep, visitMarker ); |
318 | } |
319 | |
320 | public String toString() |
321 | { |
322 | return "DepthFirstFunction[delegate=" + fstep + ", visitMarker=" + visitMarker + "]"; |
323 | } |
324 | } |
325 | |
326 | /** |
327 | * Function implementation that creates breadth first iterators on demand. |
328 | */ |
329 | private static class BreadthFirstFunction<T> implements StepFunction<T,PrunableStep<T>> |
330 | { |
331 | private Function<T,? extends Iterable<T>> fstep; |
332 | private Function<? super T,? extends Object> visitMarker; |
333 | |
334 | public BreadthFirstFunction( Function<T,? extends Iterable<T>> fstep, |
335 | Function<? super T,? extends Object> visitMarker ) |
336 | { |
337 | this.fstep = fstep; |
338 | this.visitMarker = visitMarker; |
339 | } |
340 | |
341 | public Iterable<PrunableStep<T>> apply( T start ) |
342 | { |
343 | return Traversers.breadthFirstSteps( start, fstep, visitMarker ); |
344 | } |
345 | |
346 | public String toString() |
347 | { |
348 | return "BreadthFirstFunction[delegate=" + fstep + ", visitMarker=" + visitMarker + "]"; |
349 | } |
350 | } |
351 | |
352 | /** |
353 | * Function implementation that creates priority first iterators on demand. |
354 | */ |
355 | private static class PriorityTraverseFunction<T> implements StepFunction<T,PrunableStep<T>> |
356 | { |
357 | private Function<T,? extends Iterable<T>> fstep; |
358 | private Comparator<? super T> priority; |
359 | private Function<? super T,? extends Object> visitMarker; |
360 | |
361 | public PriorityTraverseFunction( Function<T,? extends Iterable<T>> fstep, |
362 | Comparator<? super T> priority, |
363 | Function<? super T,? extends Object> visitMarker ) |
364 | { |
365 | this.fstep = fstep; |
366 | this.priority = priority; |
367 | this.visitMarker = visitMarker; |
368 | } |
369 | |
370 | public Iterable<PrunableStep<T>> apply( T start ) |
371 | { |
372 | return Traversers.prioritySteps( start, fstep, priority, visitMarker ); |
373 | } |
374 | |
375 | public String toString() |
376 | { |
377 | return "PriorityTraverseFunction[delegate=" + fstep + ", priority=" + priority |
378 | + ", visitMarker=" + visitMarker + "]"; |
379 | } |
380 | } |
381 | |
382 | } |
383 | |