-
Notifications
You must be signed in to change notification settings - Fork 1
/
AVLTree.py
666 lines (560 loc) · 19.5 KB
/
AVLTree.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
#don't touch
LEFT = 0
RIGHT = 1
# Balance Binary Tree Class
class AVLTree:
def __init__(self):
self.parent = None
self.child = [None, None]
self.height = 0
def __repr__(self):
return "|{}, h:{}, b:{}, parent:{}, children [{},{}]|".format(self.name, self.height, self.compute_balance_factor(),
None if not self.parent else self.parent.name,
None if not self.child[LEFT] else self.child[LEFT].name,
None if not self.child[RIGHT] else self.child[RIGHT].name)
# starting at this node, walk down as far left
def first(self):
#current node in traversal
cur = self
while cur.child[LEFT]:
cur = cur.child[LEFT]
return cur
# starting at this node, walk down as far right
def last(self):
#current node in traversal
cur = self
while cur.child[RIGHT]:
cur = cur.child[RIGHT]
return cur
# follow parent up the tree as long as possible
def find_root(self):
root = self
while root.parent:
root = root.parent
return root
#isolates this node, takes care of child and parent pointers
def isolate(self):
# false if no parent
# isolate from parent
# decrease heights of all parents
aux = self.parent
while aux:
aux.height -= 1
aux = aux.parent
if self.parent:
if self.parent.child[LEFT] is self:
self.parent.child[LEFT] = None
else:
self.parent.child[RIGHT] = None
# Isolate from children
if self.child[LEFT]:
self.child[LEFT].parent = None
if self.child[RIGHT]:
self.child[RIGHT].parent = None
# remove children
self.child
# find successor in InOrder Traversal of InOrder
# returns this node if exists, None otherwise
def successor(self):
sub_successor = None
# if there is a right sub tree
if self.child[RIGHT]:
# go into right sub tree
cur = self.child[RIGHT]
sub_successor = cur
# then go left as far as possible as this will be successor
while cur:
sub_successor = cur
cur = cur.child[LEFT]
# if no sub_successor, need to go up through parent and check if successor exists there
# otherwise return this sub_successor
if sub_successor:
return sub_successor
else:
# check if parent exists
if self.parent:
# if it is the left child
if self is self.parent.child[LEFT]:
#successor in Inorder traversal is then the parent
return self.parent
# is right child
else:
cur = self.parent
while cur and cur.parent:
# keep going up to parent until we find
# that our parent is a left node, return the parent of this parent
if cur is cur.parent.child[LEFT]:
return cur.parent
cur = cur.parent
# if we get here we return None
return None
# find predecessor in InOrder Traversal of InOrder
# returns this node if exists, None otherwise
def predecessor(self):
sub_predecessor = None
# if there is a left sub tree
if self.child[LEFT]:
# go into left sub tree
cur = self.child[LEFT]
sub_predecessor = cur
# then go right as far as possible as this will be successor
while cur:
sub_predecessor = cur
cur = cur.child[RIGHT]
# if no sub_predecessor, need to go up through parent and check if predecessor exists there
# otherwise return this sub_predecessor
if sub_predecessor:
return sub_predecessor
else:
# check if parent exists
if self.parent:
# if it is the right child
if self is self.parent.child[RIGHT]:
#predecessor in Inorder traversal is then the parent
return self.parent
# is left child
else:
cur = self.parent
# check if there is parent
while cur and cur.parent:
# keep going up to parent until we find
# that our parent is a right node, return the parent of this parent
if cur is cur.parent.child[RIGHT]:
return cur.parent
cur = cur.parent
# if we get here we return None
return None
def cyclic_pred(self):
c_pred = self.last() if self is self.first() else self.predecessor()
return c_pred
def cyclic_succ(self):
c_succ = self.first() if self is self.last() else self.successor()
return c_succ
#this will be called in the weighted class
def after_rot(self):
pass
# this will be called in the weighted class
def init(self):
pass
# finds child with larger height
def tallerChild(self):
left_h = -1
right_h = -1
if self.child[LEFT]:
left_h = self.child[LEFT].height
if self.child[RIGHT]:
right_h = self.child[RIGHT].height
if right_h > left_h:
return self.child[RIGHT]
else:
return self.child[LEFT]
# this will be called in the weighted class
def update_subtree_weight(self):
pass
# returns the in-order list of nodes
def in_order(self):
accum = [self]
if self.child[LEFT]:
accum = self.child[LEFT].in_order() + accum
if self.child[RIGHT]:
accum = accum + self.child[RIGHT].in_order()
return accum
def compute_balance_factor(self):
right_h = -1 if not self.child[RIGHT] else self.child[RIGHT].height
left_h = -1 if not self.child[LEFT] else self.child[LEFT].height
return right_h - left_h
def update_height(self):
right_h = -1 if not self.child[RIGHT] else self.child[RIGHT].height
left_h = -1 if not self.child[LEFT] else self.child[LEFT].height
self.height = max(right_h, left_h) + 1
aux = self.parent
while aux:
right_h = -1 if not aux.child[RIGHT] else aux.child[RIGHT].height
left_h = -1 if not aux.child[LEFT] else aux.child[LEFT].height
aux.height = max(right_h, left_h) + 1
aux = aux.parent
# split is same as in RBBTREE, but with additional rebalancing of new tree
def split(start_node, direction, dummy):
if not start_node:
t1 = None
t2 = None
return t1, t2
dummy.child[LEFT] = None
dummy.child[RIGHT] = None
# we want to add dummy node in manner where we don't cut off and part
# of the tree, and maintains InOrder of our tree, rotating dummy up until it
# replaces our root, where we then can isolate dummy creating two split
# trees
# split after our start node, t1 contains start node
if(direction == RIGHT):
sub_successor = None
if start_node.child[RIGHT]:
cur = start_node.child[RIGHT]
sub_successor = cur
while cur:
sub_successor = cur
cur = cur.child[LEFT]
if not sub_successor:
#None to right of start node, so replace with dummy
start_node.child[RIGHT] = dummy
dummy.parent = start_node
else:
# store dummy as left child of subtree successor which is immediately after our start node
# as this is always None, sub_successor does not have a right child by def
sub_successor.child[LEFT] = dummy
dummy.parent = sub_successor
# split before our start node, t1 does not contain start node
else:
sub_predecessor = None
if start_node.child[LEFT]:
cur = start_node.child[LEFT]
sub_predecessor = cur
while cur:
sub_predecessor = cur
cur = cur.child[RIGHT]
if not sub_predecessor:
# None at left child so replace that with dummy
start_node.child[LEFT] = dummy
dummy.parent = start_node
else:
# store dummy as right child of subtree predecessor which is immediately before our start node
# as this is always None, sub_predecessor does not have a right child by def
sub_predecessor.child[RIGHT] = dummy
dummy.parent = sub_predecessor
# for derived classes
dummy.init()
#rotate dummy until it becomes root
while dummy.parent:
p = dummy.parent
rotation_direction = RIGHT if p.child[LEFT] is dummy else LEFT
#print(p)
rotate(p, rotation_direction)
t1 = dummy.child[LEFT]
t2 = dummy.child[RIGHT]
dummy.isolate()
if t1:
rebalance(t1)
t1 = t1.find_root()
if t2:
rebalance(t2)
t2 = t2.find_root()
return t1, t2
def smaller(u, v):
# if u or v = None
if not u or not v:
return False
# if they are the same also return false
if u is v:
return False
# get height of u
height_u = 0
cur_u = u
while cur_u.parent:
height_u += 1
cur_u = cur_u.parent
# get height of v
height_v = 0
cur_v = v
while cur_v.parent:
height_v += 1
cur_v = cur_v.parent
#case where they have different root, then we can't determine smaller
if cur_u is not cur_v:
return False
# paths from root
u_path = []
v_path = []
# construct paths in term of Lefts and Rights from root
cur_u = u
while cur_u.parent:
if cur_u.parent.child[LEFT] is cur_u:
#build path bottom up, first index will be path starting from root
u_path = [LEFT] + u_path
else:
u_path = [RIGHT] + u_path
cur_u = cur_u.parent
cur_v = v
while cur_v.parent:
if cur_v.parent.child[LEFT] is cur_v:
#build path bottom up, first index will be path starting from root
v_path = [LEFT] + v_path
else:
v_path = [RIGHT] + v_path
cur_v = cur_v.parent
#we compare paths such that the one that is most left is the one that will be before in Inorder
i = 0
# we find index of first difference in paths
while i < height_u and i < height_v:
if u_path[i] != v_path[i]:
break
i += 1
# if we have not reached end of path u, and step at i is LEFT for u, then u is more left than v
if i < height_u and u_path[i] == LEFT:
return True
# if we have not reach end of path v, and step at i is RIGHT for v, v is more right than u
elif i < height_v and v_path[i] == RIGHT:
return True
else:
return False
def join(t1, t2, dummy):
if not t1 or not t2:
if t1:
return t1
elif t2:
return t2
else:
return None
dummy.init()
# figure out heights of two trees
if t1.height > t2.height + 1:
joinRight(t1, t2, dummy)
elif t2.height > t1.height + 1:
joinLeft(t1, t2, dummy)
else:
dummy.child[LEFT] = t1
dummy.child[RIGHT] = t2
t1.parent = dummy
t2.parent = dummy
dummy.update_height()
# we want to remove dummy node, keeping balance intact
delete_dummy(dummy)
return t1.find_root()
# delete node from the tree it belongs to, handling appropriate rebalancing
# after the deletion
def delete_dummy(dummy):
if not dummy.child[LEFT] and not dummy.child[RIGHT]:
parent = dummy.parent
dummy.isolate()
parent.update_subtree_weight()
parent.update_height()
rebalance(parent)
elif not dummy.child[LEFT]:
parent = dummy.parent
if parent.child[LEFT] is dummy:
parent.child[LEFT] = dummy.child[RIGHT]
if parent.child[LEFT]:
parent.child[LEFT].parent = parent
else:
parent.child[RIGHT] = dummy.child[RIGHT]
if parent.child[RIGHT]:
parent.child[RIGHT].parent = parent
dummy.child[RIGHT] = None
dummy.child[LEFT] = None
dummy.parent = None
parent.update_subtree_weight()
parent.update_height()
rebalance(parent)
elif not dummy.child[RIGHT]:
parent = dummy.parent
if parent.child[LEFT] is dummy:
parent.child[LEFT] = dummy.child[LEFT]
if parent.child[LEFT]:
parent.child[LEFT].parent = parent
else:
parent.child[RIGHT] = dummy.child[LEFT]
if parent.child[RIGHT]:
parent.child[RIGHT].parent = parent
dummy.child[RIGHT] = None
dummy.child[LEFT] = None
dummy.parent = None
parent.update_subtree_weight()
parent.update_height()
rebalance(parent)
else: # case where dummy has two children
succ = dummy.successor()
parent = succ.parent
succ.child[LEFT] = dummy.child[LEFT]
if succ.child[LEFT]:
succ.child[LEFT].parent = succ
if dummy.parent:
if dummy.parent.child[RIGHT] is dummy:
dummy.parent.child[RIGHT] = succ
else:
dummy.parent.child[LEFT] = succ
succ.parent = dummy.parent
if parent is not dummy:
parent.child[LEFT] = succ.child[RIGHT]
if parent.child[LEFT]:
parent.child[LEFT].parent = parent
succ.child[RIGHT] = dummy.child[RIGHT]
if succ.child[RIGHT]:
succ.child[RIGHT].parent = succ
dummy.child[RIGHT] = None
dummy.child[LEFT] = None
dummy.parent = None
succ.update_height()
parent.update_height()
parent.update_subtree_weight()
succ.update_subtree_weight()
rebalance(parent)
def rebalance(parent):
# rebalance the tree moving up the tree until it is fully balanced
while (parent):
if abs(parent.compute_balance_factor()) > 1:
x = parent
y = x.tallerChild()
z = y.tallerChild()
parent = trinode_restructure(x,y,z)
parent.update_subtree_weight()
parent = parent.parent
def trinode_restructure(x,y,z):
# figure out the four possible cases for the rotation
zLeft = (z is y.child[LEFT])
yLeft = (y is x.child[LEFT])
# Left configuration
if zLeft and yLeft:
rotate(x, RIGHT)
return y
# Left Right configuration
elif not zLeft and yLeft:
rotate(y, LEFT)
rotate(x, RIGHT)
return y
# Right Left configuration
elif zLeft and not yLeft:
rotate(y, RIGHT)
rotate(x, LEFT)
return y
# Right configurationhah
else:
rotate(x, LEFT)
return y
# helper function for joining AVL tree where the right tree has smaller height
def joinRight(t1, t2, dummy):
t1_left = t1.child[LEFT]
t1_right = t1.child[RIGHT]
if height(t1_right) <= height(t2) + 1:
dummy.child[LEFT] = t1_right
dummy.child[RIGHT] = t2
if t1_right:
t1_right.parent = dummy
t2.parent = dummy
dummy.update_subtree_weight()
dummy.update_height()
if height(dummy) <= height(t1_left) + 1:
t1.child[LEFT] = t1_left
t1.child[RIGHT] = dummy
if t1_left:
t1_left.parent = t1
dummy.parent = t1
t1.update_height()
t1.update_subtree_weight()
return t1
else:
t_ = rotate(dummy, RIGHT)
t1.child[LEFT] = t1_left
t1.child[RIGHT] = t_
if t1_left:
t1_left.parent = t1
t_.parent = t1
t1.update_height()
t1.update_subtree_weight()
return rotate(t1, LEFT)
else:
t_ = joinRight(t1_right, t2, dummy)
t1.child[LEFT] = t1_left
t1.child[RIGHT] = t_
if t1_left:
t1_left.parent = t1
t_.parent = t1
t1.update_subtree_weight()
t1.update_height()
if height(t_) <= height(t1_left) + 1:
return t1
else:
return rotate(t1, LEFT)
# helper function for joining AVL tree where the left tree has smaller height
def joinLeft(t1, t2, dummy):
t2_left = t2.child[LEFT]
t2_right = t2.child[RIGHT]
if height(t2_left) <= height(t1) + 1:
dummy.child[LEFT] = t1
dummy.child[RIGHT] = t2_left
if t2_left:
t2_left.parent = dummy
t1.parent = dummy
dummy.update_subtree_weight()
dummy.update_height()
if height(dummy) <= height(t2_right) + 1:
t2.child[LEFT] = dummy
t2.child[RIGHT] = t2_right
if t2_right:
t2_right.parent = t2
dummy.parent = t2
t2.update_subtree_weight()
t2.update_height()
return t2
else:
t_ = rotate(dummy, LEFT)
t2.child[LEFT] = t_
t2.child[RIGHT] = t2_right
if t2_right:
t2_right.parent = t2
t_.parent = t2
t2.update_subtree_weight()
t2.update_height()
return rotate(t2, RIGHT)
else:
t_ = joinLeft(t1, t2_left, dummy)
t2.child[LEFT] = t_
t2.child[RIGHT] = t2_right
if t2_right:
t2_right.parent = t2
t_.parent = t2
t2.update_subtree_weight()
t2.update_height()
if height(t_) <= height(t2_right) + 1:
return t2
else:
return rotate(t2, RIGHT)
def rotate(r_parent, rotation_direction):
r_child = r_parent.child[1 - rotation_direction]
mid_tree = r_child.child[rotation_direction]
# move mid tree to opposite side of child of parent
r_parent.child[1 - rotation_direction] = mid_tree
if mid_tree:
mid_tree.parent = r_parent
# assign child the parent of its parent
r_child.parent = r_parent.parent
# update this new parent to replace its child which was
# r_parent to be r_child
if r_child.parent:
if (r_child.parent.child[LEFT] is r_parent):
r_child.parent.child[LEFT] = r_child
else:
r_child.parent.child[RIGHT] = r_child
# rotate parent to be child of child in direction
r_child.child[rotation_direction] = r_parent
r_parent.parent = r_child
# update r_parent first as it is below r_child now, propagate weights up
r_parent.update_height()
r_child.update_height()
# fix additional information in derived classes
r_parent.after_rot()
#this is new root
return r_child
# compute height in O(1) time
def height(root):
if not root:
return -1
r = -1
l = -1
if root.child[LEFT]:
l = root.child[LEFT].height
if root.child[RIGHT]:
r = root.child[RIGHT].height
return 1 + max(l, r)
# levelorder traversal from root
def print_tree(root):
h = height(root)
for i in range(1, h+ 1):
_print_tree(root, i)
print("")
def _print_tree(root, level):
if not root:
return root
if level == 1:
print(root, end = ' ')
elif level > 1:
_print_tree(root.child[LEFT], level - 1)
_print_tree(root.child[RIGHT], level - 1)