-
Notifications
You must be signed in to change notification settings - Fork 13
/
temporal_revision.py
171 lines (156 loc) · 6.39 KB
/
temporal_revision.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
# Copyright (c) 2019 kamyu. All rights reserved.
#
# Facebook Hacker Cup 2019 Final Round - Temporal Revision
# https://www.facebook.com/hackercup/problem/1165177510537433/
#
# Time: O((S + N) * logN + (M + K) * (log*)(N)), p.s. (log*)(N) <= 5
# Space: O(NlogN), due to skip-list of tree node ancestors
#
from collections import defaultdict
from functools import partial
# Template:
# https://github.com/kamyu104/LeetCode-Solutions/blob/master/Python/accounts-merge.py
class UnionFind(object):
def __init__(self, n):
self.set = range(n)
def get_id(self):
self.set.append(len(self.set))
return len(self.set)-1
def find_set(self, x):
stk = []
while self.set[x] != x: # path compression.
stk.append(x)
x = self.set[x]
while stk:
self.set[stk.pop()] = x
return x
def union_set(self, x, y):
x_root, y_root = map(self.find_set, (x, y))
if x_root == y_root:
return False
self.set[min(x_root, y_root)] = max(x_root, y_root)
return True
# Template:
# https://github.com/kamyu104/FacebookHackerCup-2019/blob/master/Final%20Round/little_boat_on_the_sea.py
def find_binary_tree_infos(N, children):
def preprocess(L, P, C, curr, parent):
# ancestors of the node i
P[curr].append(parent)
i = 0
while P[curr][i] != -1:
P[curr].append(P[P[curr][i]][i] if i < len(P[P[curr][i]]) else -1)
i += 1
# the subtree of the node i is represented by traversal index L[i]..R[i]
C[0] += 1
L[curr] = C[0]
def divide(stk, L, R, P, C, curr, parent):
stk.append(partial(postprocess, R, C, curr))
for i in reversed(xrange(len(children[curr]))):
child = children[curr][i]
stk.append(partial(divide, stk, L, R, P, C, child, curr))
stk.append(partial(preprocess, L, P, C, curr, parent))
def postprocess(R, C, curr):
R[curr] = C[0]
L, R, P, C = [0]*N, [0]*N, [[] for _ in xrange(N)], [-1]
stk = []
stk.append(partial(divide, stk, L, R, P, C, N-1, -1)) # root is the last node id: N-1
while stk:
stk.pop()()
return L, R, P
def build_binary_tree(N, A, B, E, V, activity_type):
union_find = UnionFind(N)
children = defaultdict(list)
when = defaultdict(lambda x: len(E))
curr_active_cnt_from_node = defaultdict(int)
for i in xrange(N):
curr_active_cnt_from_node[i] = int(activity_type[i] == OPEN)
dp1, dp2 = defaultdict(int), defaultdict(int)
for i in reversed(xrange(len(E))): # O((M+K)*alpha(N))
if E[i] == 1:
x, y = union_find.find_set(A[V[i]]), union_find.find_set(B[V[i]])
if x == y:
continue
v = union_find.get_id()
union_find.union_set(v, x), union_find.union_set(v, y)
children[v], when[v] = [x, y], i
curr_active_cnt_from_node[v] = curr_active_cnt_from_node[x] + curr_active_cnt_from_node[y]
dp1[v] = max(dp1[x], dp1[y])
dp2[v] = max(dp2[x], dp2[y], dp1[x]+dp1[y])
else:
v, x = union_find.get_id(), union_find.find_set(V[i])
union_find.union_set(v, x)
children[v], when[v] = [x], i
curr_active_cnt_from_node[v] = curr_active_cnt_from_node[x] + (-1 if E[i] == 2 else 1) # -1 means not yet activated, +1 means activated
dp1[v] = int(E[i] == 2) + dp1[x]
dp2[v] = int(E[i] == 2) + dp2[x]
return len(union_find.set), children, when, curr_active_cnt_from_node, dp1, dp2
def query_binary_tree(E, V, end_hr, children, when, curr_active_cnt_from_node, dp1, dp2, L, R, P, X, Y):
def is_ancestor(L, R, a, b): # includes itself
return L[a] <= L[b] <= R[b] <= R[a]
v = X
for i in reversed(xrange(len(P[v]))): # O(logN)
if i < len(P[v]) and P[v][i] != -1 and when[P[v][i]] >= Y:
v = P[v][i]
stk = [v]
while P[v][0] != -1: # O(1)
v = P[v][0]
if when[v] < Y-24:
break
stk.append(v)
stk = stk[::-1]
assert(len(stk) <= 24+1)
active_cnt_to_the_end_from_hr = [0]*len(stk)
accumulated_cnt = dp1[stk[-1]]
for i in reversed(xrange(len(stk)-1)):
active_cnt_to_the_end_from_hr[i] = accumulated_cnt
h = when[stk[i]]
if E[h] == 2 and end_hr[V[h]] >= Y and is_ancestor(L, R, stk[-1], V[h]):
accumulated_cnt += 1
max_cnt, curr_cnt = 0, curr_active_cnt_from_node[stk[0]]
for i in xrange(len(stk)-1):
h = when[stk[i]]
curr_cnt += int(E[h] == 2)
if len(children[stk[i]]) == 2:
other_sibling = children[stk[i]][0] if children[stk[i]][0] != stk[i+1] else children[stk[i]][1]
max_cnt = max(max_cnt, curr_cnt + dp1[other_sibling] + active_cnt_to_the_end_from_hr[i])
return max(max_cnt, curr_cnt + dp2[stk[-1]])
def temporal_revision():
N, M, K, S = map(int, raw_input().strip().split())
A, B = [0]*M, [0]*M
E, V = [0]*K, [0]*K
for i in xrange(M):
A[i], B[i] = map(int, raw_input().strip().split())
A[i], B[i] = A[i]-1, B[i]-1
activity_type, is_collapsed, end_hr = [0]*N, [False]*M, [0]*N
for i in xrange(K):
E[i], V[i] = map(int, raw_input().strip().split())
V[i] -= 1
if E[i] == 1:
is_collapsed[V[i]] = True
elif E[i] == 2:
activity_type[V[i]] |= OPEN
else: # E[i] == 3
activity_type[V[i]] |= CLOSE
end_hr[V[i]] = i
for i in xrange(M): # make remaining conduits as branch points
if is_collapsed[i] == False:
E.append(1), V.append(i)
for i in xrange(N):
if activity_type[i] == OPEN:
end_hr[i] = len(E) # set the max time to end activity
node_count, children, when, curr_active_cnt_from_node, dp1, dp2 = build_binary_tree(N, A, B, E, V, activity_type)
L, R, P = find_binary_tree_infos(node_count, children)
result, ans = 0, 0
for _ in xrange(S):
X, Y = map(int, raw_input().strip().split())
X, Y = X^ans, Y^ans
assert(1 <= X <= N and 1 <= Y <= K)
X, Y = X-1, Y-1
ans = query_binary_tree(E, V, end_hr,
children, when, curr_active_cnt_from_node, dp1, dp2,
L, R, P, X, Y)
result += ans
return result
OPEN, CLOSE = range(1, 3)
for case in xrange(input()):
print 'Case #%d: %s' % (case+1, temporal_revision())