-
Notifications
You must be signed in to change notification settings - Fork 0
/
ordering.py
188 lines (160 loc) · 5.68 KB
/
ordering.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
import os
from textcompare import *
# run in command line to allow easy opening of files in sequence
def read(directory_path, name_list):
for name in name_list:
print 'Press enter to open ' + name + '...'
raw_input()
file_path = os.path.join(directory_path, name + '.txt')
os.system('open "' + file_path + '"')
# order1: sort by sum of all outgoing edges
def order1(graph):
ids = graph.get_ids()
nodes = graph.get_nodes()
edges = graph.get_edges()
ids.sort(key=lambda id: -sum([edges[key] for key in edges.keys() if key[0] == id]))
return ids
# order2: sort by sum of all incoming edges
def order2(graph):
ids = graph.get_ids()
nodes = graph.get_nodes()
edges = graph.get_edges()
ids.sort(key=lambda id: -sum([edges[key] for key in edges.keys() if key[1] == id]))
return ids
# order3: sort by sum of all outgoing edges recursively
def order3(graph):
ids = graph.get_ids()
n = len(ids)
nodes = graph.get_nodes()
edges = graph.get_edges()
order = []
for i in range(n):
scores = [(id, -sum([edges[key] for key in edges.keys() if (key[0] == id and key[1] in ids)])) for id in ids]
scores.sort(key=lambda x: x[1])
cur_id = scores[0][0]
order.append(cur_id)
ids = [id for id in ids if id != cur_id]
return order
# order4: sort by sum of all incoming edges recursively
def order4(graph):
ids = graph.get_ids()
n = len(ids)
nodes = graph.get_nodes()
edges = graph.get_edges()
order = []
for i in range(n):
scores = [(id, -sum([edges[key] for key in edges.keys() if (key[1] == id and key[1] in ids)])) for id in ids]
scores.sort(key=lambda x: x[1])
cur_id = scores[0][0]
order.append(cur_id)
ids = [id for id in ids if id != cur_id]
return order
# path1: greedily choose next node as smallest distance
def path1(graph, start_id, end_id):
ids = set(graph.get_ids())
nodes = graph.get_nodes()
edges = graph.get_edges()
path_ids = [start_id]
cur = start_id
while cur != end_id:
ids.remove(cur)
min_distance = float('inf')
min_id = None
for id in ids:
if (cur, id) in edges and edges[(cur, id)] != -1 and edges[(cur, id)] < min_distance:
min_id = id
min_distance = edges[(cur, id)]
if not min_id: return []
path_ids.append(min_id)
cur = min_id
return path_ids
# path2: shortest overall distance path with Dijkstra
def path2(graph, start_id, end_id):
ids = set(graph.get_ids())
nodes = graph.get_nodes()
edges = graph.get_edges()
distance = {}
previous = {}
max_distance = max(graph.get_edges().values())
for id in ids:
distance[id] = max_distance
previous[id] = None
distance[start_id] = 0
while ids:
cur = sorted([(key, distance[key]) for key in distance.keys() if key in ids], key=lambda x: x[1])[0][0]
ids.remove(cur)
for id in ids:
if (cur, id) in edges and edges[(cur, id)] >= 0:
alt = distance[cur] + edges[(cur, id)]
if alt < distance[id]:
distance[id] = alt
previous[id] = cur
path_ids = [end_id]
cur = end_id
while True:
cur = previous[cur]
path_ids.append(cur)
if cur == start_id:
break
path_ids = list(reversed(path_ids))
return path_ids
# path3: shortest overall distance path with Dijkstra using max distance cutoff
def path3(graph, start_id, end_id):
PERCENTILE = 0.8
ids = set(graph.get_ids())
nodes = graph.get_nodes()
edges = graph.get_edges()
distances = sorted([value for value in edges.values() if value >= 0])
distance_cutoff = distances[int(len(distances) * PERCENTILE)]
distance = {}
previous = {}
max_distance = max(graph.get_edges().values())
for id in ids:
distance[id] = max_distance
previous[id] = None
distance[start_id] = 0
while ids:
cur = sorted([(key, distance[key]) for key in distance.keys() if key in ids], key=lambda x: x[1])[0][0]
ids.remove(cur)
for id in ids:
if (cur, id) in edges and edges[(cur, id)] >= 0 and edges[(cur, id)] <= distance_cutoff:
alt = distance[cur] + edges[(cur, id)]
if alt < distance[id]:
distance[id] = alt
previous[id] = cur
path_ids = [end_id]
cur = end_id
while True:
if cur not in previous: return []
cur = previous[cur]
if cur == None:
return []
path_ids.append(cur)
if cur == start_id:
break
path_ids = list(reversed(path_ids))
path_names = [nodes[id] for id in path_ids]
return path_ids
if __name__ == '__main__':
### Set up graph
directory_path = "/Users/{0}/Dropbox (MIT)/children's books/books/".format(os.environ['USER'])
graph = make_graph(directory_path = directory_path, metric_name = 'Tversky index')
graph.prune_edges()
ids = set(graph.get_ids())
nodes = graph.get_nodes()
edges = graph.get_edges()
## Get list of names and ids
ids = order1(graph)
names1 = list(reversed([nodes[id] for id in ids]))
print names1
ids = order2(graph)
names2 = list(reversed([nodes[id] for id in ids]))
print names2
ids = order3(graph)
names3 = list(reversed([nodes[id] for id in ids]))
print names3
ids = order4(graph)
names4 = list(reversed([nodes[id] for id in ids]))
print names4
print sum([1 if names2[i] == names4[i] else 0 for i in range(len(names1))])
print sum([1 if names1[i] == names3[i] else 0 for i in range(len(names1))])