Mercurial > hg > GlobalNeighbors

annotate tests/test_distance.py @ 1:1b94f3bf97e5

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
* limit distance function * start gridding * improve unicode handling
author Jeff Hammel <k0scist@gmail.com>
date Sat, 24 Jun 2017 14:02:14 -0700
parents 5dba84370182
children 49aae0c0293b
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
1 #!/usr/bin/env python
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
2
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
3 """
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
4 test distance calculation
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
5 """
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
6
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
7 import math
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
8 import os
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
9 import unittest
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
10 from globalneighbors import distance
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
11 from globalneighbors.constants import Rearth
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
12 from globalneighbors.locations import locations
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
13 from globalneighbors.read import read_city_list
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
14 from globalneighbors.schema import primary_key
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
15
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
16 here = os.path.dirname(os.path.abspath(__file__))
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
17 data = os.path.join(here, 'data')
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
18 full_tsv_lines = 149092
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
19
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
20 class DistanceTests(unittest.TestCase):
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
21
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
22 # created with
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
23 # head -n 10 cities1000.txt > GlobalNeighbors/tests/data/sample.tsv
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
24 test_tsv = os.path.join(data, 'sample.tsv')
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
25 test_tsv_lines = 10
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
26
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
27 # full dataset: test with caution
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
28 full_tsv = os.path.join(data, 'cities1000.txt')
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
29 full_tsv_lines = 149092
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
30
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
31 def test_haversine(self):
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
32
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
33 # a simple canned case
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
34 # equator to pole
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
35 lat1 = 0.
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
36 lat2 = 90.
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
37 lon2 = 70. # undefined, technically
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
38 expected_distance = 0.5*math.pi
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
39
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
40 for lon1 in range(-135, 135, 15):
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
41 radians = [distance.deg_to_rad(degrees)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
42 for degrees in (lat1, lon2, lat2, lon2)]
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
43 error = (distance.haversine(*radians) == expected_distance)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
44 assert error < 1e-4
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
45
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
46 def test_distance(self):
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
47 """test distance between two known cities"""
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
48
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
49 # Source:https://en.wikipedia.org/wiki/List_of_cities_by_latitude
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
50 # http://www.distancefromto.net/distance-from-new-york-to-chicago-us
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
51 chicago = (40.71278,
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
52 -74.00594)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
53
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
54 new_york = (41.85003,
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
55 -87.65005)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
56 ref_distance = 1149.
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
57
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
58 args = [distance.deg_to_rad(i) for i in
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
59 list(chicago) + list(new_york)]
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
60 calculated = distance.haversine(*args, r=Rearth)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
61
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
62 # Allow some error for circular projection approximation
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
63 error = abs(calculated - ref_distance)/ref_distance
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
64
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
65 assert error < 0.01
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
66
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
67 def test_distances(self):
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
68 """"ensure disances monotonically decay"""
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
69
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
70 # parse the data
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
71 assert os.path.exists(self.test_tsv)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
72 cities = read_city_list(self.test_tsv)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
73 assert len(cities) == self.test_tsv_lines
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
74 city_locations = locations(cities)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
75 assert len(city_locations) == self.test_tsv_lines
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
76
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
77 # calculate all the neighbors
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
78 # WARNING: n*2 algorithm Too computationally intensive
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
79 # for full data set
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
80 for key, value in distance.calculate_distances(city_locations,
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
81 r=Rearth):
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
82 # for now, just make sure we can iterate over them
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
83 pass
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
84
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
85 def test_neighbors(self):
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
86
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
87 # parse the data
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
88 tsv = os.path.join(data, 'sample.tsv')
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
89 assert os.path.exists(tsv)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
90 cities = read_city_list(tsv)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
91 city_locations = locations(cities)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
92
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
93 # calculate the neighbors
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
94 neighbors = distance.calculate_neighbors(city_locations,
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
95 k=self.test_tsv_lines)
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
96 assert len(neighbors) == self.test_tsv_lines
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
97
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
98 # ensure distance increases for each thing
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
99 for src, value in neighbors.items():
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
100 distances = [i[-1] for i in value]
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
101 assert len(distances) == self.test_tsv_lines - 1
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
102 for i in range(1, len(distances)):
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
103 assert distances[i] >= distances[i-1]
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
104
1
1b94f3bf97e5 * limit distance function
Jeff Hammel <k0scist@gmail.com>
parents: 0
diff changeset
105
0
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
106 if __name__ == '__main__':
5dba84370182 initial commit; half-working prototype
Jeff Hammel <k0scist@gmail.com>
parents:
diff changeset
107 unittest.main()