class MatcherTestCase(unittest.TestCase):
"""Tests the various Matcher subclasses"""
all_matchers = [GeogMatcher]
def setUp(self):
self.sim = Simulator()
self.sim.simulate(10, 13579)
def test_matcher(self):
"""Test various matcher subclasses"""
def test_empty_matcher(MatcherSub):
"""Test empty Matcher returns valid empty results"""
## Test that an empty set returns valid, emtpy results
matcher_sub = MatcherSub(set([]))
unmatched = matcher_sub.get_unmatched()
best_matches = matcher_sub.get_best_matches()
self.assertEqual(unmatched, set([]));
self.assertEqual(best_matches, set([]));
def test_matcher_matches(MatcherSub, expected_matches, expected_unmatches):
"""Spin up Matcher subclass and confirm expected results"""
users = self.sim.get_users()
self.assertEqual(len(users),10, "Sanity check simulator user count")
matcher_sub = MatcherSub(users,seed=1234)
best_matches = set([(tup[0].user_id, tup[1].user_id) for tup in matcher_sub.get_best_matches() ])
unmatched = set([user.user_id for user in matcher_sub.get_unmatched()])
print best_matches
print unmatched
for match in expected_matches:
self.assertIn(match, best_matches, "Expected match not found - perhaps the algo or simulator changed?")
for unmatch in expected_unmatches:
self.assertIn(unmatch, unmatched, "Expected unmatched not found - perhaps the algo or simulator changed?")
## Test empty matcher return valid results
for MatcherSub in self.all_matchers:
test_empty_matcher(MatcherSub)
## For each Matcher subclass, work out what the answers should be
## given the simulator's users, and record the answers here
geog_matched = [(7736498, 2837615), (5072324, 795755), (4454245, 3882723)]
geog_unmatched = [1510961, 8148259, 6505708, 2432669]
## Feed the expected results into 3-tuples below to be tested
## (MatcherSubclass , expected_match_results, expected_unmatched_results)
test_cases = [
(GeogMatcher, geog_matched,geog_unmatched)
]
## Test full matchers return correct results
for match_case in test_cases:
test_matcher_matches(*match_case)
def get_users(self):
"""Read in the survey data or simulate new data"""
self.users = self.get_users_parser(APIParser, self.survey_api_survey_id)
print self.users
if self.users is None:
if os.path.isfile(self.survey_data_filename):
self.users = self.get_users_parser(TextParser, self.survey_data_filename)
else:
print "Simulating data"
## Simulate some number of survey responses
sim = Simulator()
sim.simulate(200) ## By adding a seed parameter, you can have reproducible sims
self.users = sim.get_users()
# ## Print some summaries
# for user in users:
# user.to_string()
# Print some descriptive stats
print "# Users", len(self.users), [u.name for u in self.users]
class MatcherTestCase(unittest.TestCase):
"""Tests the various Matcher subclasses"""
all_matchers = [GeogMatcher]
def setUp(self):
self.sim = Simulator()
self.sim.simulate(10, 13579)
def test_matcher(self):
"""Test various matcher subclasses"""
def test_empty_matcher(MatcherSub):
"""Test empty Matcher returns valid empty results"""
## Test that an empty set returns valid, emtpy results
matcher_sub = MatcherSub(set([]))
unmatched = matcher_sub.get_unmatched()
best_matches = matcher_sub.get_best_matches()
self.assertEqual(unmatched, set([]))
self.assertEqual(best_matches, set([]))
def test_matcher_matches(MatcherSub, expected_matches,
expected_unmatches):
"""Spin up Matcher subclass and confirm expected results"""
users = self.sim.get_users()
self.assertEqual(len(users), 10,
"Sanity check simulator user count")
matcher_sub = MatcherSub(users, seed=1234)
best_matches = set([(tup[0].user_id, tup[1].user_id)
for tup in matcher_sub.get_best_matches()])
unmatched = set(
[user.user_id for user in matcher_sub.get_unmatched()])
print best_matches
print unmatched
for match in expected_matches:
self.assertIn(
match, best_matches,
"Expected match not found - perhaps the algo or simulator changed?"
)
for unmatch in expected_unmatches:
self.assertIn(
unmatch, unmatched,
"Expected unmatched not found - perhaps the algo or simulator changed?"
)
## Test empty matcher return valid results
for MatcherSub in self.all_matchers:
test_empty_matcher(MatcherSub)
## For each Matcher subclass, work out what the answers should be
## given the simulator's users, and record the answers here
geog_matched = [(7736498, 2837615), (5072324, 795755),
(4454245, 3882723)]
geog_unmatched = [1510961, 8148259, 6505708, 2432669]
## Feed the expected results into 3-tuples below to be tested
## (MatcherSubclass , expected_match_results, expected_unmatched_results)
test_cases = [(GeogMatcher, geog_matched, geog_unmatched)]
## Test full matchers return correct results
for match_case in test_cases:
test_matcher_matches(*match_case)
class SimulatorTestCase(unittest.TestCase):
def setUp(self):
self.sim = Simulator()
def tearDown(self):
self.sim = None
def test_simulator(self):
self.assertEqual(self.sim.get_users(),set([]))
self.assertEqual(self.sim.CAREER_STAGES, {"Student": 10, "Entry": 15, "Early":25, "Middle": 40, "Late": 10})
def test_simulate(self):
"""Tests the simulate and clear_users methods
Since multiple calls to simulate continue to grow the group,
the test demonstrates that the group grows. It also tests that
the users set can be cleared.
"""
## Simulate the first 5 users
self.sim.simulate(5,1234)
users = list(sorted(
self.sim.get_users(),
cmp=lambda x,y : cmp(x.user_id,y.user_id)
))
print users
expected_results = {
4 : (8698081,"Melinda Thompson"),
3 : (8453420, "Freddie Bahlmann"),
2 : (8198783, "Gary Paddock"),
1 : (3966593, "Kevin Wilson"),
0 : (67423, "Scott Lopez")
}
for i in range(5):
self.assertEqual((users[i].user_id,users[i].name),expected_results[i])
## Clear the simulated users, and resimulate the same users
self.sim.clear_users()
self.sim.simulate(5,1234)
users = list(sorted(
self.sim.get_users(),
cmp=lambda x,y : cmp(x.user_id,y.user_id)
))
print users
expected_results = {
4 : (8698081,"Melinda Thompson"),
3 : (8453420, "Freddie Bahlmann"),
2 : (8198783, "Gary Paddock"),
1 : (3966593, "Kevin Wilson"),
0 : (67423, "Scott Lopez")
}
for i in range(5):
self.assertEqual((users[i].user_id,users[i].name),expected_results[i])
## Add 5 more users on top of existing users.
## Since the same seed is used, it should be tempted
## to duplicate the previous users, but the simulate
## method should prevent this from happening
self.sim.simulate(5,1234)
users = list(sorted(
self.sim.get_users(),
cmp=lambda x,y : cmp(x.user_id,y.user_id)
))
print users
expected_results = {
9 : (8698081,"Melinda Thompson"),
8 : (8698080, "Angela Jones"),
7 : (8453420, "Freddie Bahlmann"),
6 : (8453419, "Marilynn Orr"),
5 : (8198783, "Gary Paddock"),
4 : (8198781, "Tiffany Mchaney"),
3 : (3966593, "Kevin Wilson"),
2 : (3966592, "Donna Yu"),
1 : (67424, "Patrice Newell"),
0 : (67423, "Scott Lopez")
}
for i in range(5):
self.assertEqual((users[i].user_id,users[i].name),expected_results[i])
def test_blank_simulator(self):
blankSim = Simulator()
self.assertEqual(blankSim.get_users(),set([]))
self.assertEqual(blankSim.CAREER_STAGES, {"Student": 10, "Entry": 15, "Early":25, "Middle": 40, "Late": 10})
