def test_random_noisy_interactions(self):
intrs = random_noisy_interactions(
self.params['n_noisy_interactions'],
self.params['min_time'],
self.params['max_time'],
self.params['n_total_participants'],
self.params['n_topics'],
self.params['topic_noise'],
)
assert_equal(self.params['n_noisy_interactions'],
len(intrs))
topic_mean = np.mean([i['topics'] for i in intrs], axis=0)
self.seems_like_uniform_distribution(topic_mean)
np.testing.assert_almost_equal(
305,
np.mean([i['timestamp'] for i in intrs]),
decimal=-3
)
freq = Counter(itertools.chain(
*[[i['sender_id']] + i['recipient_ids']
for i in intrs]
))
freq = np.array(freq.values(),
dtype=np.float64)
self.seems_like_uniform_distribution(freq / freq.sum())
def test_random_noisy_interactions_with_taboo(self):
intrs = random_noisy_interactions(
self.params['n_noisy_interactions'],
self.params['min_time'],
self.params['max_time'],
self.params['n_total_participants'],
self.params['n_topics'],
self.params['topic_noise'],
taboo_topics=range(9)
)
topics = set([i['topics'].argmax() for i in intrs])
assert_equal(set(range(10)), topics)
