def main():
aggregated_recommendations = aggregate_recommendations(verbose=False)
stats = count_occurrences(aggregated_recommendations,
verbose=False)
app_ids, pb_occurrences_dict, rb_occurrences_dict = summarize_occurrences(aggregated_recommendations,
stats,
chosen_num_occurrences=max(stats.keys()),
verbose=True)
for app_id in app_ids:
data = extract_data_with_equal_release_recency_bias(app_id,
aggregated_recommendations,
rb_occurrences_dict,
verbose=True)
data = extract_data_with_equal_popularity_bias(app_id,
aggregated_recommendations,
pb_occurrences_dict,
verbose=True)
data = extract_data(app_id,
aggregated_recommendations,
verbose=True)
return
def test_get_total_num_apps(self):
aggregated_recommendations = inverse_problem.aggregate_recommendations(
verbose=True)
stats = inverse_problem.count_occurrences(aggregated_recommendations,
verbose=True)
total_num_apps = inverse_problem.get_total_num_apps(stats,
verbose=True)
expected_total_num_apps = len(aggregated_recommendations)
self.assertEqual(total_num_apps, expected_total_num_apps)
def test_get_total_num_occurrences(self):
aggregated_recommendations = inverse_problem.aggregate_recommendations(
verbose=True)
stats = inverse_problem.count_occurrences(aggregated_recommendations,
verbose=True)
total_num_occurrences = inverse_problem.get_total_num_occurrences(
stats, verbose=True)
num_rankings, ranking_size = inverse_problem.count_rankings(
verbose=True)
expected_total_num_occurrences = num_rankings * ranking_size
self.assertEqual(total_num_occurrences, expected_total_num_occurrences)
def test_extract_data(self):
aggregated_recommendations = inverse_problem.aggregate_recommendations(
verbose=False)
app_ids, pb_occurrences_dict, rb_occurrences_dict = inverse_problem.summarize_occurrences(
aggregated_recommendations, verbose=False)
for app_id in app_ids:
data = extract_regression_data.extract_data(
app_id, aggregated_recommendations, verbose=True)
X = data['X']
y = data['y']
self.assertEqual(len(X), len(y))
def test_extract_data_with_equal_popularity_bias(self):
aggregated_recommendations = inverse_problem.aggregate_recommendations(
verbose=False)
app_ids, pb_occurrences_dict, rb_occurrences_dict = inverse_problem.summarize_occurrences(
aggregated_recommendations, verbose=False)
for app_id in app_ids:
data = extract_regression_data.extract_data_with_equal_popularity_bias(
app_id,
aggregated_recommendations,
pb_occurrences_dict,
verbose=True)
for pb_argmax in data:
X = data[pb_argmax]['X']
y = data[pb_argmax]['y']
self.assertEqual(len(X), len(y))
def test_summarize_occurrences(self):
aggregated_recommendations = inverse_problem.aggregate_recommendations(
verbose=True)
app_ids, pb_occurrences_dict, rb_occurrences_dict = inverse_problem.summarize_occurrences(
aggregated_recommendations, verbose=True)
self.assertGreater(len(app_ids), 0)
self.assertEqual(len(app_ids), len(pb_occurrences_dict))
self.assertEqual(len(app_ids), len(rb_occurrences_dict))
pb_val = inverse_problem.get_popularity_bias_range()
rb_val = inverse_problem.get_release_recency_bias_range()
self.assertTrue(
all(
len(occurrences) == len(pb_val)
for occurrences in pb_occurrences_dict.values()))
self.assertTrue(
all(
len(occurrences) == len(rb_val)
for occurrences in rb_occurrences_dict.values()))
def test_count_occurrences(self):
aggregated_recommendations = inverse_problem.aggregate_recommendations(
verbose=True)
stats = inverse_problem.count_occurrences(aggregated_recommendations,
verbose=True)
self.assertGreater(len(stats), 0)
def test_aggregate_recommendations(self):
aggregated_recommendations = inverse_problem.aggregate_recommendations(
verbose=True)
self.assertGreater(len(aggregated_recommendations), 0)