def test_extract_vectors_from_dict_returns_expected_numpy_arrays():
given = {
(2, "Walking"): ((186, 15.5), (6, 0.5)),
(2, "Standing"): ((0, 0), (6, 0.5))
}
expected = (np.array([186, 6, 0, 6]), np.array([15.5, 0.5, 0.0, 0.5]))
result = features.extract_vectors_from_dict(given)
assert_array_equal(result[0], expected[0])
assert_array_equal(result[1], expected[1])
def feature_histograms_for_activities(interval_features, activities, bins):
for activity in activities:
activity_intervals = parse.collect_dict_values_by_key_content(interval_features, activity)
feature_vectors = features.extract_vectors_from_dict(activity_intervals)
for i, vector in enumerate(feature_vectors):
plt.figure(figsize=(15, 5))
plt.hist(
vector,
bins=bins,
normed=1, facecolor='blue', alpha=0.5)
plt.title(activity + " x_" + str(i))
plt.show()
def test_extract_vectors_from_dict_handles_three_entry_case():
given = {
(1, 'Downstairs'): (
(85, 12),
(81, 12),
(70, 12),
),
}
expected = (np.array([85, 81, 70]), np.array([12, 12, 12]))
result = features.extract_vectors_from_dict(given)
assert_array_equal(result[0], expected[0])
assert_array_equal(result[1], expected[1])
def test_extract_vectors_from_dict_handles_keys_mapping_to_empty_tuples():
given = {
(2, "Walking"): ((186, 15.5), (6, 0.5)),
(2, "Jogging"): (),
(2, "Standing"): ((0, 0), (6, 0.5)),
(3, "Standing"): ((0, 0), (6, 0.5))
}
expected = (np.array([186, 6, 0, 6, 0,
6]), np.array([15.5, 0.5, 0.0, 0.5, 0, 0.5]))
result = features.extract_vectors_from_dict(given)
assert_array_equal(result[0], expected[0])
assert_array_equal(result[1], expected[1])
def feature_means_and_variances(
data: Dict[Tuple[int, str], Sequence[Sequence[float]]],
activities: Set[str]
) -> Dict[str, Sequence[Tuple[float, float]]]:
out = dict()
for activity in activities:
activity_features = parse.collect_dict_values_by_key_content(data, activity)
value_vectors = features.extract_vectors_from_dict(activity_features)
means_and_variances = []
for vector in value_vectors:
means_and_variances.append((float(np.mean(vector)), float(np.var(vector))))
out[activity] = means_and_variances
return out
def feature_scatter_for_activities(interval_features, activities, colours):
fig, ax = plt.subplots(figsize=(10, 10))
for i, activity in enumerate(activities):
activity_intervals = parse.collect_dict_values_by_key_content(interval_features, activity)
x = features.extract_vectors_from_dict(activity_intervals)
ax.scatter(x[0], x[1], c=colours[i], alpha=0.4, marker='.', label=activity)
ax.set_xlim((0, 200))
ax.set_ylim((0, 40))
x0, x1 = ax.get_xlim()
y0, y1 = ax.get_ylim()
ax.set_aspect(abs(x1 - x0) / abs(y1 - y0))
ax.set_title("")
plt.xlabel("mean absolute magnitude change per second")
plt.ylabel("mean angle change per second")
ax.legend()
plt.show()