def dtw_compare(label_json, input_json):
with open(label_json) as f:
label = json.load(f)
with open(input_json) as f:
ip_data = json.load(f)
label_list = []
ip_list = []
score = []
score_json = {}
for _ in range(17):
temp_x = []
temp_y = []
for frame in range(len(label)):
temp = label[frame]['keypoints']
temp_x.append(temp[_ * 3])
temp_y.append(temp[_ * 3 + 1])
label_list.append(temp_x)
label_list.append(temp_y)
for _ in range(17):
temp_x = []
temp_y = []
for frame in range(len(ip_data)):
temp = ip_data[frame]['keypoints']
temp_x.append(temp[_ * 3])
temp_y.append(temp[_ * 3 + 1])
ip_list.append(temp_x)
ip_list.append(temp_y)
for _ in range(17):
score_x = dtw.distance(label_list[_ * 2], ip_list[_ * 2])
score_y = dtw.distance(label_list[_ * 2 + 1], ip_list[_ * 2 + 1])
score_temp = []
score_temp.append(100 - (score_x * 100))
score_temp.append(100 - (score_y * 100))
score.append(np.mean(score_temp))
score_json['Total_Score'] = np.mean(score)
score_json['Head'] = np.mean(score[0:5])
score_json['LShoulder'] = score[5]
score_json['RShoulder'] = score[6]
score_json['LElbow'] = score[7]
score_json['RElbow'] = score[8]
score_json['LWrist'] = score[9]
score_json['RWrist'] = score[10]
score_json['LHip'] = score[11]
score_json['RHip'] = score[12]
score_json['LKnee'] = score[13]
score_json['Rknee'] = score[14]
score_json['LAnkle'] = score[15]
score_json['RAnkle'] = score[16]
for (k, v) in score_json.items():
print(k, ' : ', v, '\n')
return score_json
def test_overflowdistance():
maxvalthirtytwobit = 2147483647
s = np.array([
[maxvalthirtytwobit, maxvalthirtytwobit, 1, 2, 1, 0, 1, 0, 0],
[1., 2, 0, 0, 0, 0, 0, 1, 0]])
d1 = dtw.distance(s[0], s[1], use_c=False)
d2 = dtw.distance(s[0], s[1], use_c=True)
print(d1)
print(d2)
def test_penalty_cyclicalshift():
s1 = np.array([0., 1, 2, 1, 0, 1, 2, 1, 0, 1, 2, 1, 0])
s2 = np.array([2., 1, 0, 1, 2, 1, 0, 1, 2, 1, 0, 1, 2])
# plt.plot(s1)
# plt.plot(s2)
# plt.show(block=True)
d1 = dtw.distance(s1, s2)
d2 = dtw.distance(s1, s2, penalty=1)
assert (d1) == pytest.approx(math.sqrt(10))
assert (d2) == pytest.approx(math.sqrt(14))
def get_distances(keypoints1, keypoints2):
distances = list({i: {
"dist_x": [],
"dist_y": []
}} for i in range(len(keypoints1)))
for i in range(len(distances)):
distances[i][i]["dist_x"].append(
dtw.distance(keypoints1[i][i]["x"], keypoints2[i][i]["x"]))
distances[i][i]["dist_y"].append(
dtw.distance(keypoints1[i][i]["y"], keypoints2[i][i]["y"]))
return distances
def test_psi_dtw_1b():
with util_numpy.test_uses_numpy() as np:
x = np.arange(0, 20, .5)
s1 = np.sin(x)
s2 = np.sin(x - 1)
d = dtw.distance(s1, s2, psi=2)
np.testing.assert_equal(d, 0.0)
def get_dtw_wrapping_path(data, col1, col2):
'''
input:
data: dataframe 源数据
col1: 列名
col2：列名
output: 输出图像，上方是col1，下方是col2
'''
indicators = [i for i in data.columns if i not in 'date']
array_subset = data[indicators].values
array_subset_zscore = stats.zscore(array_subset)
array_subset_zscore_T = array_subset_zscore.T
x_idx = indicators.index(col1)
y_idx = indicators.index(col2)
# x = array_for_dtw_zscore_T[col1,:]
# y = array_for_dtw_zscore_T[col2,:]
x = array_subset_zscore_T[x_idx, :]
y = array_subset_zscore_T[y_idx, :]
path = dtw.warping_path(x, y)
outname = col1 + 'vs' + col2
ds_xy = dtw.distance(x, y)
dtwvis.plot_warping(
x,
y,
path,
filename=
"D:/Pythoncode/JD_mart/operation_flow_distribution/DTW_for_business/results/%s.png"
% outname)
print("%s 和 %s 的DTW距离: %2.4f" % (col1, col2, ds_xy))
def test_bug2():
with util_numpy.test_uses_numpy() as np:
s1 = np.array([0, 0, 1, 2, 1, 0, 1, 0, 0], dtype=np.double)
s2 = np.array([0.0, 1, 2, 0, 0, 0, 0, 0, 0])
d1a = dtw.distance_fast(s1, s2, window=2)
d1b = dtw.distance(s1, s2, window=2)
if directory:
fn = directory / "warpingpaths.png"
else:
file = tempfile.NamedTemporaryFile()
fn = Path(file.name + "_warpingpaths.png")
d2, paths = dtw.warping_paths(s1, s2, window=2)
best_path = dtw.best_path(paths)
if not dtwvis.test_without_visualization():
dtwvis.plot_warpingpaths(s1,
s2,
paths,
best_path,
filename=fn,
shownumbers=False)
print("Figure saved to", fn)
assert d1a == pytest.approx(d2)
assert d1b == pytest.approx(d2)
def filter_cuts(shifted_observed_runs: List[int],
expected_runs: List[int]) -> Tuple[List[int], List[int], int]:
"""
Applies dynamic time warping to select cuts in the observed pixel sequence of whitespace runs according to the expected runs
:param shifted_observed_runs: histogram of whitespace runs encoded as follows: position = start of run, value = length of run
:param expected_runs: histogram of expected word cuts (see: `expected_runs_for_line`)
:return: A triplet `(cuts, cuts_indices, distance)` where:
* cuts: list of x-coordinates of word cuts
* cuts_indices: indices of selected cuts
* distance: DTW distance between the two input sequences
"""
path = dtw.warping_path(expected_runs, shifted_observed_runs)
distance = dtw.distance(expected_runs, shifted_observed_runs)
runs_indices = [i for i, x in enumerate(shifted_observed_runs) if x > 0]
runs_indices.insert(0, 0)
cuts = []
cuts_indices = []
for i, j in path:
if expected_runs[i] > 0:
cuts.append(j)
index_found = j in runs_indices
if not index_found:
print(
f"DTW associated expected peak in {i} to zero value in {j}",
file=sys.stderr)
else:
cuts_indices.append(runs_indices.index(j))
return cuts, cuts_indices, distance
def test_distance3_a():
dist_opts = {"penalty": 0.005, "max_step": 0.011, "window": 3}
s = np.array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.005, 0.01, 0.015, 0.02, 0.01, 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.])
p = np.array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.005, 0.01, 0.015, 0.02, 0.01, 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.])
d1 = dtw.distance(s, p, **dist_opts)
d2 = dtw_c.distance_nogil(s, p, **dist_opts)
assert d1 == pytest.approx(d2)
def _calculate_dtw_over_time_window(window, minimum_close, maximum_close,
pattern, start_window, end_window,
threshold):
calculations = []
minimum_value_pattern, maximum_value_pattern = pattern.get_min_max()
normalized_data = [
normalize_function_paper(close_price, minimum_close, maximum_close,
minimum_value_pattern, maximum_value_pattern)
for close_price in window
]
# distance, path = fastdtw(normalized_data, pattern.get_pattern(), dist=euclidean)
# alignment = dtw(normalized_data, pattern.get_pattern(), keep_internals=True)
distance = dtw.distance(normalized_data, pattern.get_pattern())
path = dtw.warping_path(normalized_data, pattern.get_pattern())
if distance < threshold:
calculations.append({
"pattern_name": pattern.__class__.__name__,
"start_window": start_window,
"start_date": list_of_dates[start_window],
"end_window": end_window,
"end_date": list_of_dates[end_window],
"distance": distance,
"path": path,
"normalized_data": normalized_data,
"pattern": pattern.get_pattern()
})
return calculations
def dtwDistance(mainInput, targetInput, mainFile, targetFile):
distance = dtw.distance(mainInput, targetInput)
resultDict = {}
resultDict['mainFile'] = mainFile
resultDict['targetFile'] = targetFile
resultDict['distance'] = distance
return resultDict
def dtwdis(input_points, model_points, i, j):
input_pts = input_points.reshape(2 * j,)
model_pts = model_points.reshape(2 * i,)
model_pts = model_pts / np.linalg.norm(model_pts)
input_pts = input_pts / np.linalg.norm(input_pts)
return percentage_score(dtw.distance(model_pts, input_pts))
def test_distance1_b():
dist_opts = {}
s1 = np.array([0., 0.01, 0., 0.01, 0., 0., 0., 0.01, 0.01, 0.02, 0., 0.])
s2 = np.array([0., 0.02, 0.02, 0., 0., 0.01, 0.01, 0., 0., 0., 0.])
d1 = dtw.distance(s1, s2, **dist_opts)
d2 = dtw_c.distance_nogil(s1, s2, **dist_opts)
assert d1 == d2
assert d1 == pytest.approx(0.02)
def test_distance2_c():
dist_opts = {}
s1 = np.array([0.0, 0.0, 2.0, 1.0, 1.0, 0.0, 0.0])
s2 = np.array([0.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0])
d1 = dtw.distance(s1, s2, **dist_opts)
d2 = dtw_c.distance_nogil(s1, s2, **dist_opts)
assert d1 == d2
assert d1 == pytest.approx(1.0)
def test_distance2_c():
with util_numpy.test_uses_numpy() as np:
dist_opts = {}
s1 = np.array([0.0, 0.0, 2.0, 1.0, 1.0, 0.0, 0.0])
s2 = np.array([0.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0])
d1 = dtw.distance(s1, s2, **dist_opts)
d2 = dtw.distance_fast(s1, s2, **dist_opts)
assert d1 == d2
assert d1 == pytest.approx(1.0)
def nearest_to_centroid(xs):
centroid = mean(xs)
dists = []
for x in xs:
dists.append(dtw.distance(x, centroid))
min_i = np.argmin(dists)
return xs[min_i]
def test_distance2_bb():
dist_opts = {'max_step': 0.1}
s1 = np.array([0.0, 0.0, 2.0, 1.0, 1.0, 0.0, 0.0])
s2 = np.array([0.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0])
d1 = dtw.distance(s1, s2, **dist_opts)
d2 = dtw_c.distance_nogil(s1, s2, **dist_opts)
print(d1, d2)
assert d1 == d2
assert d1 == pytest.approx(np.inf)
def test_distance1_a():
# dist_opts = {'max_dist': 0.201, 'max_step': 0.011, 'max_length_diff': 8, 'window': 3}
dist_opts = {'window': 3}
s1 = np.array([0., 0.01, 0., 0.01, 0., 0., 0., 0.01, 0.01, 0.02, 0., 0.])
s2 = np.array([0., 0.02, 0.02, 0., 0., 0.01, 0.01, 0., 0., 0., 0.])
d1 = dtw.distance(s1, s2, **dist_opts)
d2 = dtw_c.distance_nogil(s1, s2, **dist_opts)
assert d1 == d2
assert d1 == pytest.approx(0.02)
def test_distance2_bb():
with util_numpy.test_uses_numpy() as np:
dist_opts = {'max_step': 0.1}
s1 = np.array([0.0, 0.0, 2.0, 1.0, 1.0, 0.0, 0.0])
s2 = np.array([0.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0])
d1 = dtw.distance(s1, s2, **dist_opts)
d2 = dtw.distance_fast(s1, s2, **dist_opts)
print(d1, d2)
assert d1 == d2
assert d1 == pytest.approx(np.inf)
def test_distance1_a():
with util_numpy.test_uses_numpy() as np:
# dist_opts = {'max_dist': 0.201, 'max_step': 0.011, 'max_length_diff': 8, 'window': 3}
dist_opts = {'window': 3}
s1 = np.array(
[0., 0.01, 0., 0.01, 0., 0., 0., 0.01, 0.01, 0.02, 0., 0.])
s2 = np.array([0., 0.02, 0.02, 0., 0., 0.01, 0.01, 0., 0., 0., 0.])
d1 = dtw.distance(s1, s2, **dist_opts)
d2 = dtw.distance_fast(s1, s2, **dist_opts)
print("X")
assert d1 == d2
assert d1 == pytest.approx(0.02)
def computeSimilaryByDTW(self):
if util.getDayCounts(self._df) < 49:
print("days: ", util.getDayCounts(self._df))
print("The duration is too short!!!")
return
self._dayDf, self._weekDf = util.getRecordByDay(self._df)
if self.isEmptyGas():
return
#print(self_dayDf)
weekMean = util.getWeekMeanValue(self._weekDf)
weekMean = util.normalizeByZeroOne(weekMean)
dis = [
dtw.distance(weekMean, features.Stable_Feature),
dtw.distance(weekMean, features.Concave_Feature),
dtw.distance(weekMean, features.Convexity_Feature),
dtw.distance(weekMean, features.Slop_Feature),
dtw.distance(weekMean, features.Slop_Reverse_Feature),
dtw.distance(weekMean, features.V_Feature)
]
for i in range(len(dis)):
if dis[i] < self._similar:
self._similar = dis[i]
self._category = i
def test_distance1_b():
with util_numpy.test_uses_numpy() as np:
dist_opts = {}
s1 = np.array(
[0., 0.01, 0., 0.01, 0., 0., 0., 0.01, 0.01, 0.02, 0., 0.])
s2 = np.array([0., 0.02, 0.02, 0., 0., 0.01, 0.01, 0., 0., 0., 0.])
d1 = dtw.distance(s1, s2, **dist_opts)
d2 = dtw.distance_fast(s1, s2, **dist_opts)
d3, wps = dtw.warping_paths(s1, s2, **dist_opts)
print(np.power(wps, 2))
assert d1 == d2
assert d1 == d3
assert d1 == pytest.approx(0.02)
def distance_control(kpi_series_normal, kpi_series_abnormal):
time_series_normal = []
for row in kpi_series_normal:
temp = float(row[1])
time_series_normal.append(temp)
print(time_series_normal)
time_series_abnormal = []
for row in kpi_series_abnormal:
temp = float(row[1])
time_series_abnormal.append(temp)
print(time_series_abnormal)
distance = dtw.distance(time_series_normal, time_series_abnormal)
print(distance)
return distance
def compute_dtw(county_covid_cases_delta, mobility_type_data, mobility_data_type_index):
print(county_covid_cases_delta.shape)
print(mobility_type_data.shape)
mobility_data_type_name = ['recreation', 'groecry', 'park', 'transit', 'work', 'residential']
distance = dtw_visualize.distance(county_covid_cases_delta, mobility_type_data)
print(distance)
with open('results/confirmed_cases_dtw_distance.csv','a') as csvfile:
csvfile.write(mobility_data_type_name[mobility_data_type_index] + " " + str(distance) + "\n")
d, paths = dtw_visualize.warping_paths(county_covid_cases_delta, mobility_type_data, window=25, psi=2)
best_path = dtw_visualize.best_path(paths)
dtwvis.plot_warpingpaths(county_covid_cases_delta, mobility_type_data, paths, best_path, filename="results/dtw_" + mobility_data_type_name[mobility_data_type_index] +".png")
plt.title("confirmed cases vs " + mobility_data_type_name[mobility_data_type_index] + " DTW")
def find_dynamic_time_warp(df, time_var, y_var, end_training_period):
texto(' ')
st.markdown('---')
texto(f'Dynamic time warp con {y_var}', 17)
texto('Sugerencia: series con mayor distancia son diferentes entre sí.',
12, 'grey')
df_filtered = df.query(f'{time_var} <= "{end_training_period}"')
new_frame = pd.DataFrame(
index=[m for m in df.columns if m != time_var and 'scaled' not in m])
for col in df.columns:
if 'scaled' not in col and col != time_var:
new_frame.loc[col, 'dtw distance'] = dtw.distance(
df_filtered[y_var].values, df_filtered[col].values)
new_frame.sort_values(by='dtw distance', ascending=False, inplace=True)
st.write(new_frame)
def test_distance3_a():
with util_numpy.test_uses_numpy() as np:
dist_opts = {"penalty": 0.005, "max_step": 0.011, "window": 3}
s = np.array([
0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.005, 0.01,
0.015, 0.02, 0.01, 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
0.
])
p = np.array([
0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.005, 0.01, 0.015,
0.02, 0.01, 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,
0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.
])
d1 = dtw.distance(s, p, **dist_opts)
d2 = dtw.distance_fast(s, p, **dist_opts)
assert d1 == pytest.approx(d2)
def compute_dwt_dist_between_ts_and_list(single_ts, ts_list, r):
"""
This function computes the list of DTW distances between a single time-series and a list of time-series.
:param single_ts: single time-series
:type single_ts: 1d array
:param ts_list: list of time-series
:type ts_list: list of 1d array
:param r: distance upper bound (for DTW distance computation)
:type r: float
:return: list of DTW distances
:rtype: list of float
"""
dist_list = []
for ts in ts_list:
dist = dtw.distance(single_ts, ts, max_dist=r)
dist_list.append(dist)
return dist_list
def getdisance(data, datalen):
# print data
dismatrix = [[0 for i in range(datalen)] for i in range(datalen)]
for i in range(datalen):
for j in range(datalen):
use_dtw = 0
dis = 0
if use_dtw == 0:
for l in range(len(data[0])):
dis = dis + (data[i][l] - data[j][l]) * (data[i][l] -
data[j][l])
dis = math.sqrt(dis)
else:
dis = dtw.distance(data[i], data[j])
dismatrix[i][j] = dis
dismatrix[j][i] = dis
return dismatrix
def make_sequence(k, model, dist_type, df_test, label_df_train, label_df_test):
arr_test_full, arr_test_norm = get_array(df_test, 24)
for i in range(len(arr_test_full)):
distance = []
for cn in range(k):
if dist_type == 'ed':
dist = np.linalg.norm(arr_test_full[i] -
model.cluster_centers_[cn])
else:
dist = dtw.distance(arr_test_full[i],
model.cluster_centers_[cn])
distance.append(dist)
cluster = np.argmin(distance)
label_df_test['label'][i] = cluster
cluster_sequence = pd.concat([label_df_train, label_df_test], axis=0)
cluster_sequence.sort_index(inplace=True)
return cluster_sequence
def main():
s1 = np.array([0., 0, 1, 2, 1, 0, 1, 0, 0, 0, 2, 1, 0, 0])
s2 = np.array([0., 1, 2, 3, 1, 0, 0, 0, 2, 1, 0, 0, 0])
path = dtw.warping_path(s1, s2)
dtwvis.plot_warping(s1, s2, path)
plt.figure(1)
plt.subplot(211)
plt.title('Timeseries: s1 & s2')
plt.plot(s1)
plt.subplot(212)
plt.plot(s2)
plt.show()
dist = dtw.distance(s1, s2)
print(dist)
plt.figure(2)
d, paths = dtw.warping_paths(s1, s2, window=3, psi=2)
best_path = dtw.best_path(paths)
dtwvis.plot_warpingpaths(s1, s2, paths, best_path)