-
Notifications
You must be signed in to change notification settings - Fork 0
/
gpx_stats.py
344 lines (266 loc) · 11 KB
/
gpx_stats.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
import collections
from typing import List, Dict, Union, Generator, Iterable
import statistics
from gpxpy import parse
from gpxpy.gpx import GPX, GPXTrackSegment
import numpy as np
from config import DataPreparationConfig
from gpx_data_utils import gpx_segment_to_array
class PathFeature:
"Wrapper class for storing array data in pd.DataFrame."
def __init__(self, data: np.array):
self.array_data = data
@property
def shape(self):
return self.array_data.shape
@property
def data(self):
return self.array_data
def convert_path_to_feature(segment: GPXTrackSegment, num_points_path: int) -> np.array:
"""
Convert GPX track segment to feature.
:param: segment: GPX track segment
:param: num_points_path: Maximal length of track points in path feature
:return: Rotated and normalized GPX track segment
"""
def _get_rotation_matrix(phi):
cos_phi, sin_phi = np.cos(phi), np.sin(phi)
return np.array([np.array([cos_phi, -sin_phi]), np.array([sin_phi, cos_phi])])
data = np.zeros((num_points_path, 3))
assert len(segment.points) > 0, "Path does not contain any points"
assert (
len(segment.points) <= num_points_path
), f"Path too long, got {len(segment.points)} and expected less than {num_points_path}"
data[: len(segment.points)] = gpx_segment_to_array(segment)
# Normalize coordinates
data[: len(segment.points)] -= data[0]
# Compute center of gravity of path
center = np.sum(data, axis=0) / len(segment.points)
phi = np.arctan2(center[1], center[0])
m = _get_rotation_matrix(
-phi
) # Return points with angle that maps center to x-axis
for idx in range(len(segment.points)):
data[idx][0:2] = np.dot(m, data[idx][0:2])
return data
def convert_paths_to_array(path_features: List[PathFeature]) -> np.array:
"""
Convert a list of PathFeature objects to an array.
:param: path_features: List of PathFeature objects
:return: Array of shape (len(path_features), path_features[0].shape)
"""
return np.stack([path_data.data for path_data in path_features], axis=0)
class GpxSegmentStats(object):
"Object collecting statistical properties of a GPX segment."
def __init__(self, segment: GPXTrackSegment, num_points_path: int = 25) -> None:
"""
Construct GpxSegmentStats object.
:param: segment: GPX track segment
:param: num_points_path: Maximal length of points in path features
"""
self.name = getattr(segment, "name", "NotAvailable")
self.length2d = segment.length_2d()
self.length3d = segment.length_3d()
self.duration = (
float(segment.get_duration())
if segment.get_duration() is not None
else float(-1)
)
segment_moving_data = segment.get_moving_data()
assert segment_moving_data
self.moving_time = segment_moving_data.moving_time
self.stopped_time = segment_moving_data.stopped_time
self.total_uphill = segment.get_uphill_downhill().uphill
self.total_downhill = segment.get_uphill_downhill().downhill
self.path = convert_path_to_feature(segment, num_points_path)
def to_list(self) -> List[Union[float, PathFeature]]:
"Convert GpxSegmentStats object data to list."
entry = [
self.length2d,
self.length3d,
self.duration,
self.moving_time,
self.stopped_time,
self.total_uphill,
self.total_downhill,
self.path,
]
return entry
def to_dict(self) -> Dict[str, Union[float, PathFeature]]:
"Convert GpxSegmentStats object data to dictionary."
entry = {
"Length2d": self.length2d,
"Length3d": self.length3d,
"MovingTime": self.moving_time,
"StoppedTime": self.stopped_time,
"Duration": self.duration,
"TotalUphill": self.total_uphill,
"TotalDownhill": self.total_downhill,
"Path": self.path,
}
return entry
@classmethod
def get_header(cls):
"Return names of data entries."
return [
"Length2d",
"Length3d",
"Duration",
"MovingTime",
"StoppedTime",
"TotalUphill",
"TotalDownhill",
"Path",
]
def parse_gpx_files(file_name_list: List[str]) -> Generator[GPX, None, None]:
"""
Parse GPX files and yield their content
:param: file_name_list: List of GPX file names
"""
for file_name in file_name_list:
with open(file_name, "r") as gpx_file:
yield parse(gpx_file)
def get_segments(gpx_file_content_list: Iterable[GPX]) -> List[GPXTrackSegment]:
gpx_segments_list = []
for gpx_file_content in gpx_file_content_list:
for track in gpx_file_content.tracks:
for segment in track.segments:
gpx_segments_list.append(segment)
print("Finished reading", len(gpx_segments_list), "segments.")
return gpx_segments_list
def parse_gpx_files_return_segments(file_name_list: List[str]) -> List[GPXTrackSegment]:
"""
Parse GPX files and return list of GPX track segments
:param: file_name_list: List of GPX files
:return: List of GPXTrackSegment objects
"""
gpx_file_contents: Iterable[GPX] = parse_gpx_files(file_name_list)
return get_segments(gpx_file_contents)
def smoothen_coordinates(
segments_list: List[GPXTrackSegment], window_size: int = 3
) -> None:
"""
Smoothen coordinates of GPX track segment inplace
This function smoothens the coordinates in a GPX track, keeping other track information like time stamps
:param: segments_list: List of GPX track segments
:param: window_size: Size of moving window for averaging
"""
assert window_size % 2, "Window size should be an odd number, {} given.".format(
window_size
)
half_window_size = (window_size - 1) // 2
for idx in range(len(segments_list)):
cloned_segment = segments_list[idx].clone()
if len(segments_list[idx].points) >= window_size:
for i in range(
half_window_size, len(segments_list[idx].points) - half_window_size
):
segments_list[idx].points[i].longitude = statistics.mean(
[
point.longitude
for point in cloned_segment.points[
(i - half_window_size) : (i + half_window_size + 1)
]
]
)
segments_list[idx].points[i].latitude = statistics.mean(
[
point.latitude
for point in cloned_segment.points[
(i - half_window_size) : (i + half_window_size + 1)
]
]
)
# Smoothen elevations only if all points have elevation information
elevations = [
point.elevation
for point in cloned_segment.points[
(i - half_window_size) : (i + half_window_size + 1)
]
]
if all(elevations):
segments_list[idx].points[i].elevation = statistics.mean(elevations)
del segments_list[idx].points[
:half_window_size
] # Delete first and last points
del segments_list[idx].points[-half_window_size:]
def filter_segments(
gpx_segments_list: List[GPXTrackSegment], min_distance_m: float = 5
) -> List[GPXTrackSegment]:
"""
Filter points from GPX track segments that are too close to each other
:param: gpx_segments_list: List of GPX track segments to be processed
:param: min_distance_m: Minimum distance between consecutive track points after filtering
:return: List of GPX track segments that are all shorter than max_length_m
"""
gpx_filtered_segments_list = []
for segment in gpx_segments_list:
points = [segment.points[0]]
last_kept = segment.points[0]
for point in segment.points[1:]:
if GPXTrackSegment([last_kept, point]).length_2d() > min_distance_m:
points.append(point)
last_kept = point
gpx_filtered_segments_list.append(GPXTrackSegment(points))
return gpx_filtered_segments_list
def split_segments_by_length(
gpx_segments_list: List[GPXTrackSegment], *, max_length_m: float
) -> List[GPXTrackSegment]:
"""
Split GPX track segments until all are shorter than max_length_m
:param: gpx_segments_list: List of GPX track segments to be processed
:param: max_length_m: Maximum length of track segment above which the segment should be split
:return: List of GPX track segments that are all shorter than max_length_m
"""
gpx_split_segments_list = []
buffer = collections.deque(gpx_segments_list)
while len(buffer) > 0:
segment = buffer.popleft()
if segment.length_2d() < max_length_m:
gpx_split_segments_list.append(segment)
else:
buffer.append(GPXTrackSegment(segment.points[: len(segment.points) // 2]))
buffer.append(GPXTrackSegment(segment.points[len(segment.points) // 2 :]))
return gpx_split_segments_list
def extract_stats(
gpx_segments_list: List[GPXTrackSegment], num_points_path: int = 25
) -> List[GpxSegmentStats]:
"""
Extract some properties of GPX track segments
:param: gpx_segments_list: List of GPX track segments to be processed
:param: num_points_path: Maximum number of points in path features
:return: List of GPX track segments that are all shorter than max_length_m
"""
gpx_stats = []
for segment in gpx_segments_list:
segment_stats = GpxSegmentStats(segment, num_points_path)
if segment_stats.moving_time >= segment_stats.stopped_time:
gpx_stats.append(segment_stats)
return gpx_stats
def filter_bad_segments(
gpx_segments_list: List[GPXTrackSegment],
data_preparation_config: DataPreparationConfig,
) -> List[GPXTrackSegment]:
"""
Filter segments with obviously insensible data
This function filters segments with unreasonable data, for example large differences in elevation, arising from
bad measurements.
Args:
gpx_segments_list: List of GPX segments
data_preparation_config: Config for data preparation
Returns:
Filtered list of GPX segments
"""
def _elevation_predicate(
segment: GPXTrackSegment, max_elevation_diff: float
) -> bool:
uphill, downhill = segment.get_uphill_downhill()
return uphill < max_elevation_diff and downhill < max_elevation_diff
return list(
filter(
lambda segment: _elevation_predicate(
segment, data_preparation_config.max_elevation_diff_m
),
gpx_segments_list,
)
)