-
Notifications
You must be signed in to change notification settings - Fork 2
/
consolidation.py
414 lines (351 loc) · 18.4 KB
/
consolidation.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
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
"""Functions that are used for the consolidation of tempTIMDs into one TIMD.
The main function in this file is consolidate_temp_timds(), which is
called in calculate_timd.py."""
# External imports
import collections
import numpy as np
# No internal imports
def consolidate_times(times):
"""Takes in multiple time options and consolidates them into one time.
times is a dictionary of each scout to their respective time value."""
times = {scout: time for scout, time in times.items() if time is not None}
# Creates a list of the times in the form of floats instead of their
# tempTIMD format of strings. Does this in order to use them for
# calculations.
float_list = [float(time) for time in times.values()]
# Finds the mean and standard deviation for the array of floats,
# these metrics are used for the reciprocal z-scores used later on.
mean = np.mean(float_list)
std = np.std(float_list)
# If the standard deviation is zero, all the times are the same, so
# it just returns the mean.
if std == 0:
return round(mean, 1)
# If one of the float_list items is equivalent to the mean, it's
# weight will be undefined, so we can just return the mean.
if mean in float_list:
return mean
# Creates a list of tuples with the first item as the time and the
# second item as the weight (squared reciprocal of the z-score for
# each time). These values are how much each time is weighted when
# calculating the final weighted average. The lower the value on
# this list the time is, the farther away from the mean it is, and
# the less it is weighted.
reciprocal_zscores = [(number, (1 / ((mean - number) / std)) ** 2)
for number in float_list]
# Multiplies each time by its corresponding reciprocal z-score
# value, creating a weighted time.
weighted_times = [number * zscore_weight for number, zscore_weight
in reciprocal_zscores]
# Adds up all the weighted times and divides it by the sum of the
# reciprocal_zscore_list. Does this in order to get a reasonable
# time, if this step is not taken, the weighted time makes no sense.
weighted_average = sum(weighted_times) / sum([zscore[1] for zscore \
in reciprocal_zscores])
# Formats each average to a float with one decimal place.
return round(weighted_average, 1)
def convert_float_time(time):
"""Converts a time from a string to a float.
time is the time that needs to be converted."""
# If an asterisk is in the time, the time is in the wrong time
# period. If the asterisk time is in teleop, the time period is
# supposed to be in sandstorm, so it sets the time to the lowest
# time in sandstorm, and vice versa for when the time is in
# sandstorm.
if '*' in time:
if float(time[:-1]) >= 135.1:
return 135.0
else:
return 135.1
else:
return float(time)
def max_occurrences(comparison_list, sprking):
"""Takes in a dictionary of scouts to their value and returns the majority.
If there is no clear winner, returns the value for the best spr
scout.
comparison_list is a dictionary of each of the scouts to their input
on a specific decision (value for a data field, amount of actions,
etc).
sprking is the scout with the best spr out of the scouts, used if
there is no clear majority."""
# If the sprking is not part of the comparison_list, another scout
# is randomly selected.
if sprking not in list(comparison_list.keys()):
correct_scout = list(comparison_list.keys())[-1]
else:
correct_scout = sprking
# Each item in the list to how many times it appeared in the list.
# Uses the collections module to count how many appearances each
# item has in the list.
occurence_list = dict(collections.Counter(comparison_list.values()))
# Handling for an empty occurrence list.
if len(occurence_list.values()) == 0:
return None
# If the highest occurrence on the occurrence list is the same as
# the lowest occurrence, the correct value for the data point is
# the value output by the scout with the best spr. This triggers
# both when all the scout values are the same (The max and min
# would both be three) and when all the scout values are
# different (The max and min are both 1). In the case of any
# other scenario, the max is trusted because it would suggest
# the max is the 2 in a 2 scout versus 1 split decision.
elif max(occurence_list.values()) == min(occurence_list.values()):
return comparison_list[correct_scout]
else:
return max(occurence_list, key=occurence_list.get)
def consolidate_timeline_action(temp_timd_timelines, action_type, sprking):
"""Takes an action type out of timelines and consolidates it separately.
Returns a consolidated timeline only made up of the action type that
was passed as action_type.
input_timelines is the dictionary of the scouts to their specific
timelines.
action_type is the action type that the function is consolidating.
sprking is the scout with the best spr out of the scouts, used when
max_occurrences is called."""
# The dictionary of three timelines with only the types specified
# in the function.
simplified_timelines = {scout: [] for scout in temp_timd_timelines.keys()}
# Takes the three different timelines and cuts out any types of
# data points which are not the specified types.
for scout, timeline in temp_timd_timelines.items():
for action in timeline:
if action.get('type') == action_type:
simplified_timelines[scout].append(action)
# For each action in each scouts list of actions, the time is
# converted from a string to a float.
for scout, simplified_timeline in simplified_timelines.items():
for action in simplified_timeline:
action['time'] = convert_float_time(action['time'])
# Scouts to the amount of actions of the specified type are in the
# timeline.
count_timelines = {scout: len(timeline) for
scout, timeline in simplified_timelines.items()}
# Finds the majority amount of actions in the timeline to see
# which amount of actions is the correct amount.
majority_length = max_occurrences(count_timelines, sprking)
# Creates a dictionary of scouts to their timelines which follow the
# majority length of timeline.
correct_length_timelines = {}
for scout, timeline_length in count_timelines.items():
if timeline_length == majority_length:
correct_length_timelines[scout] = simplified_timelines[scout]
# If there are scouts that don't agree with the majority timeline,
# creates a time_reference to line up against.
time_reference = {}
if sprking in correct_length_timelines.keys():
correct_scout = sprking
else:
correct_scout = list(correct_length_timelines.keys())[-1]
reference_timeline = correct_length_timelines[correct_scout]
time_reference[correct_scout] = [action['time'] for action in
reference_timeline]
# If there are scouts that do not agree with the correct timeline
# length, find out which of their action times agree with the time
# reference the best, and line it up against the reference in the
# correct_length_timelines dictionary.
for scout in simplified_timelines.keys():
if scout not in correct_length_timelines.keys():
correct_length_timelines[scout] = [{} for action in
range(majority_length)]
# In order to find the best option for timings, it sets
# up a matrix of time differences between each action in
# each tempTIMD.
timings = np.zeros((len(simplified_timelines[scout]),
majority_length))
for false_index, false_action in \
enumerate(simplified_timelines[scout]):
for comparison_index, comparison_action in \
enumerate(list(time_reference.values())[0]):
timings[false_index][comparison_index] = \
abs(float(comparison_action) -
float(false_action['time']))
# Once the matrix of timing differences has been
# created, the lowest difference is used to line up the
# incorrect length timeline with the correct length
# timeline. To avoid one action being compared with multiple
# other actions, all other instances of the action (The row
# and column) are set to 200 to signify that it has been
# used. 200 is used because it is higher than any possible
# time difference.
if timings.size > 0:
# The loop runs until there are no more time differences
# in the matrix less than 200.
while timings.min() < 200:
# lowest_index is in the format of ([y coordinate],
# [x coordinate]), which requires lowest_index[1][0]
# to get the x coordinate, and lowest_index[0][0]
# for the y coordinate.
lowest_index = np.where(timings == timings.min())
correct_length_timelines[scout][lowest_index[1][0]] = \
simplified_timelines[scout][lowest_index[0][0]]
timings[int(lowest_index[0][0])] = \
np.full([1, len(timings[0])], 200)
for row in range(len(timings)):
timings[row][int(lowest_index[1][0])] = 200
final_simplified_timd = [{} for action in range(majority_length)]
# Iterates through the sprking's timeline to compare all the actions.
# If the majority 'type' for the action is None, the majority of
# scouts did not record this action, and this action should not
# appear in the consolidated TIMD.
for action_index, action in enumerate(correct_length_timelines[sprking]):
comparison_dict = {scout: timeline[action_index] for scout,
timeline in correct_length_timelines.items()}
types = {scout: action.get('type') for scout, action in
comparison_dict.items()}
if max_occurrences(types, sprking) is None:
# Skips current iteration
continue
# Deletes scouts that did not record this action.
for scout in list(comparison_dict):
if comparison_dict[scout] == {}:
comparison_dict.pop(scout)
# All of the possible keys for a tempTIMD for this action.
keys = set()
for action in comparison_dict.values():
for key in action.keys():
keys.add(key)
for key in keys:
# For every key that isn't time, which can't realistically
# have a majority, the majority opinion is set to the final
# timd.
scout_to_keys = {scout: action.get(key) for scout,
action in comparison_dict.items()}
if key == 'time':
# If the key is time, finds the correct time using the
# consolidate_times algorithm.
final_simplified_timd[action_index]['time'] = \
consolidate_times(scout_to_keys)
else:
# For every key in the dictionary other than time, it just
# takes the majority value for the key.
final_simplified_timd[action_index][key] = \
max_occurrences(scout_to_keys, sprking)
# Returns the final created timeline
return final_simplified_timd
def climb_consolidation(input_timelines, sprking):
"""Takes climb out of the timelines of the tempTIMDs and consolidates it.
Returns a timeline only with climb inside it to add to the final
timeline for the timd.
input_timelines is the dictionary of the scouts to their specific
timelines.
sprking is the scout with the best spr out of the scouts, used when
max_occurrences is called. More info in the docstring for
max_occurrences.
"""
# Scout name to climb dictionary.
simplified_timelines = {}
# Fills in 'simplified_timelines' with the scout and the climb
# dictionary from the three tempTIMDs.
for scout, timeline in input_timelines.items():
for action in timeline:
if action.get('type') == 'climb':
simplified_timelines[scout] = action
# Returns None if no climb was recorded.
if simplified_timelines == {}:
return None
final_simplified_timd = {'type': 'climb', 'attempted': {}, 'actual': {}}
# Consolidates time first
final_simplified_timd['time'] = consolidate_times({
scout: convert_float_time(climb['time']) for scout,
climb in simplified_timelines.items()})
for key in ['attempted', 'actual']:
for robot in ['self', 'robot1', 'robot2']:
final_simplified_timd[key][robot] = max_occurrences({
scout: climb[key][robot] for scout, climb in
simplified_timelines.items()}, sprking)
# Returns the final created timeline
return final_simplified_timd
def consolidate_temp_timds(temp_timds):
"""Consolidates between 1-3 temp_timds into one final timd.
This is the main function of consolidation.py, and is called by
calculate_timd.py.
temp_timds is a dictionary with scout names as keys and their
respective tempTIMD as a value.
"""
# 'sprking' is the scout with the best (lowest) SPR
#TODO: Implement spr system
sprking = list(temp_timds.keys())[0]
final_timd = {}
# Iterates through the keys of the best scout's tempTIMD and
# consolidates each data_field one at a time.
for data_field in list(temp_timds[sprking]):
if data_field == 'timeline':
# In order to compute the timeline properly, it is split
# into a list of the timelines.
timelines = {}
for scout, temp_timd in temp_timds.items():
temp_timeline = temp_timd.get('timeline', [])
timelines[scout] = temp_timeline
# If the list of timelines only includes one timeline, that
# timeline is taken as the correct one and put into the
# final TIMD.
if len(timelines.values()) == 1:
# Converts all times to floats and removes asterisk to
# put it into the format of a timd.
final_timeline = []
for action in timelines[sprking]:
action_time = action.get('time')
# Takes the time before the asterisk, if there is no
# asterisk, .split() still returns a list, a list of
# only the time, meaning [0] works in both
# instances.
action['time'] = float(action_time.split('*')[0])
final_timd['timeline'] = timelines[sprking]
# If the list has more than one tempTIMD, the process for
# computation has to be split up by each of the types of
# actions in the timeline.
else:
# Creates the final timeline which is passed as the
# timeline for the final timd at the end of
# consolidation.
final_timeline = []
# Separates all the basic actions out and consolidates
# them one at a time. All the actions are consolidated
# separately so that the timings on each action are
# split apart, making it more easy to line up, identify,
# and consolidate the timeline.
for action_type in ['pinningFoul', 'incap', 'unincap', \
'drop', 'startDefense', 'endDefense', \
'placement', 'intake']:
final_timeline += consolidate_timeline_action(
timelines, action_type, sprking)
# Also consolidates climb separately in order to
# separate it from intakes and placements. Climb needs a
# separate function because of its relatively strange
# structure.
climb = climb_consolidation(timelines, sprking)
if climb is not None:
final_timeline.append(climb)
# Deletes any blank actions.
final_timeline = [action for action in final_timeline if
action != {}]
# Once the timeline is finally completed, it is sorted
# by time, and added to the final timd.
final_timd['timeline'] = sorted(final_timeline, \
key=lambda action: action['time'], reverse=True)
# When consolidating non-timed keys, it is easy to consolidate
# them, as you can simply find which value is the most common in
# the set of three possibilities. The other data_fields that
# are not included in this set, such as timerStarted, are scout
# diagnostics, and not included in the final TIMD.
elif data_field not in ['timeline', 'timerStarted',
'currentCycle', 'scoutID', 'scoutName',
'appVersion', 'assignmentMode',
'assignmentFileTimestamp',
'matchesNotScouted']:
# Creates a dictionary of each scout to the key from their
# tempTIMD to compare against each other. (Code note - This
# code is using .get and not simply referencing the key out
# of the dictionary because .get doesn't error out when the
# key doesn't exist. Instead, .get returns NoneType).
data_field_comparison_list = {}
for scout, temp_timd in temp_timds.items():
temp_data_field = temp_timd.get(data_field)
if temp_data_field is not None:
data_field_comparison_list[scout] = temp_data_field
# Uses the max_occurrences function to find the correct value
# for the data field.
data_occurence_max = max_occurrences(
data_field_comparison_list, sprking)
final_timd[data_field] = data_occurence_max
return final_timd