-
Notifications
You must be signed in to change notification settings - Fork 1
/
FAA_DataPointLoader_no_image_1.py
356 lines (315 loc) · 24 KB
/
FAA_DataPointLoader_no_image_1.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
from pprint import pprint
import pickle
import time
from datetime import datetime, timedelta
import json
import math
import random
import numpy as np
import pandas as pd
import sys
sys.path.append('./')
from ProgressIndicator import ProgressIndicator as API
import warnings
warnings.filterwarnings('ignore')
class FAA_DataPointLoader_no_image:
# choose between DROP or ONEHOT
# time, 'UNIX' or 'COMPLETE'
# complete: also,month,day,day_of_week
def __init__(self,fn_dep,fn_arr,_fn_wthr,
observation_window_duration,prediction_horizon,
c_weatherCondition = 'DROP', # 'ONEHOT'
c_tailNumber = 'DROP', # 'ONEHOT'
c_flightNumber = 'DROP', # ONEHOT
c_airport = 'DROP', # ONEHOT
c_time = 'UNIX', # COMPLETE (no year)
c_carrier_code = 'DROP', # ONEHOT
c_icao = 'DROP', # (always drop)
c_callsign = 'DROP', # (always drop)
normalization = 'MINMAX01', # NONE, STANDARD, (WHITE)
verbose = False):
time_start = time.time()
self.fn_dep = fn_dep
self.fn_arr = fn_arr
self._fn_wthr = _fn_wthr
self.observation_window_duration = observation_window_duration
self.prediction_horizon = prediction_horizon
self.c_weatherCondition = c_weatherCondition
self.c_tailNumber = c_tailNumber
self.c_flightNumber = c_flightNumber
self.c_airport = c_airport
self.c_time = c_time
self.c_carrier_code = c_carrier_code
self.c_icao = c_icao
self.c_callsign = c_callsign
self.normalization = normalization
self.verbose = verbose
weather = pd.read_csv(_fn_wthr) # data cleaning here
arrival = pd.read_csv(fn_arr)
depart = pd.read_csv(fn_dep)
weather['Wind_is_CALM'] = weather.Wind
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("CALM", "1")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("WSW", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("VAR", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("SSE", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("NNW", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("SSW", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("ESE", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("WNW", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("ENE", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("NNE", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("SW", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("SE", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("NW", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("NE", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("W", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("E", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("S", "0")
weather['Wind_is_CALM'] = weather.Wind_is_CALM.str.replace("N", "0")
weather['Wind_is_CALM'] = weather['Wind_is_CALM'].astype(float)
weather['Wind_is_VAR'] = weather.Wind
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("VAR", "1")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("WSW", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("CALM", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("SSE", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("NNW", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("SSW", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("ESE", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("WNW", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("ENE", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("NNE", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("SW", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("SE", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("NW", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("NE", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("W", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("E", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("S", "0")
weather['Wind_is_VAR'] = weather.Wind_is_VAR.str.replace("N", "0")
weather['Wind_is_VAR'] = weather['Wind_is_VAR'].astype(float)
# convert "Wind" from direction to radian
weather['Wind'] = weather.Wind.str.replace("WSW", str(247.5*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("CALM", str(0))
weather['Wind'] = weather.Wind.str.replace("VAR", str(0))
weather['Wind'] = weather.Wind.str.replace("SSE", str(157.5*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("NNW", str(337.5*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("SSW", str(202.5*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("ESE", str(112.5*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("WNW", str(292.5*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("ENE", str(67.5*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("NNE", str(22.5*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("SW", str(225*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("SE", str(135*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("NW", str(315*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("NE", str(45*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("W", str(270*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("E", str(90*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("S", str(180*math.pi/180))
weather['Wind'] = weather.Wind.str.replace("N", str(0*math.pi/180))
weather['Wind'] = weather['Wind'].astype(float)
weather.DateTime = pd.to_datetime(weather.DateTime)
if self.c_time == 'COMPLETE':
weather['year'], weather['month'], weather['day_of_month'] = weather['DateTime'].dt.year, weather['DateTime'].dt.month, weather['DateTime'].dt.day
#weather['year'] = (weather.year - weather.year.min()) / (weather.year.max() - weather.year.min())
weather['Date'] = weather['month'].apply(str) + '/' + weather['day_of_month'].apply(str) + '/' + weather['year'].apply(str)
#weather['Date'] = (weather.Date - weather.Date.min()) / (weather.Date.max() - weather.Date.min())
weather['day_of_year'] = weather['DateTime'].dt.dayofyear
#weather['day_of_year'] = (weather.day_of_year - weather.day_of_year.min()) / (weather.day_of_year.max() - weather.day_of_year.min())
weather['day_of_week'] = weather['DateTime'].dt.dayofweek
#weather['day_of_week'] = (weather.day_of_week - weather.day_of_week.min()) / (weather.day_of_week.max() - weather.day_of_week.min())
weather['week_of_year'] = weather['DateTime'].dt.weekofyear
#weather['week_of_year'] = (weather.week_of_year - weather.week_of_year.min()) / (weather.week_of_year.max() - weather.week_of_year.min())
weather['hour_in_day'] = weather['DateTime'].dt.hour
#weather['hour_in_day'] = (weather.hour_in_day - weather.hour_in_day.min()) / (weather.hour_in_day.max() - weather.hour_in_day.min())
weather['minute_in_hour'] = weather['DateTime'].dt.minute
#weather['minute_in_hour'] = (weather.minute_in_hour - weather.minute_in_hour.min()) / (weather.minute_in_hour.max() - weather.minute_in_hour.min())
weather['time_in_unix'] = weather['DateTime'].apply(lambda x: datetime.timestamp(x))
elif self.c_time == 'UNIX':
weather['time_in_unix'] = weather['DateTime'].apply(lambda x: datetime.timestamp(x))
if self.c_weatherCondition == 'ONEHOT':
weather = weather.join(pd.get_dummies(weather.Condition))
elif self.c_weatherCondition == 'DROP':
# make it drop
pass
weather = weather.drop(['Date','Unnamed: 0','Condition','Time'], axis=1)
weather['Temperature'] = (weather.Temperature - weather.Temperature.min()) / (weather.Temperature.max() - weather.Temperature.min())
weather['DewPoint'] = (weather.DewPoint - weather.DewPoint.min()) / (weather.DewPoint.max() - weather.DewPoint.min())
weather['Wind'] = (weather.Wind - weather.Wind.min()) / (weather.Wind.max() - weather.Wind.min())
weather['WindSpeed'] = (weather.WindSpeed - weather.WindSpeed.min()) / (weather.WindSpeed.max() - weather.WindSpeed.min())
weather['WindGust'] = (weather.WindGust - weather.WindGust.min()) / (weather.WindGust.max() - weather.WindGust.min())
weather['Pressure'] = (weather.Pressure - weather.Pressure.min()) / (weather.Pressure.max() - weather.Pressure.min())
weather['temp_unix'] = (weather.time_in_unix - weather.time_in_unix.min()) / (weather.time_in_unix.max() - weather.time_in_unix.min())
self.weather = weather
if verbose: print('weather DONE, START arrival',time.time()-time_start)
# rename columns
arrival.rename(columns={'Date (MM/DD/YYYY)':'Date', 'Tail Number':'Tail_Number', 'Scheduled Arrival Time':'Scheduled_Arrival_Time',
'Carrier Code':'Carrier_Code', 'Origin Airport':'Origin_Airport', 'Actual Arrival Time':'Actual_Arrival_Time',
'Wheels-on Time':'Wheels_on_Time', 'Scheduled Elapsed Time (Minutes)':'Scheduled_Elapsed_Time_Minutes',
'Actual Elapsed Time (Minutes)':'Actual_Elapsed_Time_Minutes', 'Arrival Delay (Minutes)':'Arrival_Delay_Minutes',
'Taxi-In time (Minutes)':'Taxi_In_time_Minutes', 'Delay Carrier (Minutes)':'Delay_Carrier_Minutes',
'Delay Weather (Minutes)':'Delay_Weather_Minutes',
'Delay National Aviation System (Minutes)':'Delay_National_Aviation_System_Minutes',
'Delay Security (Minutes)':'Delay_Security_Minutes',
'Delay Late Aircraft Arrival (Minutes)':'Delay_Late_Aircraft_Arrival_Minutes'}, inplace = True)
arrival['DateTime'] = arrival['Date'] + ' ' + arrival['Scheduled_Arrival_Time']
arrival.DateTime = pd.to_datetime(arrival.DateTime)
if self.c_time == 'COMPLETE':
arrival['year'], arrival['month'] = arrival['DateTime'].dt.year, arrival['DateTime'].dt.month
arrival['day_of_month'] = arrival['DateTime'].dt.day
arrival['Date'] = arrival['month'].apply(str) + '/' + arrival['day_of_month'].apply(str) + '/' + arrival['year'].apply(str)
arrival['day_of_year'], arrival['day_of_week'] = arrival['DateTime'].dt.dayofyear, arrival['DateTime'].dt.dayofweek
arrival['week_of_year'] = arrival['DateTime'].dt.weekofyear
arrival['hour_in_day_SAT'], arrival['minute_in_hour_SAT'] = arrival['DateTime'].dt.hour, arrival['DateTime'].dt.minute
arrival['time_in_unix_SAT'] = arrival['DateTime'].apply(lambda x: datetime.timestamp(x))
arrival['Actual_Arrival_Time'] = arrival.Actual_Arrival_Time.str.replace("24:00", "23:59")
arrival['DateTime_AAT'] = arrival['Date'] + ' ' + arrival['Actual_Arrival_Time']
arrival.DateTime_AAT = pd.to_datetime(arrival.DateTime_AAT)
arrival['hour_in_day_AAT'], arrival['minute_in_hour_AAT'] = arrival['DateTime_AAT'].dt.hour, arrival['DateTime_AAT'].dt.minute
arrival['time_in_unix_AAT'] = arrival['DateTime_AAT'].apply(lambda x: datetime.timestamp(x))
arrival['Wheels_on_Time'] = arrival.Wheels_on_Time.str.replace("24:00", "23:59")
arrival['DateTime_WOnT'] = arrival['Date'] + ' ' + arrival['Wheels_on_Time']
arrival.DateTime_WOnT = pd.to_datetime(arrival.DateTime_WOnT)
arrival['hour_in_day_WOnT'], arrival['minute_in_hour_WOnT'] = arrival['DateTime_WOnT'].dt.hour, arrival['DateTime_WOnT'].dt.minute
arrival['time_in_unix_WOnT'] = arrival['DateTime_WOnT'].apply(lambda x: datetime.timestamp(x))
elif self.c_time == 'UNIX':
arrival['time_in_unix_SAT'] = arrival['DateTime'].apply(lambda x: datetime.timestamp(x))
#arrival['time_in_unix_SAT'] = pd.to_datetime(arrival.DateTime).values.astype(float)//10**9
arrival['Actual_Arrival_Time'] = arrival.Actual_Arrival_Time.str.replace("24:00", "23:59")
arrival['DateTime_AAT'] = arrival['Date'] + ' ' + arrival['Actual_Arrival_Time']
arrival.DateTime_AAT = pd.to_datetime(arrival.DateTime_AAT)
arrival['time_in_unix_AAT'] = arrival['DateTime_AAT'].apply(lambda x: datetime.timestamp(x))
#arrival['time_in_unix_AAT'] = pd.to_datetime(arrival.DateTime_AAT).values.astype(float)//10**9
arrival['Wheels_on_Time'] = arrival.Wheels_on_Time.str.replace("24:00", "23:59")
arrival['DateTime_WOnT'] = arrival['Date'] + ' ' + arrival['Wheels_on_Time']
arrival.DateTime_WOnT = pd.to_datetime(arrival.DateTime_WOnT)
arrival['time_in_unix_WOnT'] = arrival['DateTime_WOnT'].apply(lambda x: datetime.timestamp(x))
#arrival['time_in_unix_WOnT'] = pd.to_datetime(arrival.DateTime_WOnT).values.astype(float)//10**9
if self.c_carrier_code == 'ONEHOT':
arrival = arrival.join(pd.get_dummies(arrival.Carrier_Code))
elif self.c_carrier_code == 'DROP':
# make it drop
pass
if self.c_airport == 'ONEHOT':
arrival = arrival.join(pd.get_dummies(arrival.Origin_Airport))
elif self.c_airport == 'DROP':
# make it drop
pass
if self.c_tailNumber == 'ONEHOT':
arrival = arrival.join(pd.get_dummies(arrival.Tail_Number))
elif self.c_tailNumber == 'DROP':
# make it drop
pass
arrival = arrival.drop(['Flight Number','Date','Scheduled_Arrival_Time','Actual_Arrival_Time','Wheels_on_Time','DateTime_AAT','DateTime_WOnT',
'Carrier_Code', 'Origin_Airport','ICAO','Callsign'], axis=1)
arrival['Scheduled_Elapsed_Time_Minutes'] = (arrival.Scheduled_Elapsed_Time_Minutes - arrival.Scheduled_Elapsed_Time_Minutes.min()) / (arrival.Scheduled_Elapsed_Time_Minutes.max() - arrival.Scheduled_Elapsed_Time_Minutes.min())
arrival['Actual_Elapsed_Time_Minutes'] = (arrival.Actual_Elapsed_Time_Minutes - arrival.Actual_Elapsed_Time_Minutes.min()) / (arrival.Actual_Elapsed_Time_Minutes.max() - arrival.Actual_Elapsed_Time_Minutes.min())
arrival['Arrival_Delay_Minutes'] = (arrival.Arrival_Delay_Minutes - arrival.Arrival_Delay_Minutes.min()) / (arrival.Arrival_Delay_Minutes.max() - arrival.Arrival_Delay_Minutes.min())
arrival['Taxi_In_time_Minutes'] = (arrival.Taxi_In_time_Minutes - arrival.Taxi_In_time_Minutes.min()) / (arrival.Taxi_In_time_Minutes.max() - arrival.Taxi_In_time_Minutes.min())
arrival['Delay_Carrier_Minutes'] = (arrival.Delay_Carrier_Minutes - arrival.Delay_Carrier_Minutes.min()) / (arrival.Delay_Carrier_Minutes.max() - arrival.Delay_Carrier_Minutes.min())
arrival['Delay_Weather_Minutes'] = (arrival.Delay_Weather_Minutes - arrival.Delay_Weather_Minutes.min()) / (arrival.Delay_Weather_Minutes.max() - arrival.Delay_Weather_Minutes.min())
arrival['Delay_National_Aviation_System_Minutes'] = (arrival.Delay_National_Aviation_System_Minutes - arrival.Delay_National_Aviation_System_Minutes.min()) / (arrival.Delay_National_Aviation_System_Minutes.max() - arrival.Delay_National_Aviation_System_Minutes.min())
arrival['Delay_Security_Minutes'] = (arrival.Delay_Security_Minutes - arrival.Delay_Security_Minutes.min()) / (arrival.Delay_Security_Minutes.max() - arrival.Delay_Security_Minutes.min())
arrival['Delay_Late_Aircraft_Arrival_Minutes'] = (arrival.Delay_Late_Aircraft_Arrival_Minutes - arrival.Delay_Late_Aircraft_Arrival_Minutes.min()) / (arrival.Delay_Late_Aircraft_Arrival_Minutes.max() - arrival.Delay_Late_Aircraft_Arrival_Minutes.min())
arrival['time_in_unix_SAT'] = (arrival.time_in_unix_SAT - arrival.time_in_unix_SAT.min()) / (arrival.time_in_unix_SAT.max() - arrival.time_in_unix_SAT.min())
arrival['time_in_unix_WOnT'] = (arrival.time_in_unix_WOnT - arrival.time_in_unix_WOnT.min()) / (arrival.time_in_unix_WOnT.max() - arrival.time_in_unix_WOnT.min())
arrival['temp_AAT'] = (arrival.time_in_unix_AAT - arrival.time_in_unix_AAT.min()) / (arrival.time_in_unix_AAT.max() - arrival.time_in_unix_AAT.min())
self.arrival = arrival
if verbose: print('arrival DONE, START depart',time.time()-time_start)
# rename "Date (MM/DD/YYYY)" & "Tail Number" columns
depart.rename(columns={'Date (MM/DD/YYYY)':'Date', 'Tail Number':'Tail_Number', 'Scheduled departure time':'Scheduled_departure_time',
'Actual departure time':'Actual_departure_time', 'Wheels-off time':'Wheels_off_time', 'Carrier Code':'Carrier_Code',
'Destination Airport':'Destination_Airport', 'Scheduled elapsed time (Minutes)':'Scheduled_elapsed_time_Minutes',
'Departure delay (Minutes)':'Departure_delay_Minutes', 'Delay Carrier (Minutes)':'Delay_Carrier_Minutes',
'Delay Weather (Minutes)':'Delay_Weather_Minutes', 'Delay National Aviation System (Minutes)':'Delay_National_Aviation_System_Minutes',
'Delay Security (Minutes)':'Delay_Security_Minutes', 'Delay Late Aircraft Arrival (Minutes)':'Delay_Late_Aircraft_Arrival_Minutes'}, inplace = True)
# correct day and month and create a new column 'DateTime'
depart['DateTime'] = depart['Date'] + ' ' + depart['Scheduled_departure_time']
depart.DateTime = pd.to_datetime(depart.DateTime)
depart['year'], depart['month'], depart['day_of_month'] = depart['DateTime'].dt.year, depart['DateTime'].dt.day, depart['DateTime'].dt.month
depart['Date'] = depart['month'].apply(str) + '/' + depart['day_of_month'].apply(str) + '/' + depart['year'].apply(str)
depart['DateTime'] = depart['Date'] + ' ' + depart['Scheduled_departure_time']
depart.DateTime = pd.to_datetime(depart.DateTime)
if self.c_time == 'COMPLETE':
depart['day_of_year'], depart['day_of_week'] = depart['DateTime'].dt.dayofyear, depart['DateTime'].dt.dayofweek
depart['week_of_year'] = depart['DateTime'].dt.weekofyear
depart['hour_in_day_SDT'], depart['minute_in_hour_SDT'] = depart['DateTime'].dt.hour, depart['DateTime'].dt.minute
depart['time_in_unix_SDT'] = depart['DateTime'].apply(lambda x: datetime.timestamp(x))
elif self.c_time == 'UNIX':
depart['time_in_unix_SDT'] = depart['DateTime'].apply(lambda x: datetime.timestamp(x))
depart = depart.drop(['year','month','day_of_month'], axis=1)
if self.c_carrier_code == 'ONEHOT':
depart = depart.join(pd.get_dummies(depart.Carrier_Code))
elif self.c_carrier_code == 'DROP':
pass
if self.c_airport == 'ONEHOT':
depart = depart.join(pd.get_dummies(depart.Destination_Airport))
elif self.c_airport == 'DROP':
# make it drop
pass
if self.c_tailNumber == 'ONEHOT':
depart = depart.join(pd.get_dummies(depart.Tail_Number))
elif self.c_tailNumber == 'DROP':
# make it drop
pass
depart = depart.drop(['Flight Number','Date','Scheduled_departure_time','Actual_departure_time','Wheels_off_time', 'Carrier_Code',
'Destination_Airport','ICAO','Callsign','Actual elapsed time (Minutes)','Taxi-Out time (Minutes)'], axis=1)
depart['Scheduled_elapsed_time_Minutes'] = (depart.Scheduled_elapsed_time_Minutes - depart.Scheduled_elapsed_time_Minutes.min()) / (depart.Scheduled_elapsed_time_Minutes.max() - depart.Scheduled_elapsed_time_Minutes.min())
depart['Departure_delay_Minutes'] = (depart.Departure_delay_Minutes - depart.Departure_delay_Minutes.min()) / (depart.Departure_delay_Minutes.max() - depart.Departure_delay_Minutes.min())
depart['Delay_Carrier_Minutes'] = (depart.Delay_Carrier_Minutes - depart.Delay_Carrier_Minutes.min()) / (depart.Delay_Carrier_Minutes.max() - depart.Delay_Carrier_Minutes.min())
depart['Delay_Weather_Minutes'] = (depart.Delay_Weather_Minutes - depart.Delay_Weather_Minutes.min()) / (depart.Delay_Weather_Minutes.max() - depart.Delay_Weather_Minutes.min())
depart['Delay_National_Aviation_System_Minutes'] = (depart.Delay_National_Aviation_System_Minutes - depart.Delay_National_Aviation_System_Minutes.min()) / (depart.Delay_National_Aviation_System_Minutes.max() - depart.Delay_National_Aviation_System_Minutes.min())
depart['Delay_Security_Minutes'] = (depart.Delay_Security_Minutes - depart.Delay_Security_Minutes.min()) / (depart.Delay_Security_Minutes.max() - depart.Delay_Security_Minutes.min())
depart['Delay_Late_Aircraft_Arrival_Minutes'] = (depart.Delay_Late_Aircraft_Arrival_Minutes - depart.Delay_Late_Aircraft_Arrival_Minutes.min()) / (depart.Delay_Late_Aircraft_Arrival_Minutes.max() - depart.Delay_Late_Aircraft_Arrival_Minutes.min())
depart['temp_SDT'] = (depart.time_in_unix_SDT - depart.time_in_unix_SDT.min()) / (depart.time_in_unix_SDT.max() - depart.time_in_unix_SDT.min())
self.depart = depart
if verbose: print('depart DONE, continue...',time.time()-time_start)
if (self.c_time == 'UNIX' and self.c_carrier_code == 'DROP' and self.c_airport == 'DROP' and self.c_tailNumber == 'DROP' and self.c_weatherCondition == 'DROP'):
self.x_col = 24
self.y_col = 6
else:
raise NotImplementedError(self.c_time, self.c_carrier_code, self.c_airport, self.c_tailNumber, self.c_weatherCondition)
tlen = len(depart)
self.X = np.empty((tlen,self.x_col))
self.Y = np.empty((tlen,self.y_col))
if verbose:
api = API(tlen)
for i in range(tlen):
self.X[i,:],self.Y[i,:] = self.getItem_in_np(i)
if verbose: api.toc('matching...')
return
def __len__(self): # index available value
length = len(self.depart)
return length
def __getitem__(self,key):
return self.X[key,:],self.Y[key,:]
def getItem_in_df(self,indexNum):
df_y = self.depart.query('index == @indexNum')
tailNumb = df_y["Tail_Number"].values[0]
matchTime = df_y["time_in_unix_SDT"].values[0]
timePoint = matchTime - self.prediction_horizon
label_y = df_y.loc[:,["Departure_delay_Minutes", "Delay_Carrier_Minutes", "Delay_Weather_Minutes",
"Delay_National_Aviation_System_Minutes", "Delay_Security_Minutes", "Delay_Late_Aircraft_Arrival_Minutes"]]
x_dep_temp = df_y.drop(["Departure_delay_Minutes", "Delay_Carrier_Minutes", "Delay_Weather_Minutes",
"Delay_National_Aviation_System_Minutes", "Delay_Security_Minutes", "Delay_Late_Aircraft_Arrival_Minutes",
"Tail_Number","DateTime"], axis=1)
df_xarr = self.arrival.query('Tail_Number == @tailNumb & time_in_unix_AAT <= @matchTime').tail(1)
x_arr_temp = df_xarr.drop(["Tail_Number","DateTime"], axis=1)
# unix_time_timepoint - 7200 < unix_time_weather < unix_time_timepoint
df_xwea = self.weather.query('time_in_unix <= @timePoint & time_in_unix >= @timePoint - 7200').tail(1)
x_wea_temp = df_xwea.drop('DateTime', axis=1)
x_dep_temp['time_in_unix_SDT'] = x_dep_temp['temp_SDT']
x_arr_temp['time_in_unix_AAT'] = x_arr_temp['temp_AAT']
x_wea_temp['time_in_unix'] = x_wea_temp['temp_unix']
x_dep_temp = x_dep_temp.drop('temp_SDT', axis=1)
x_arr_temp = x_arr_temp.drop('temp_AAT', axis=1)
x_wea_temp = x_wea_temp.drop('temp_unix', axis=1)
label_x = pd.concat([x_dep_temp.reset_index(drop=True), x_arr_temp.reset_index(drop=True), x_wea_temp.reset_index(drop=True)], axis=1)
label_x = label_x.fillna(0)
return label_x, label_y
def getItem_in_np(self, indexNum):
x,y = self.getItem_in_df(indexNum)
return np.array(x).flatten(), np.array(y).flatten()
def getFeatureShape(self):
return len(self),self.x_col
def getLabelShape(self):
return len(self),self.y_col