def od_merge_mergeable(child_trip, father_trip):
'''check if this pair of pairs is od-merge mergable and admissible-od-merge mergable'''
if child_trip['start_point']['time'] >= father_trip['start_point']['time']:
return False, False, -1
t_walkLeg1 = geo.distBetween(
child_trip['start_point']['lat'], child_trip['start_point']['lng'],
father_trip['start_point']['lat'],
father_trip['start_point']['lng']) / Constants.PACE
if child_trip['start_point']['time'] + t_walkLeg1 > father_trip[
'start_point']['time']:
return False, False, -1
t_walkLeg2 = geo.distBetween(
child_trip['end_point']['lat'], child_trip['end_point']['lng'],
father_trip['end_point']['lat'],
father_trip['end_point']['lng']) / Constants.PACE
delay = father_trip['end_point']['time'] + t_walkLeg2 - child_trip[
'end_point']['time'] - child_trip['tt']
'''
if delay<0:
print delay, 'f=(%d~%d), c=(%d~%d)'%(father_trip['start_point']['time'], father_trip['end_point']['time'], child_trip['start_point']['time'],child_trip['end_point']['time'])
delay+=864000
'''
#if father_trip['end_point']['time']- child_trip['start_point']['time']+ t_walkLeg2 / PACE > child_trip['tt'] * (1 + DELTA):
if delay > Constants.DELTA:
return True, False, delay
#print father_trip['tt'], t_walkLeg1, t_walkLeg2, child_trip['start_point']['time'], father_trip['start_point']['time'], child_trip['tt'], father_trip['tt']+(t_walkLeg1+t_walkLeg2)/PACE+(father_trip['start_point']['time']-child_trip['start_point']['time']-t_walkLeg1/PACE), child_trip['tt']*(1+DELTA/100)
return True, True, delay
def is_valid_trip(trip_file_name, debug=False):
"""this function determines whether an extracted trip is valid based on some heuristics"""
#filename=Constants.PROCESSED_DIR+dirname+'/trip_trajectory/'+str(trip_id)+'.txt'
#gps_point=[]
trip_file=open(trip_file_name, 'r')
lines=trip_file.readlines()
count=0
for idx in range(len(lines)):
_, lon, lat, _, _, time=lines[idx].split(',')
lon=float(lon)
lat=float(lat)
time=int(time)
if count==5: #if the max distance from the start point does not increase in 5 consecutive steps, then disregard the trip
return False
if idx==0:
start_time=time
start_point=(lat, lon)
max_dist=0
#gps_point.append((0, lat, lon))
else:
dist=geo.distBetween(lat, lon, start_point[0], start_point[1])
if debug:
print dist
if dist<max_dist:
count+=1
else:
max_dist=dist
#gps_point.append((dist, lat, lon))
if time-start_time<30: #if travel time < half minute, disregard the trip
return False
return True
def is_valid_trip(trip_file_name, debug=False):
"""this function determines whether an extracted trip is valid based on some heuristics"""
#filename=Constants.PROCESSED_DIR+dirname+'/trip_trajectory/'+str(trip_id)+'.txt'
#gps_point=[]
trip_file = open(trip_file_name, 'r')
lines = trip_file.readlines()
count = 0
for idx in range(len(lines)):
_, lon, lat, _, _, time = lines[idx].split(',')
lon = float(lon)
lat = float(lat)
time = int(time)
if count == 5: #if the max distance from the start point does not increase in 5 consecutive steps, then disregard the trip
return False
if idx == 0:
start_time = time
start_point = (lat, lon)
max_dist = 0
#gps_point.append((0, lat, lon))
else:
dist = geo.distBetween(lat, lon, start_point[0], start_point[1])
if debug:
print dist
if dist < max_dist:
count += 1
else:
max_dist = dist
#gps_point.append((dist, lat, lon))
if time - start_time < 30: #if travel time < half minute, disregard the trip
return False
return True
def od_merge_mergeable(child_trip, father_trip):
'''check if this pair of pairs is od-merge mergable and admissible-od-merge mergable'''
if child_trip['start_point']['time'] >= father_trip['start_point']['time']:
return False, False, -1
t_walkLeg1 = geo.distBetween(child_trip['start_point']['lat'], child_trip['start_point']['lng'], father_trip['start_point']['lat'], father_trip['start_point']['lng']) / Constants.PACE
if child_trip['start_point']['time'] + t_walkLeg1 > father_trip['start_point']['time']:
return False, False, -1
t_walkLeg2 = geo.distBetween(child_trip['end_point']['lat'], child_trip['end_point']['lng'], father_trip['end_point']['lat'], father_trip['end_point']['lng']) / Constants.PACE
delay = father_trip['end_point']['time'] + t_walkLeg2 - child_trip['end_point']['time']-child_trip['tt']
'''
if delay<0:
print delay, 'f=(%d~%d), c=(%d~%d)'%(father_trip['start_point']['time'], father_trip['end_point']['time'], child_trip['start_point']['time'],child_trip['end_point']['time'])
delay+=864000
'''
#if father_trip['end_point']['time']- child_trip['start_point']['time']+ t_walkLeg2 / PACE > child_trip['tt'] * (1 + DELTA):
if delay > Constants.DELTA:
return True, False, delay
#print father_trip['tt'], t_walkLeg1, t_walkLeg2, child_trip['start_point']['time'], father_trip['start_point']['time'], child_trip['tt'], father_trip['tt']+(t_walkLeg1+t_walkLeg2)/PACE+(father_trip['start_point']['time']-child_trip['start_point']['time']-t_walkLeg1/PACE), child_trip['tt']*(1+DELTA/100)
return True, True, delay
def extract_trip_from_one_file(file_name, trip_dir, trip_meta):
#print file_name
f = open(file_name)
# Using a DictReader instead
r = csv.DictReader(f, ['taxi_id', 'timestamp', 'lat', 'lng', 'unknown1', 'unknown2', 'occupied'])
trips = []
max_lat = 0.0
min_lat = 180
max_lng = 0.0
min_lng = 180
lat = []
lng = []
timestamp = []
trip = []
travelDistance = 0.0
trip_meta_file = open(trip_meta, 'a+')
trip_id = extract_trip_from_one_file.count
for line in r:
if "{occupied}".format(**line) == '1':
day_of_week = int(time.strftime("%w", time.strptime("{timestamp}".format(**line), "%Y-%m-%d %H:%M:%S")))
hour_of_day = int(time.strftime("%H", time.strptime("{timestamp}".format(**line), "%Y-%m-%d %H:%M:%S")))
minute = int(time.strftime("%M", time.strptime("{timestamp}".format(**line), "%Y-%m-%d %H:%M:%S")))
sec = int(time.strftime("%S", time.strptime("{timestamp}".format(**line), "%Y-%m-%d %H:%M:%S")))
sec_of_day = hour_of_day * 3600 + minute * 60 + sec
trip.append(("{taxi_id}".format(**line), float("{lat}".format(**line)), float("{lng}".format(**line)), day_of_week, hour_of_day, sec_of_day))
lat.append(float("{lat}".format(**line)))
lng.append(float("{lng}".format(**line)))
timestamp.append(sec_of_day)
if len(lat) > 1:
travelDistance += geo.distBetween(lat[-1], lng[-1], lat[-2], lng[-2])
else:
if trip:
if timestamp[-1] >= timestamp[0]:
travelTime = timestamp[-1] - timestamp[0]
else: # a trip crossed a day
travelTime = timestamp[-1] - timestamp[0] + 60 * 60 * 24
if travelTime > 0 and travelDistance > 0: #filter trips with zero travel time/distance
trips.append((trip, lat, lng, timestamp, travelDistance, travelTime))
#generate a trip file
trip_file = open(trip_dir + str(trip_id) + '.txt', 'w')
for line in trip:
print >> trip_file, ','.join([str(field) for field in line])
trip_file.close()
#update trip meta
trip_meta_file.write(','.join([str(trip_id), str(trip[0][0]), str(lat[0]), str(lng[0]), str(timestamp[0]), str(lat[-1]), str(lng[-1]), str(timestamp[-1]), str(travelTime), str(travelDistance)])+'\n')
trip_id += 1
max_lat = max(max(lat), max_lat)
min_lat = min(min(lat), min_lat)
max_lng = max(max(lng), max_lng)
min_lng = min(min(lng), min_lng)
lat = []
lng = []
timestamp = []
trip = []
travelDistance = 0.0
trip_meta_file.close()
extract_trip_from_one_file.count = trip_id
return trips, max_lat, min_lat, max_lng, min_lng
def extract_trip_from_one_file(file_name, trip_dir, trip_meta):
#print file_name
f = open(file_name)
# Using a DictReader instead
r = csv.DictReader(f, [
'taxi_id', 'timestamp', 'lat', 'lng', 'unknown1', 'unknown2',
'occupied'
])
trips = []
max_lat = 0.0
min_lat = 180
max_lng = 0.0
min_lng = 180
lat = []
lng = []
timestamp = []
trip = []
travelDistance = 0.0
trip_meta_file = open(trip_meta, 'a+')
trip_id = extract_trip_from_one_file.count
for line in r:
if "{occupied}".format(**line) == '1':
day_of_week = int(
time.strftime(
"%w",
time.strptime("{timestamp}".format(**line),
"%Y-%m-%d %H:%M:%S")))
hour_of_day = int(
time.strftime(
"%H",
time.strptime("{timestamp}".format(**line),
"%Y-%m-%d %H:%M:%S")))
minute = int(
time.strftime(
"%M",
time.strptime("{timestamp}".format(**line),
"%Y-%m-%d %H:%M:%S")))
sec = int(
time.strftime(
"%S",
time.strptime("{timestamp}".format(**line),
"%Y-%m-%d %H:%M:%S")))
sec_of_day = hour_of_day * 3600 + minute * 60 + sec
trip.append(
("{taxi_id}".format(**line), float("{lat}".format(**line)),
float("{lng}".format(**line)), day_of_week, hour_of_day,
sec_of_day))
lat.append(float("{lat}".format(**line)))
lng.append(float("{lng}".format(**line)))
timestamp.append(sec_of_day)
if len(lat) > 1:
travelDistance += geo.distBetween(lat[-1], lng[-1], lat[-2],
lng[-2])
else:
if trip:
if timestamp[-1] >= timestamp[0]:
travelTime = timestamp[-1] - timestamp[0]
else: # a trip crossed a day
travelTime = timestamp[-1] - timestamp[0] + 60 * 60 * 24
if travelTime > 0 and travelDistance > 0: #filter trips with zero travel time/distance
trips.append((trip, lat, lng, timestamp, travelDistance,
travelTime))
#generate a trip file
trip_file = open(trip_dir + str(trip_id) + '.txt', 'w')
for line in trip:
print >> trip_file, ','.join(
[str(field) for field in line])
trip_file.close()
#update trip meta
trip_meta_file.write(','.join([
str(trip_id),
str(trip[0][0]),
str(lat[0]),
str(lng[0]),
str(timestamp[0]),
str(lat[-1]),
str(lng[-1]),
str(timestamp[-1]),
str(travelTime),
str(travelDistance)
]) + '\n')
trip_id += 1
max_lat = max(max(lat), max_lat)
min_lat = min(min(lat), min_lat)
max_lng = max(max(lng), max_lng)
min_lng = min(min(lng), min_lng)
lat = []
lng = []
timestamp = []
trip = []
travelDistance = 0.0
trip_meta_file.close()
extract_trip_from_one_file.count = trip_id
return trips, max_lat, min_lat, max_lng, min_lng
