def cal_dist(name,GROUND_TRUTH): distance=[150.453623567463] df=pd.read_csv(name) l=len(df) i=0 while i+1<l: lat1=df.iloc[i]['latitude'] long1=df.iloc[i][' longitude'] lat2=df.iloc[i+1]['latitude'] long2=df.iloc[i+1][' longitude'] dist=get_spherical_distance(lat1,lat2,long1,long2) distance.append(dist) i+=1 df['cons_dist']=distance df.to_csv(name,index=False) insert_time(name) #print(GROUND_TRUTH) dec_zone(name,GROUND_TRUTH)
def compare_ground_truth(ground_truth_file, bus_stop_file,
GROUND_TRUTH_THRESHOLD):
gt = []
bus_stops = []
text = open(ground_truth_file, "r")
gt_1 = text.readlines()[1:]
for gt_point in gt_1:
gt_point = gt_point.replace('\n', '')
k = gt_point.split(',')
#print(k)
#print('\n')
for i in k:
gt.append(float(i))
#print(gt)
#print('\n')
text.close()
text = open(bus_stop_file, "r")
bus_stops_1 = text.readlines()[1:]
for gt_point in bus_stops_1:
gt_point = gt_point.replace('\n', '')
k = gt_point.split(',')
#print(k)
#print('\n')
cnt = 0
for i in k:
if cnt != 2:
bus_stops.append(float(i))
cnt += 1
#bus_stops.append(k)
#print(bus_stops)
text.close()
#bus_stops = read_file()
df = pd.read_csv(bus_stop_file)
#print "GT: ",len(gt)
#print "all bus_stops: ",len(bus_stops)
#print gt[0]
#print bus_stops[0]
#detected = []
#false_negative=[]
#gt_dict={}
used = []
stop = []
l_gt = len(gt)
l_bs = len(bus_stops)
cnt_gt = 0
#for gt_point in gt:
while cnt_gt < l_gt:
#distances = []
#min_point = None
#min_dist = 100000000000000
cnt_bs = 0
#for bs_point in bus_stops:
while cnt_bs < l_bs:
#pas
#print("p")
if bus_stops[cnt_bs] in used and bus_stops[cnt_bs + 1] in used:
#print("l")
cnt_bs += 2
continue
lats1, longs1 = gt[cnt_gt], gt[cnt_gt + 1]
lat1, long1 = float(lats1), float(longs1)
lats2, longs2 = bus_stops[cnt_bs], bus_stops[cnt_bs + 1]
lat2, long2 = float(lats2), float(longs2)
#print(lats1,",",longs1,",",lats2,",",longs2)
#print(lat1,",",long1,",",lat2,",",long2)
dist = get_spherical_distance(lat1, lat2, long1, long2)
#print(dist)
#the stoppage at the closest distance goes at the end of the list
if dist < GROUND_TRUTH_THRESHOLD:
#min_dist = dist
#min_point = bs_point
stop.append(cnt_bs / 2)
used.append(bus_stops[cnt_bs])
used.append(bus_stops[cnt_bs + 1])
#print(used)
#else:
# stop.append("0")
cnt_bs += 2
#print("b")
#print(l_gt)
#print(l_bs)
#print(cnt_bs)
cnt_gt += 2
#print(cnt_gt)
#print("a")
#print("s")
#print(len(stop))
#print(stop)
stop_ac = []
k = 0
while k < l_bs / 2:
stop_ac.append(0)
#print(len(stop_ac))
k += 1
j = 0
while j < l_bs / 2:
#pass
if j in stop:
stop_ac[j] = 1
j += 1
#print(stop_ac)
df['bs_predict'] = stop_ac
#print(df)
df.to_csv(bus_stop_file, index=False)
#gt_dict[gt_point[0]]= points
'''if min_dist < GROUND_TRUTH_THRESHOLD:
def dec_zone(name,GROUND_TRUTH):
#print(GROUND_TRUTH)
zone=[]
zone_class=[]
l_market=[]
l_nc=[]
l_highway=[]
df=pd.read_csv(name)
if '54feet' in GROUND_TRUTH:
Lat_market=23.5631833333
Long_market=87.28356
Lat_nc=23.5440616667
Long_nc=87.2887366667
Lat_high=23.4946766667
Long_high=87.3168283333
l_market.append(get_spherical_distance(Lat_market,Lat_nc,Long_market,Long_nc))
l_nc.append(get_spherical_distance(Lat_nc,Lat_high,Long_nc,Long_high))
l=len(df)
#print(l)
i=0
while i<l:
#print("a ",i)
Lat_a=df.iloc[i]['latitude']
Long_a=df.iloc[i][' longitude']
#if get_spherical_distance(Lat_a,Lat_market,Long_a,Long_market)+get_spherical_distance(Lat_a,Lat_nc,Long_a,Long_nc)==l_market:
if get_spherical_distance(Lat_a,Lat_market,Long_a,Long_market)<l_market[0]:
zone.append('Market')
zone_class.append('1')
#elif get_spherical_distance(Lat_a,Lat_nc,Long_a,Long_nc)+get_spherical_distance(Lat_a,Lat_high,Long_a,Long_high)==l_nc:
elif get_spherical_distance(Lat_a,Lat_nc,Long_a,Long_nc)<l_nc[0]:
zone.append('Normal city')
zone_class.append('2')
else:
zone.append('Highway')
zone_class.append('3')
#print(len(zone))
#print(zone)
i+=1
if 'azone' in GROUND_TRUTH:
Lat_market=[23.5481833333]
Long_market=[87.3024633333]
Lat_nc=[23.564295,23.5074966667]
Long_nc=[87.28288,87.3101183333]
Lat_high=[23.5702466667,23.5355883333]
Long_high=[87.3025383333,87.2982133333]
Lat=23.494395
Long=87.317125
l_market.append(get_spherical_distance(Lat_market[0],Lat_high[1],Long_market[0],Long_high[1]))
#l_market.append(get_spherical_distance(Lat_market[1],Lat_high[0],Long_market[1],Long_high[0]))
l_nc.append(get_spherical_distance(Lat_nc[0],Lat_high[0],Long_nc[0],Long_high[0]))
l_nc.append(get_spherical_distance(Lat_nc[1],Lat,Long_nc[1],Long))
l_highway.append(get_spherical_distance(Lat_high[0],Lat_market[0],Long_high[0],Long_market[0]))
l_highway.append(get_spherical_distance(Lat_high[1],Lat_nc[1],Long_high[1],Long_nc[1]))
#df=pd.read_csv(name)
l=len(df)
#print(l)
i=0
while i<l:
#print("a ",i)
Lat_a=df.iloc[i]['latitude']
Long_a=df.iloc[i][' longitude']
#if get_spherical_distance(Lat_a,Lat_market,Long_a,Long_market)+get_spherical_distance(Lat_a,Lat_nc,Long_a,Long_nc)==l_market:
if get_spherical_distance(Lat_a,Lat_market[0],Long_a,Long_market[0])<l_market[0]:
zone.append('Market')
zone_class.append('1')
#elif get_spherical_distance(Lat_a,Lat_nc,Long_a,Long_nc)+get_spherical_distance(Lat_a,Lat_high,Long_a,Long_high)==l_nc:
elif get_spherical_distance(Lat_a,Lat_nc[0],Long_a,Long_nc[0])<l_nc[0] or get_spherical_distance(Lat_a,Lat_nc[1],Long_a,Long_nc[1])<l_nc[1]:
zone.append('Normal city')
zone_class.append('2')
else:
zone.append('Highway')
zone_class.append('3')
#rint(len(zone))
#print(zone)
i+=1
if 'ukhra' in GROUND_TRUTH:
Lat_market=[23.56451,23.5887733333]
Long_market=[87.2823233333,87.2085283333]
Lat_nc=[23.6109329268,23.6407509281]
Long_nc=[87.217555195,87.2265615539]
Lat_high=[23.5505989912,23.6276064719]
Long_high=[87.2686366159,87.2219748885]
Lat=23.6523530924
Long=87.2431245943
l_market.append(get_spherical_distance(Lat_market[0],Lat_high[0],Long_market[0],Long_high[0]))
l_market.append(get_spherical_distance(Lat_market[1],Lat_high[0],Long_market[1],Long_high[0]))
l_nc.append(get_spherical_distance(Lat_nc[0],Lat_high[1],Long_nc[0],Long_high[1]))
l_nc.append(get_spherical_distance(Lat_nc[1],Lat,Long_nc[1],Long))
l_highway.append(get_spherical_distance(Lat_high[0],Lat_market[0],Long_high[0],Long_market[0]))
l_highway.append(get_spherical_distance(Lat_high[1],Lat_nc[0],Long_high[1],Long_nc[0]))
#df=pd.read_csv(name)
l=len(df)
#print(l)
i=0
while i<l:
#print("a ",i)
Lat_a=df.iloc[i]['latitude']
Long_a=df.iloc[i][' longitude']
#if get_spherical_distance(Lat_a,Lat_market,Long_a,Long_market)+get_spherical_distance(Lat_a,Lat_nc,Long_a,Long_nc)==l_market:
if get_spherical_distance(Lat_a,Lat_market[0],Long_a,Long_market[0])<l_market[0] or get_spherical_distance(Lat_a,Lat_market[1],Long_a,Long_market[1])<l_market[1]:
zone.append('Market')
zone_class.append('1')
#elif get_spherical_distance(Lat_a,Lat_nc,Long_a,Long_nc)+get_spherical_distance(Lat_a,Lat_high,Long_a,Long_high)==l_nc:
elif get_spherical_distance(Lat_a,Lat_nc[0],Long_a,Long_nc[0])<l_nc[0] or get_spherical_distance(Lat_a,Lat_nc[1],Long_a,Long_nc[1])<l_nc[1]:
zone.append('Normal city')
zone_class.append('2')
else:
zone.append('Highway')
zone_class.append('3')
#print(len(zone))
#print(zone)
i+=1
if '8B' in GROUND_TRUTH:
Lat_market=[23.56451333,23.52723683,23.5158748,23.49778284]
Long_market=[87.28308833,87.3419682,87.3548276,87.33484252]
Lat_nc=[23.56451333,23.53399268,23.52339667]
Long_nc=[87.28308833,87.34188019,87.33937167]
Lat_high=[23.54631667,23.50425167]
Long_high=[87.31860333,87.35286333]
Lat=23.50697833
Long=87.31082333
l_market.append(get_spherical_distance(Lat_market[0],Lat_high[0],Long_market[0],Long_high[0]))
l_market.append(get_spherical_distance(Lat_market[1],Lat_nc[2],Long_market[1],Long_nc[2]))
l_market.append(get_spherical_distance(Lat_market[2],Lat_high[1],Long_market[2],Long_high[1]))
l_market.append(get_spherical_distance(Lat_market[3],Lat,Long_market[3],Long))
l_nc.append(get_spherical_distance(Lat_nc[0],Lat_market[0],Long_nc[0],Long_market[0]))
l_nc.append(get_spherical_distance(Lat_nc[1],Lat_market[2],Long_nc[1],Long_market[2]))
l_nc.append(get_spherical_distance(Lat_nc[2],Lat_market[2],Long_nc[2],Long_market[2]))
l_highway.append(get_spherical_distance(Lat_high[0],Lat_nc[1],Long_high[0],Long_nc[1]))
l_highway.append(get_spherical_distance(Lat_high[1],Lat_market[3],Long_high[1],Long_market[3]))
#df=pd.read_csv(name)
l=len(df)
#print(l)
i=0
while i<l:
#print("a ",i)
Lat_a=df.iloc[i]['latitude']
Long_a=df.iloc[i][' longitude']
#if get_spherical_distance(Lat_a,Lat_market,Long_a,Long_market)+get_spherical_distance(Lat_a,Lat_nc,Long_a,Long_nc)==l_market:
if get_spherical_distance(Lat_a,Lat_market[0],Long_a,Long_market[0])<l_market[0] or get_spherical_distance(Lat_a,Lat_market[1],Long_a,Long_market[1])<l_market[1] or get_spherical_distance(Lat_a,Lat_market[2],Long_a,Long_market[2])<l_market[2] or get_spherical_distance(Lat_a,Lat_market[3],Long_a,Long_market[3])<l_market[3]:
zone.append('Market')
zone_class.append('1')
#elif get_spherical_distance(Lat_a,Lat_nc,Long_a,Long_nc)+get_spherical_distance(Lat_a,Lat_high,Long_a,Long_high)==l_nc:
elif get_spherical_distance(Lat_a,Lat_nc[0],Long_a,Long_nc[0])<l_nc[0] or get_spherical_distance(Lat_a,Lat_nc[1],Long_a,Long_nc[1])<l_nc[1] or get_spherical_distance(Lat_a,Lat_nc[2],Long_a,Long_nc[2])<l_nc[2]:
zone.append('Normal city')
zone_class.append('2')
else:
zone.append('Highway')
zone_class.append('3')
#print(len(zone))
#print(zone)
i+=1
#WSSprint(zone)
#print(zone)
#print(len(zone))
df['Zone']=zone
df['Zone_class']=zone_class
df.to_csv(name,index=False)
def compare_ground_truth(ground_truth_file, bus_stop_file, OUTPUT_FOLDER,
threshold, GROUND_TRUTH_THRESHOLD, FP_DISTANCE):
gt = read_file(ground_truth_file)
bus_stops = read_file(bus_stop_file)
print "GT: ", len(gt)
print "all bus_stops: ", len(bus_stops)
print gt[0]
print bus_stops[0]
detected = []
false_negative = []
gt_dict = {}
used = []
for gt_point in gt:
distances = []
min_point = None
min_dist = 100000000000000
points = []
for bs_point in bus_stops:
if bs_point in used:
continue
lat1, long1 = gt_point[1:]
lat2, long2 = bs_point[:2]
dist = get_spherical_distance(lat1, lat2, long1, long2)
#the stoppage at the closest distance goes at the end of the list
if dist < min_dist and dist < GROUND_TRUTH_THRESHOLD:
min_dist = dist
min_point = bs_point
points.append(bs_point)
used.append(bs_point)
gt_dict[gt_point[0]] = points
if min_dist < GROUND_TRUTH_THRESHOLD:
detected.append([gt_point[0]] + min_point)
else:
false_negative.append(gt_point)
false_positive = bus_stops[:]
# for i in gt_dict:
# gt_dict[i].sort()
for i, j in gt_dict.items():
print i, len(j)
for gt_id, bs_stops in gt_dict.items():
for point in bs_stops[:-1]:
bus_stops.remove(point)
# print "$ ",len(gt_dict)
for gt_id, bs_stops in gt_dict.items():
# print "#################"
for point in bs_stops:
# print "Deleting ",point
# print gt_id,point
false_positive.remove(point)
##removing false positives who are within 100m distance of each other
false_positive_copy = false_positive[:]
for i in xrange(len(false_positive_copy)):
for j in xrange(i + 1, len(false_positive_copy)):
lat1, long1 = false_positive_copy[i][:2]
lat2, long2 = false_positive_copy[j][:2]
dist = get_spherical_distance(lat1, lat2, long1, long2)
if dist < FP_DISTANCE:
if false_positive_copy[j] in false_positive:
print i, j, dist
print "removing: ", false_positive_copy[j]
false_positive.remove(false_positive_copy[j])
print "effective bus_stops: ", len(bus_stops)
print "detected ", len(detected)
# for bs_point in bus_stops:
# found = False
# for detected_point in detected:
# if bs_point[:2] == detected_point[1:3]:
# found = True
# break
# if found == False:
# false_positive.append(bs_point)
print "FP: ", len(false_positive), "(", len(false_positive) * 100.0 / len(
bus_stops), "%", ")"
print "FN: ", len(
false_negative), "(", len(false_negative) * 100.0 / len(gt), "%", ")"
return detected, false_positive, false_negative
if '54feet' in GROUND_TRUTH and s==22:
s+=1
continue
df=pd.read_csv(input_file)
p=len(df)
j=0
while j<p:
check.append([float(df.iloc[j][0]),float(df.iloc[j][1])])
j+=1
s+=1
i=0
while i<length:
for j in check:
if get_spherical_distance(grd[i][0],j[0],grd[i][1],j[1])<10:
test[i]=1
i+=1
i=0
missing=[]
while i<length:
if test[i]==0:
cnt+=1
missing.append(i)
i+=1
text=open('record_local.txt','a')
text.write(GROUND_TRUTH+'\n')