python -m timeit 'from utils import LCSS; from numpy.random import random_sample; lcss = LCSS(metric = "cityblock", epsilon = 0.1); lcss.compute(random_sample((1000,2)), random_sample((2000, 2)))'
10 loops, best of 3: 1.6 sec per loop

python -m timeit 'from utils import LCSS; from numpy.random import random_sample; lcss = LCSS(similarityFunc = lambda x,y: (abs(x[0]-y[0]) <= 0.1) and (abs(x[1]-y[1]) <= 0.1)); lcss.compute(random_sample((1000,2)), random_sample((2000, 2)))' 
10 loops, best of 3: 3.48 sec per loop
########
Performance report using ../../Data/norway/2_video/20130606-0830.sqlite
####
test query positions
####
no index

real	7m31.857s
user	2m20.962s
sys	5m23.521s
with index on trajectory_id

real	7m43.479s
user	2m25.793s
sys	5m27.585s
with index on trajectory_id and frame_number

real	7m36.917s
user	2m21.628s
sys	5m26.639s
with index on the 2 columns trajectory_id

real	7m28.447s
user	2m15.028s
sys	5m26.697s
query all feature positions, order by id and frame_number (index/no index)

real	0m16.219s
user	0m15.988s
sys	0m0.245s

real	0m16.185s
user	0m15.927s
sys	0m0.272s
####
test query object positions
####
no index

real	0m22.119s
user	0m20.147s
sys	0m1.406s
index on trajectory_id

real	0m21.850s
user	0m19.834s
sys	0m1.401s
index on trajectory_id in positions and objects_features

real	0m25.066s
user	0m23.315s
sys	0m1.265s
Cleaning up!

##############################
indexed: positions et velocities on trajectory_id
time feature-based-tracking ~/Research/Data/nyc/125th-hancock/tracking-125.cfg --database-filename tmp-notindexed.sqlite --gf >> log.txt && time feature-based-tracking ~/Research/Data/nyc/125th-hancock/tracking-125.cfg --database-filename tmp.sqlite --gf >> log-notindexed.txt
feature-based-tracking ~/Research/Data/nyc/125th-hancock/tracking-125.cfg      19985.11s user 24.37s system 99% cpu 5:34:01.61 total
feature-based-tracking ~/Research/Data/nyc/125th-hancock/tracking-125.cfg      20246.41s user 25.64s system 99% cpu 5:38:38.87 total