def detect_stops(dates, network, **param):
"""
Main function to be run to learn the mixture distribution from the TSpot
Input:
tspots_data: list of list of TSpot.
Each list of TSpot represents a trajectory provided by the PIF
network: network object
"""
learned_mixtures = {}
tspots_data = [ttob_seq for date in dates
for ttob_seq in pip_fun.getDayTSpotsInterpolated(date,
network)]
tspots_cut_link = [pip_fun.seqGroupBy(tspots, keyf=lambda tsp:tsp.spot.linkId) for tspots in tspots_data]
tspots_groups_per_link = pip_fun.groupby([link_tspots for traj_tspots in tspots_cut_link
for link_tspots in traj_tspots], lambda tspots:tspots[0].spot.linkId)
for lid, link_obs in tspots_groups_per_link:
stopping_obs = ([detect_stops_on_link_traj(traj, **param)
if len(traj) >= 2
else None
for traj in link_obs])
non_stop_mix = compute_tt_from_tspots(
[traj for i, traj in enumerate(link_obs) if stopping_obs[i] is False],
lid,
network,
avg_ff_tt=None,
**param)
stop_mix = compute_tt_from_tspots(
[traj for i, traj in enumerate(link_obs) if stopping_obs[i] is True],
lid,
network,
avg_ff_tt=non_stop_mix[0],
**param)
mixt_dist = compute_mixture(non_stop_mix, stop_mix, param['non_stopping_default'])
learned_mixtures[lid] = mixt_dist
return learned_mixtures