def add_schedule_adherence_stats_for_route(all_stats, timestamp_intervals, route_config, history_df, timetable_df):
route_id = route_config.id
timetable_df = timetable_df.sort_values('DEPARTURE_TIME', axis=0)
history_df = history_df.sort_values('DEPARTURE_TIME', axis=0)
history_sid_values = history_df['SID'].values
history_did_values = history_df['DID'].values
timetable_sid_values = timetable_df['SID'].values
timetable_did_values = timetable_df['DID'].values
for dir_info in route_config.get_direction_infos():
dir_id = dir_info.id
stop_ids = dir_info.get_stop_ids()
for i, stop_id in enumerate(stop_ids):
stop_history_df = history_df[(history_sid_values == stop_id) & (history_did_values == dir_id)]
stop_timetable_df = timetable_df[(timetable_sid_values == stop_id) & (timetable_did_values == dir_id)]
stop_history_departure_times = stop_history_df['DEPARTURE_TIME'].values
stop_timetable_departure_times = stop_timetable_df['DEPARTURE_TIME'].values
comparison_df = timetables.match_schedule_to_actual_times(
stop_timetable_departure_times,
stop_history_departure_times,
#late_sec=180
)
comparison_df['DEPARTURE_TIME'] = stop_timetable_departure_times
for interval_index, (start_time, end_time) in enumerate(timestamp_intervals):
interval_comparison_df = comparison_df
if start_time is not None:
interval_comparison_df = interval_comparison_df[interval_comparison_df['DEPARTURE_TIME'] >= start_time]
if end_time is not None:
interval_comparison_df = interval_comparison_df[interval_comparison_df['DEPARTURE_TIME'] < end_time]
num_scheduled_departures = len(interval_comparison_df)
if num_scheduled_departures > 0:
on_time_rate = round(np.sum(interval_comparison_df['on_time']) / num_scheduled_departures, 3)
dir_stats = all_stats['combined'][interval_index][route_id]['directions'][dir_id]
dir_stats['onTimeRates'][stop_id] = on_time_rate
for interval_index, _ in enumerate(timestamp_intervals):
add_median_schedule_adherence_stats_for_direction(
all_stats['combined'][interval_index][route_id]['directions'][dir_id]
)
early_sec = early_min * 60
late_sec = late_min * 60
timetable_df['scheduled_headway'] = np.r_[
np.nan,
metrics.compute_headway_minutes(timetable_df['TIME'].values)]
if comparison:
history = arrival_history.get_by_date(agency_id, route_id, d)
arrivals_df = history.get_data_frame(stop_id=stop_id,
direction_id=direction_id)
comparison_df = timetables.match_schedule_to_actual_times(
timetable_df['TIME'].values,
arrivals_df['TIME'].values,
early_sec=early_sec,
late_sec=late_sec)
timetable_df = pd.concat([timetable_df, comparison_df], axis=1)
timetable_df['DATE_TIME'] = timetable_df['TIME'].apply(
lambda t: datetime.fromtimestamp(t, tz))
for row in timetable_df.itertuples():
did = row.DID
dwell_time = util.render_dwell_time(row.DEPARTURE_TIME - row.TIME)
scheduled_headway = f'{round(row.scheduled_headway, 1)}'.rjust(4)
if args.comparison: