Busiest day of the service is not exhaustive about partridge

[invisiblefunnel]

The following service_ids are missing in the output : [B.613,C.613,F.613,E.613,U.613,S.613]

Thanks for sharing your findings @praneethd7. I took a look at the linked GTFS file and it matches the results from partridge.

% curl -L -o trimet.zip https://transitfeeds.com/p/trimet/43/20220201/download
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
100 18.1M  100 18.1M    0     0  16.6M      0  0:00:01  0:00:01 --:--:-- 31.9M
% unzip trimet.zip calendar.txt calendar_dates.txt 
Archive:  trimet.zip
  inflating: calendar.txt            
  inflating: calendar_dates.txt  
% cat calendar_dates.txt | grep 20220124
A.613,20220124,1
D.613,20220124,1
W.613,20220124,1
Q.613,20220124,1
% cat calendar.txt 
service_id,monday,tuesday,wednesday,thursday,friday,saturday,sunday,start_date,end_date
B.613,0,0,0,0,0,0,0,20220123,20220205
B.614,0,0,0,0,0,0,0,20220206,20220514
C.613,0,0,0,0,0,0,0,20220123,20220205
C.614,0,0,0,0,0,0,0,20220206,20220514
A.613,0,0,0,0,0,0,0,20220123,20220205
A.614,0,0,0,0,0,0,0,20220206,20220514
F.613,0,0,0,0,0,0,0,20220123,20220205
F.614,0,0,0,0,0,0,0,20220206,20220514
E.613,0,0,0,0,0,0,0,20220123,20220205
E.614,0,0,0,0,0,0,0,20220206,20220514
D.613,0,0,0,0,0,0,0,20220123,20220205
D.614,0,0,0,0,0,0,0,20220206,20220514
W.613,0,0,0,0,0,0,0,20220123,20220205
W.614,0,0,0,0,0,0,0,20220206,20220514
Q.613,0,0,0,0,0,0,0,20220123,20220205
Q.614,0,0,0,0,0,0,0,20220206,20220514
U.613,0,0,0,0,0,0,0,20220123,20220205
U.614,0,0,0,0,0,0,0,20220206,20220514
S.613,0,0,0,0,0,0,0,20220123,20220205
S.614,0,0,0,0,0,0,0,20220206,20220514

The missing service_ids include both Light Rail & Bus route_type. For example B.613 (Light Rail) consists of the route 'MAX Red Line' that operates Monday-Friday & Weekends. This is perhaps most busiest line as it connects the Portland Airport. Also U.613 (Bus) consists of 48 routes. All the routes in this service_id can be seen [here].(http://gtfs.transitq.com/TriMet_20220201_20220201/serviceids/U.613)

Take a look at http://gtfs.transitq.com/TriMet_20220201_20220201/serviceids/A.613 and this gist showing that on 20220124 trips for the MAX Red Line are covered by service_id A.613.

[praneethd7]

Thank you @invisiblefunnel for the quick response and gist. I had an incorrect notion that service_ids in calendars.txt must be operational on all days between the start_date and end_date. Also I was unde the impression that two service_id have no overlap of routes. After your response, I realized that despite 24th January, 2022 missing service_ids : [B.613,C.613,F.613,E.613,U.613,S.613] , the routes in these service_ids are covered by ['A.613', 'D.613', 'Q.613', 'W.613'] (reported by partridge). However, I am still missing how the busiest day is actually reported. Is it the day with the maximum number of service_id in the output of _service_ids_by_date()?

[invisiblefunnel]

However, I am still missing how the busiest day is actually reported. Is it the day with the maximum number of service_id in the output of _service_ids_by_date()?

Partridge uses the number of trips to approximate busyness. The earliest date is returned if multiple dates have the same number of trips.

partridge/partridge/readers.py

Lines 57 to 60 in df3167e

	def read_busiest_date(path: str) -> Tuple[datetime.date, FrozenSet[str]]:
	"""Find the earliest date with the most trips"""
	feed = load_raw_feed(path)
	return _busiest_date(feed)

partridge/partridge/readers.py

Lines 117 to 128 in df3167e

	def _busiest_date(feed: Feed) -> Tuple[datetime.date, FrozenSet[str]]:
	service_ids_by_date = _service_ids_by_date(feed)
	trip_counts_by_date = _trip_counts_by_date(feed)
	def max_by(kv: Tuple[datetime.date, int]) -> Tuple[int, int]:
	date, count = kv
	return count, -date.toordinal()
	date, _ = max(trip_counts_by_date.items(), key=max_by)
	service_ids = service_ids_by_date[date]
	return date, service_ids

partridge/partridge/readers.py

Lines 222 to 229 in df3167e

	def _trip_counts_by_date(feed: Feed) -> Dict[datetime.date, int]:
	results: DefaultDict[datetime.date, int] = defaultdict(int)
	trips = feed.trips
	for service_ids, dates in _dates_by_service_ids(feed).items():
	trip_count = trips[trips.service_id.isin(service_ids)].shape[0]
	for date in dates:
	results[date] += trip_count
	return dict(results)

[praneethd7]

That makes sense! Thank you so much @invisiblefunnel!