ignore stop_timezone when calculating arrivals/departures 🐛✅

Apparently, stop_timezone was only ever intended for displaying local date+time.
The current operating practice has been clarified in google/transit#322.

lib/stop_times.js

@@ -179,15 +179,15 @@ WITH stop_times_based AS NOT MATERIALIZED (
 		drop_off_type,
 		shape_dist_traveled,
 		timepoint,
-		coalesce(stations.stop_timezone, stops.stop_timezone, agency.agency_timezone) as tz,
+		agency.agency_timezone as tz,
 		arrival_time,
 		(
 			make_timestamptz(
 				date_part('year', "date")::int,
 				date_part('month', "date")::int,
 				date_part('day', "date")::int,
 				12, 0, 0,
-				coalesce(stations.stop_timezone, stops.stop_timezone, agency.agency_timezone)
+				agency.agency_timezone
 			)
 			- INTERVAL '12 hours'
 			+ arrival_time
@@ -199,7 +199,7 @@ WITH stop_times_based AS NOT MATERIALIZED (
 				date_part('month', "date")::int,
 				date_part('day', "date")::int,
 				12, 0, 0,
-				coalesce(stations.stop_timezone, stops.stop_timezone, agency.agency_timezone)
+				agency.agency_timezone
 			)
 			- INTERVAL '12 hours'
 			+ departure_time
@@ -365,7 +365,7 @@ WITH stop_times_based AS NOT MATERIALIZED (
 			date_part('month'::text, "date")::integer,
 			date_part('day'::text, "date")::integer,
 			12, 0, 0::double precision,
-			coalesce(from_station_tz, from_stop_tz, agency_timezone)
+			agency_timezone
 		) - '12:00:00'::interval + departure_time AS t_departure,
 		departure_time,
 		from_stop_sequence,
@@ -380,7 +380,7 @@ WITH stop_times_based AS NOT MATERIALIZED (
 			date_part('month'::text, "date")::integer,
 			date_part('day'::text, "date")::integer,
 			12, 0, 0::double precision,
-			coalesce(to_station_tz, to_stop_tz, agency_timezone)
+			agency_timezone
 		) - '12:00:00'::interval + arrival_time AS t_arrival,
 		arrival_time,
 		to_drop_off_type,
@@ -404,10 +404,8 @@ WITH stop_times_based AS NOT MATERIALIZED (
 
 			from_stops.stop_id as from_stop_id,
 			from_stops.stop_name as from_stop_name,
-			from_stops.stop_timezone as from_stop_tz,
 			from_stations.stop_id as from_station_id,
 			from_stations.stop_name as from_station_name,
-			from_stations.stop_timezone as from_station_tz,
 			stop_times.stop_headsign as from_stop_headsign,
 			stop_times.pickup_type as from_pickup_type,
 			stop_times.departure_time as departure_time,
@@ -423,10 +421,8 @@ WITH stop_times_based AS NOT MATERIALIZED (
 			to_stop_times.stop_headsign as to_stop_headsign,
 			to_stop_times.stop_id as to_stop_id,
 			to_stops.stop_name as to_stop_name,
-			to_stops.stop_timezone as to_stop_tz,
 			to_stations.stop_id as to_station_id,
-			to_stations.stop_name as to_station_name,
-			to_stations.stop_timezone as to_station_tz
+			to_stations.stop_name as to_station_name
 		FROM "${opt.schema}".trips
 		LEFT JOIN "${opt.schema}".routes ON trips.route_id = routes.route_id
 		LEFT JOIN "${opt.schema}".agency ON (

test/amtrak-gtfs-2021-10-06.sh

@@ -29,16 +29,16 @@ and date = '2021-11-26'
 order by t_arrival
 EOF)
 
-# 2021-11-26T15:15:00-06:00
+# 2021-11-26T15:15:00-05:00
 arr1=$(psql --csv -t -c "$query" | head -n 1)
-if [[ "$arr1" != "1637961300" ]]; then
+if [[ "$arr1" != "1637957700" ]]; then
 	echo "invalid 1st t_arrival: $arr1" 1>&2
 	exit 1
 fi
 
-# 2021-11-27T13:45:00-06:00
+# 2021-11-27T13:45:00-05:00
 arrN=$(psql --csv -t -c "$query" | tail -n 1)
-if [[ "$arrN" != "1638042300" ]]; then
+if [[ "$arrN" != "1638038700" ]]; then
 	echo "invalid 2nd t_arrival: $arrN" 1>&2
 	exit 1
 fi
@@ -86,14 +86,14 @@ SELECT nr_of_active_trips
 FROM amtrak.stats_active_trips_by_hour
 WHERE "hour" = '2021-11-26T04:00-05'
 EOF)
-# Note: I'm not sure if 146 is correct, but it is in the right ballpark. 🙈
-# The following query yields 175 connections, and it doesn't contain those who depart earlier and arrive later.
+# Note: I'm not sure if 127 is correct, but it is in the right ballpark. 🙈
+# The following query yields 150 connections, and it doesn't contain those who depart earlier and arrive later.
 # SELECT DISTINCT ON (trip_id) *
 # FROM amtrak.connections
 # WHERE t_departure >= '2021-11-26T02:00-05'
 # AND t_arrival <= '2021-11-26T06:00-05'
 nrOfActiveTrips=$(psql --csv -t -c "$nrOfActiveTripsQuery" | tail -n 1)
-if [[ "$nrOfActiveTrips" != "146" ]]; then
+if [[ "$nrOfActiveTrips" != "127" ]]; then
 	echo "unexpected no. of active trips at 2021-11-26T04:00-05: $nrOfActiveTrips" 1>&2
 	exit 1
 fi