queryScheduleStopTimes(): warn against matching ambiguously without stop_{id,sequence}

Author: derhuerst

lib/match-alert.js

@@ -144,6 +144,7 @@ const createMatchAlert = (cfg) => {
 			// // todo: what if there are >1 matches? the match should be unambiguous, right?
 			// const isMatch = scheduleStopTimes => scheduleStopTimes.length > 0
 			// const scheduleStopTimes = await queryScheduleStopTimes({
+			// 	logger,
 			// 	route_id,
 			// 	start_date,
 			// 	trip_id: realtimeTripId,

lib/match-trip-update.js

@@ -121,6 +121,7 @@ const createMatchTripUpdate = (cfg) => {
 		// todo: use some StopTimeUpdate's stop_id/stop_sequence to unambiguously identify the trip "instance", then query all its stop_times
 		const isMatch = scheduleStopTimes => scheduleStopTimes.length > 0
 		const scheduleStopTimes = await queryScheduleStopTimes({
+			logger,
 			route_id,
 			start_date: startDate,
 			trip_id: realtimeTripId,

lib/match-vehicle-position.js

@@ -72,6 +72,7 @@ const createMatchVehiclePosition = (cfg) => {
 
 		const isMatch = scheduleStopTimes => scheduleStopTimes.length === 1
 		const scheduleStopTimes = await queryScheduleStopTimes({
+			logger,
 			route_id,
 			start_date: startDate,
 			trip_id: realtimeTripId,

lib/match.js

@@ -12,7 +12,7 @@ import {createMatchAlert} from './match-alert.js'
 import {createStoreAndRestoreStopTimeUpdatesFromDb} from './restore-stoptimeupdates.js'
 import {createApplyTripReplacementPeriods} from './apply-trip-replacement-periods.js'
 
-const MATCHING_LOG_LEVEL = process.env.LOG_LEVEL_MATCHING || 'error'
+const MATCHING_LOG_LEVEL = process.env.LOG_LEVEL_MATCHING || 'warn'
 
 const parseEncodedFeed = (feedEncoded) => {
 	// decode feed, validate NyctFeedHeader

lib/query-schedule-stop-times.js

@@ -66,6 +66,7 @@ const subdivisionsByRouteId = new Map([
 // todo: query & expose start_time?
 const queryScheduleStopTimes = async (cfg) => {
 	const {
+		logger,
 		route_id,
 		start_date,
 		trip_id,
@@ -84,6 +85,26 @@ const queryScheduleStopTimes = async (cfg) => {
 	ok(trip_id, 'missing/empty trip_id')
 	strictEqual(typeof isMatch, 'function', 'isMatch must be a function')
 
+	// Without stop_id, it is not possible to unambiguously identify the targeted trip "instance", as there might be two trips with the same trip_id *suffix* running by the same stop one the same date.
+	if (stop_id === null) {
+		logger.warn({
+			route_id,
+			start_date,
+			trip_id,
+			stop_id,
+			stop_sequence,
+		}, 'matching without stop_id, risking incorrectly matched trip')
+	// On top, with trips running loops (visiting a stop more than once), we need stop_sequence to uniquely identify which visit to match.
+	} else if (stop_sequence === null) {
+		logger.warn({
+			route_id,
+			start_date,
+			trip_id,
+			stop_id,
+			stop_sequence,
+		}, 'matching without stop_id, risking incorrectly matched stop_times row')
+	}
+
 	const metricsCtx = {
 		schedule_feed_digest: scheduleFeedDigestSlice,
 		route_id,