Add support for via-location visits in Raptor access and egress paths #7101

t2gran · 2025-11-28T17:23:11Z

Summary

Build on top of #7095 and #7093.

Add support for via-location visits in access and egress paths in Raptor. This enables via points to be reached as part of access or egress. Currently we do not support flex, but the Raptor code do.

This PR extends Raptor's via-location search capability to handle via points that are visited in access or egress paths, not just during transit routing.

How the code works

See the Raptor desing doc

The implementation tracks the number of via locations visited in access/egress paths and uses a chain of Raptor workers+state. When a via-location is reached the state for that stop is coped over the the next segment. Driving this is a composit-worker and stop-arrival-events at the via-stops.

Key classes:

RangeRaptorWorkerComposite - chains Raptor searches together
RouterResultPathAggregator - aggregates results from multiple segments
ViaConnectionStopArrivalEventListener - copies stop arrivals between segments
ArrivalsEventListenerMapper - builds and orders arrival listeners
AccessEgressFunctions - filters access/egress by via count

Issue

Part of #4887

Unit tests

Added comprehensive module tests:
- J04_ViaVisitWithAccessTest - tests via locations visited in access paths
- J05_ViaVisitWithEgressTest - tests via locations visited in egress paths
Updated existing tests

Documentation

✅ Design doc
✅ JavaDoc

Changelog

✅ Should be part of support for via

codecov · 2025-11-28T17:34:40Z

Codecov Report

❌ Patch coverage is 94.68085% with 10 lines in your changes missing coverage. Please review.
✅ Project coverage is 72.13%. Comparing base (fc3f984) to head (3469c83).
⚠️ Report is 23 commits behind head on dev-2.x.

Files with missing lines	Patch %	Lines
...angeraptor/support/RouterResultPathAggregator.java	73.33%	4 Missing ⚠️
...r/raptor/util/paretoset/ParetoSetWithListener.java	75.00%	0 Missing and 3 partials ⚠️
...entripplanner/raptor/api/request/SearchParams.java	95.23%	1 Missing ⚠️
...raptor/rangeraptor/RangeRaptorWorkerComposite.java	96.00%	0 Missing and 1 partial ⚠️
...lanner/raptor/rangeraptor/transit/AccessPaths.java	92.85%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@              Coverage Diff              @@
##             dev-2.x    #7101      +/-   ##
=============================================
+ Coverage      72.09%   72.13%   +0.04%     
- Complexity     20706    20757      +51     
=============================================
  Files           2241     2242       +1     
  Lines          83725    83874     +149     
  Branches        8355     8380      +25     
=============================================
+ Hits           60362    60503     +141     
- Misses         20433    20437       +4     
- Partials        2930     2934       +4

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

…tions visited in Raptor

…stener

The McStopArrivals is a critical class, building the listeners is just noice and can be done in a mapper and injected. This also remove StopArrivalParetoSet - witch only existed to take part of the listeners construction. A tiny business rule is no enforced by the the mapper, not ideal, but overall much better than it was. The mapper is responsible for the order the listeners are added.

…s from all workers.

optionsome · 2025-12-12T22:30:36Z

raptor/src/main/java/org/opentripplanner/raptor/api/request/SearchParams.java

+      .max()
+      .orElse(0);
+
+    if (maxAccessViaVisits + maxEgressViaVisits > numberOfViaVisits) {


This logic needs to be updates since the same via can be visited both in access and egress.

jessicaKoehnke · 2025-12-09T15:56:23Z

...opentripplanner/raptor/rangeraptor/multicriteria/arrivals/StopArrivalStateParetoSetTest.java

-        "Stop Arrival - w/egress",
-        StopArrivalParetoSet.of(comparator).withEgressListener(List.of(), null).build()
-      )
+      Arguments.of("Stop Arrival - regular", ParetoSet.of(comparator)),


they are both the same test case just under a different name

jessicaKoehnke · 2025-12-10T10:36:04Z

raptor/src/main/java/org/opentripplanner/raptor/rangeraptor/context/SearchContext.java

-          null
+          accessPaths.filterOnSegment(i),
+          viaConnection,
+          egressPaths.filterOnSegment(nConnections - i)


since you don't set the egress paths to null anymore, they shouldn't be nullable in the SearchContextViaSegments either

jessicaKoehnke · 2025-12-10T12:12:13Z

raptor/src/main/java/org/opentripplanner/raptor/rangeraptor/transit/AccessPaths.java

+    );
+  }
+
+  static TIntObjectMap<List<RaptorAccessEgress>> filter(


since this is used by both AccessPaths and EgressPaths I think this method might be better located in AccessEgressFunctions.

jessicaKoehnke · 2025-12-10T12:17:11Z

...java/org/opentripplanner/raptor/rangeraptor/multicriteria/configure/McRangeRaptorConfig.java

  }

  /**
   * Sets the next leg state. This is used to connect the state created by this config with the


some java doc still refers to segments as legs

jessicaKoehnke · 2025-12-15T11:03:37Z

...g/opentripplanner/raptor/rangeraptor/multicriteria/arrivals/ArrivalsEventListenerMapper.java

+
+  public static <T extends RaptorTripSchedule> TIntObjectMap<
+    ParetoSetEventListener<ArrivalView<T>>
+  > of(


It's nice to separate this logic out, but I am not sure I like the naming here. When I call xxx.of(...) I'd expect to get a new instance of type xxx

jessicaKoehnke · 2025-12-15T14:45:34Z

raptor/src/main/java/org/opentripplanner/raptor/util/paretoset/ParetoSet.java

  }

-  @Override
-  public boolean remove(Object o) {


why could this be deleted?

jessicaKoehnke · 2025-12-15T14:55:00Z

raptor/src/test/java/org/opentripplanner/raptor/util/paretoset/ParetoSetWithMarkerTest.java


    if (firstVectorAfterMarker == null) {
-      set.forEach(it -> buf.add(it.toString()));
+      set.stream().forEach(it -> buf.add(it.toString()));


why did you add .stream() here and in the else block? Using a stream before forEach is just overhead, isn't it?

jessicaKoehnke · 2025-12-15T14:58:48Z

raptor/src/main/java/org/opentripplanner/raptor/util/paretoset/ParetoSet.java

   * while using this iterator.
   */
-  private Iterator<T> tailIterator(final int startInclusive) {
+  private final Iterator<T> tailIterator(final int startInclusive) {


private methods can't be overridden, so the final indicator is redundant

jessicaKoehnke · 2025-12-15T15:00:16Z

raptor/src/main/java/org/opentripplanner/raptor/util/paretoset/ParetoSetWithListener.java

+  }
+
+  @Override
+  protected final void notifyElementAccepted(T newElement) {


the class is final, so the methods don't need to be

jessicaKoehnke · 2025-12-15T15:10:06Z

raptor/src/main/java/org/opentripplanner/raptor/util/paretoset/ParetoSetWithListener.java

+
+  @Override
+  protected final void notifyElementDropped(T element, T droppedByElement) {
+    if (eventListener != null) {


the if clause is not necessary anymore. I'm not sure how much is actually won by putting the listener stuff into a child class. It's basically just dropping this if clause and having the very basic logic of calling the eventListener in another place. Not sure.

optionsome · 2025-12-16T12:37:43Z

raptor/src/main/java/org/opentripplanner/raptor/api/model/RaptorAccessEgress.java

  /**
   * In a via-search (both pass-through and visit-via) the access/egress may contain one
-   * ore more via-locations. If so Raptor needs to know how many via-locations are included
+   * ore more via-locations. If so, Raptor needs to know how many via-locations are included


Suggested change

* ore more via-locations. If so, Raptor needs to know how many via-locations are included

* or more via-locations. If so, Raptor needs to know how many via-locations are included

optionsome · 2025-12-16T12:40:32Z

raptor/src/main/java/org/opentripplanner/raptor/api/model/RaptorAccessEgress.java

+  default void validateAccessEgressVisitVia(int numberOfViaLocations) {
+    int viaVisits = numberOfViaLocationsVisited();
+
+    if (viaVisits < 0) {
+      throw new IllegalArgumentException(
+        "Access/Egress cannot have negative via visits: " + viaVisits
+      );
+    }
+
+    if (viaVisits > numberOfViaLocations) {
+      throw new IllegalArgumentException(
+        "Access/Egress visits " +
+        viaVisits +
+        " via locations, but only " +
+        numberOfViaLocations +
+        " are defined"
+      );
+    }
+  }


I'm not sure if this kind of validation should be when you set the number of via locations instead of when you try to use them?

optionsome · 2025-12-16T12:49:32Z

...g/opentripplanner/raptor/rangeraptor/multicriteria/arrivals/ArrivalsEventListenerMapper.java

+   * the "glue" make sure new destination arrivals are added to the destination arrivals.
+   */
+  private TIntObjectMap<ParetoSetEventListener<ArrivalView<T>>> map() {
+    // debugListener, nextSearchListener, egressListener


optionsome · 2025-12-16T12:57:32Z

...src/main/java/org/opentripplanner/raptor/rangeraptor/support/RouterResultPathAggregator.java

+public class RouterResultPathAggregator<T extends RaptorTripSchedule>
+  implements RaptorRouterResult<T> {
+
+  private final RaptorRouterResult<T> master;


What does master mean in this context?

optionsome · 2025-12-16T12:59:36Z

raptor/src/main/java/org/opentripplanner/raptor/rangeraptor/transit/AccessEgressFunctions.java

      set.clear();
      set.addAll(list);
-      result.addAll(set);
+      result.addAll(set.stream().toList());


You are repeating this stream.ToList() on at least a couple of places. Should ParetoSet itself have asList method?

optionsome · 2025-12-16T13:21:28Z

raptor/src/test/java/org/opentripplanner/raptor/moduletests/J05_ViaVisitWithEgressTest.java

+    assertEquals(
+      """
+      Walk 1m ~ A ~ BUS R1 0:05 0:15 ~ C ~ BUS R2 0:18 0:20 ~ D ~ Walk 1m [0:04 0:21 17m Tₙ1 C₁2_340]
+      Walk 1m ~ A ~ BUS R1 0:05 0:10 ~ B ~ Walk 15m Vₙ1 [0:04 0:25 21m Tₙ0 C₁2_820]""",


The formatting is a bit odd with """ at the end being on the same line. Is this something that prettier forces or something you can affect? There are other instances of this formatting in this class.

t2gran added this to the 2.9 (next release) milestone Nov 28, 2025

t2gran requested a review from a team as a code owner November 28, 2025 17:23

t2gran added the !New Feature A functional feature targeting the end user. label Nov 28, 2025

t2gran force-pushed the via-access-egress-raptor-impl branch from b80158a to 9d53204 Compare November 28, 2025 17:26

t2gran assigned optionsome Nov 28, 2025

t2gran requested a review from optionsome November 28, 2025 17:28

t2gran mentioned this pull request Nov 28, 2025

Add support for via access/egress in Raptor API (API only) #7093

Merged

t2gran force-pushed the via-access-egress-raptor-impl branch from 9d53204 to 8dc0804 Compare November 28, 2025 17:41

t2gran added 13 commits December 8, 2025 13:17

feature: Implement validation for via visits in access/egress paths

0e6fb5b

feature: Add support for formatting and validating number of via loca…

c8b8b38

…tions visited in Raptor

feature: Initial support for via in access/egress

e2b4b06

refactor: RaptorConfig connectNext SEGMENT, not leg

6809332

feature: Include num-of-via in access/egress filter in Raptor

765c311

refactor: Decouple McStopArrivals <-> ViaConnectionStopArrivalEventLi…

fb43e54

…stener

refactor: Move McStopArrivals and related classes to arrivals package

b21f4df

refactor: Small cleanup of the pareto set in Raptor.

ae7c8d2

feature: Add support for via-location in egress by aggregating result…

d8931f3

…s from all workers.

docs: Add design documentation for via-location

d9c8806

refactor: Merge ParetoSet and ParetoSetWithMarker

8d817f8

refactor: Split ParetoSet and ParetoSetWithListener

3469c83

t2gran force-pushed the via-access-egress-raptor-impl branch from 26bc2e9 to 3469c83 Compare December 8, 2025 13:07

optionsome assigned jessicaKoehnke Dec 9, 2025

optionsome requested a review from jessicaKoehnke December 9, 2025 09:57

optionsome reviewed Dec 12, 2025

View reviewed changes

jessicaKoehnke requested changes Dec 15, 2025

View reviewed changes

optionsome reviewed Dec 16, 2025

View reviewed changes

	* ore more via-locations. If so, Raptor needs to know how many via-locations are included
	* or more via-locations. If so, Raptor needs to know how many via-locations are included

Add support for via-location visits in Raptor access and egress paths #7101

Are you sure you want to change the base?

Add support for via-location visits in Raptor access and egress paths #7101

Conversation

t2gran commented Nov 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Issue

Unit tests

Documentation

Changelog

Uh oh!

codecov bot commented Nov 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

t2gran commented Nov 28, 2025 •

edited

Loading

codecov bot commented Nov 28, 2025 •

edited

Loading