De-duplicate stops returned by stopsByRadius

[leonardehrenfried]

Summary

@binh-dam-ibigroup has reported that there are cases where the stopsByRadius GraphQL query returns duplicate stops. This PR removes the duplicates.

It introduces a new Util class that lets you remove duplicates from a collection even when the thing to deduplicate on is not the whole entity but a field of it.

Unit tests

Added.

Documentation

Javadoc.

[codecov]

Codecov Report

Patch coverage: 80.00% and project coverage change: +0.01% 🎉

Comparison is base (3e36162) 66.48% compared to head (8d39386) 66.49%.
Report is 22 commits behind head on dev-2.x.

Additional details and impacted files

@@              Coverage Diff              @@
##             dev-2.x    #5366      +/-   ##
=============================================
+ Coverage      66.48%   66.49%   +0.01%     
+ Complexity     15238    15236       -2     
=============================================
  Files           1787     1786       -1     
  Lines          69325    69271      -54     
  Branches        7319     7291      -28     
=============================================
- Hits           46088    46064      -24     
+ Misses         20758    20739      -19     
+ Partials        2479     2468      -11     

Files Changed	Coverage Δ
...ext/gtfsgraphqlapi/datafetchers/QueryTypeImpl.java	19.83% <0.00%> (ø)
...opentripplanner/framework/lang/PredicateUtils.java	66.66% <66.66%> (ø)
...pentripplanner/ext/geocoder/StopClusterMapper.java	96.55% <100.00%> (-0.23%)	⬇️
...routing/graphfinder/StopFinderTraverseVisitor.java	100.00% <100.00%> (ø)
...planner/routing/graphfinder/StreetGraphFinder.java	100.00% <100.00%> (ø)

... and 22 files with indirect coverage changes

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

[binh-dam-ibigroup]

The issue mentioned is indeed resolved. Thanks!

[vpaturet]

Oracle discourages the use of stateful filters (https://docs.oracle.com/en/java/javase/14/docs/api/java.base/java/util/stream/package-summary.html#Statelessness), an alternative implementation could use collectors to achieve the same effect:

return stopsFound.stream().distinct().collect(Collectors.toMap(nearbyStop -> nearbyStop.stop, Function.identity(), (n1, n2) -> n1 )).values().stream().toList();

In this particular case, the impact is limited, so approving anyway.

[leonardehrenfried]

That's an excellent point which I will implement in a follow-up PR.