Incorrect Stop Direction Compared to Java Output

[Ahmedhossamdev]

While working on stops-for-location, I found several stops where the direction field does not match the Java API output. This inconsistency affects API clients that rely on correct stop headings.

This issue may be related to differences in how direction is computed.

[Santosh69]

Hi @Ahmedhossamdev , I'm Santosh
I’ve been going through the Maglev codebase to understand this issue fully, and I’d like to contribute a fix for #135.

After tracing the stops-for-location flow end-to-end, it looks like this endpoint is the only one that doesn’t use the shared direction-calculation path. Other endpoints (like stops-for-route and trip-details) call api.calculateStopDirection, which internally uses the original Java logic (via the precomputed directions or the route-path fallback). But in stops-for-location, the handler just reads the raw stop.Direction field from the GTFS DB. That seems to be the main reason this endpoint returns different compass directions compared to the Java API.

From what I understood, the likely file needing a small update is:

internal/restapi/stops_for_location_handler.go

→ Using api.calculateStopDirection(ctx, stop.ID) here would make this endpoint consistent with the others and match the Java behavior.

Before I start preparing the patch and tests, I wanted to check with you:

Are you okay with updating stops-for-location to follow the shared direction pipeline?

And would you prefer relying on the simple calculator or the advanced one, if both are available in the current setup?

If this approach looks good to you, I’d be happy to work on this issue and open a PR.
Could you assign this one to me so I can start moving in the right direction?

Thanks!

[Ahmedhossamdev]

Hi @Santosh69,

We just need to make sure that any stop returned by any API has the same direction as the Java server.

[Santosh69]

Sure @Ahmedhossamdev, I’ll re-verify that all endpoints return the same direction as the Java server and keep everything consistent. I’ll also make sure the stops-for-location endpoint matches the Java output as expected.

I’ve finished the setup and already started working on the fix.
could you assign the issue to me?

I'm done with the setup and started working on the fix accordingly
could you assign the issue to me?

[Ahmedhossamdev]

@Santosh69 Done

[MohamedKamal000]

I’ll try to investigate why this is happening. I already have the Java OBA code set up on my machine, so I’ll compare the calculation logic between both codebases.
If the issue is still present, could you assign it to me? @Ahmedhossamdev

[GaneshPatil7517]

I am working on it...