Sumolib: Unexpected Upstream Edges From GetDownstreamEdges

Hey everyone in the SUMO community,

I've run into a bit of a puzzling situation while using the getDownstreamEdges(edge, distance, stopOnTLS, stopOnTurnaround) function in Sumolib, and I'm hoping you can help me unravel it. Basically, I'm trying to figure out why I'm seeing edges that seem to be upstream of my starting edge when I'm expecting to only see downstream edges.

The Issue: Unexpected Upstream Edges

So, here's the deal. I'm using Sumolib's getDownstreamEdges function, which, as I understand it, should return all the edges downstream of a given edge within a specified distance. I've set it up to stop at turnarounds but not at traffic lights. But, the results I'm getting are a bit…unexpected.

To give you a clearer picture, I've attached a test case (test.zip) that you can use to reproduce the issue. In my test, I'm using these commands:

edge_object = network.getEdge('-E6.65')
downstream_edges = network.getDownstreamEdges(edge_object, 1500, False, True)

What I'm expecting is a list of edges that are downstream from the edge -E6.65. However, when I run this, the output includes edges that appear to be upstream based on the network's layout. I've included some images below to illustrate what I'm seeing:

As you can see in the images, some of the returned edges seem to be leading into -E6.65, rather than flowing out of it. This is definitely throwing a wrench in my calculations and simulations.

Digging Deeper into the Problem

I've been scratching my head trying to figure out what's going on. Is it possible that I'm misunderstanding how the getDownstreamEdges function works? Am I missing a crucial setting or parameter? Or could there be a potential bug in the function itself?

I've double-checked my setup and the documentation, but I haven't been able to pinpoint the root cause. The distance parameter is set to 1500 meters, which should be more than enough to cover the downstream edges I'm interested in. I've also set stopOnTLS to False because I want the function to traverse traffic lights, and stopOnTurnaround is set to True to prevent it from going backward on the same edge. These settings should be correct for what I'm trying to achieve.

Could it be something related to the network topology itself? Perhaps there's a specific configuration in my network that's causing the function to behave unexpectedly. Or maybe there's a subtle edge connection that I'm overlooking.

I've also considered the possibility of numerical precision issues, but I'm not sure how that would manifest in this particular case. It's a long shot, but I'm open to any suggestions at this point.

Possible Causes and Solutions

Let's brainstorm some possible explanations and potential solutions for this unexpected behavior. This is where your expertise and insights would be incredibly valuable.

1. Misunderstanding the Function: It's always possible that I'm simply misunderstanding how getDownstreamEdges is intended to work. Perhaps there's a nuance in its behavior that I'm not aware of. If anyone has a deeper understanding of this function, please enlighten me!

2. Error in Setup: Maybe there's something wrong with my network setup or the way I'm calling the function. I've tried to be as careful as possible, but it's always possible to make a mistake. If you spot anything in my code or approach that seems off, please let me know.

3. Potential Issue with the Function: While I'm hesitant to jump to this conclusion, there's a chance that there could be a bug or an edge case in the getDownstreamEdges function that's causing it to return incorrect results. If this is the case, it would be important to identify the issue and report it to the SUMO developers.

4. Network Topology Quirks: As I mentioned earlier, the specific topology of my network might be contributing to the problem. Perhaps there are unusual connections or configurations that are confusing the function. It would be helpful to analyze the network structure more closely to see if anything stands out.

5. Distance Calculation Issues: It's also conceivable that there's an issue with how the distance is being calculated within the function. Maybe there's a discrepancy between the distance along the edges and the straight-line distance, or perhaps there's a rounding error that's accumulating and leading to incorrect results.

Seeking Community Wisdom

I'm really hoping that someone in the SUMO community has encountered this issue before or has some insights into what might be going on. Your experience and expertise would be greatly appreciated.

Have you ever seen getDownstreamEdges return upstream edges? Do you have any suggestions for how to debug this further? Are there any specific aspects of my setup or network that I should investigate?

Any help you can offer would be fantastic. I'm eager to get to the bottom of this and ensure that my simulations are accurate and reliable.

Next Steps and Further Investigation

To help narrow down the possibilities, I'm planning to take a few more steps in my investigation. These include:

• Simplifying the Network: I'll try creating a much simpler network with a clear and straightforward topology to see if the issue persists. This will help rule out any network-specific complexities.
• Debugging the Function: If I'm feeling brave, I might try to dive into the Sumolib source code and step through the getDownstreamEdges function to see exactly how it's calculating the downstream edges. This could be a bit of a challenge, but it might be necessary to pinpoint the problem.
• Experimenting with Parameters: I'll try varying the distance parameter and other settings to see if that has any impact on the results. This could help identify any sensitivity to specific values.

I'll keep you updated on my progress and any findings I uncover. In the meantime, please feel free to share any thoughts or suggestions you might have. Let's work together to solve this puzzle!

Thanks in advance for your help, guys. I really appreciate it!