IllegalArgumentException in ComputeCnossosRays during diffraction calculation (2 > 0) – NoiseModelling 4.0.5

[IsotoCedex]

NoiseModelling version
4.0.5

Describe the bug
Hello friends of NoiseModelling,
I am computing road maps of some areas of Spain, using NoiseModelling with Scripts.
I have divided the calculation domains by official cartographic sheets.
In some sheets it gives errors, although most are computed normally.

Please, could you help me to find the cause of the error?
Thanks very much!

To Reproduce
Below I leave the files of one of the sheets that gives an error:

• Groovy: Groovy.zip
• SourceData: SurceData.zip

log

[pool-57-thread-7] INFO org.noise_planet.noisemodelling.pathfinder.RootProgressVisitor - 88,24 %
[pool-57-thread-4] ERROR org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays - 2 > 0
java.lang.IllegalArgumentException: 2 > 0
        at java.base/java.util.Arrays.copyOfRange(Arrays.java:3990)
        at java.base/java.util.Arrays.copyOfRange(Arrays.java:3950)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.computeHEdgeDiffraction(ComputeCnossosRays.java:795)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.directPath(ComputeCnossosRays.java:320)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.directPath(ComputeCnossosRays.java:298)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.rcvSrcPropagation(ComputeCnossosRays.java:273)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.computeRaysAtPosition(ComputeCnossosRays.java:232)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays$RangeReceiversComputation.run(ComputeCnossosRays.java:1718)
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
        at java.base/java.lang.Thread.run(Thread.java:834)
Exception in thread "pool-57-thread-4" java.lang.IllegalArgumentException: 2 > 0
        at java.base/java.util.Arrays.copyOfRange(Arrays.java:3990)
        at java.base/java.util.Arrays.copyOfRange(Arrays.java:3950)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.computeHEdgeDiffraction(ComputeCnossosRays.java:795)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.directPath(ComputeCnossosRays.java:320)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.directPath(ComputeCnossosRays.java:298)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.rcvSrcPropagation(ComputeCnossosRays.java:273)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.computeRaysAtPosition(ComputeCnossosRays.java:232)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays$RangeReceiversComputation.run(ComputeCnossosRays.java:1718)
        at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1128)
        at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:628)
        at java.base/java.lang.Thread.run(Thread.java:834)

Expected behavior

java.lang.IllegalArgumentException: 2 > 0
        at java.base/java.util.Arrays.copyOfRange(Arrays.java:3990)
        at java.base/java.util.Arrays.copyOfRange(Arrays.java:3950)
        at org.noise_planet.noisemodelling.pathfinder.ComputeCnossosRays.computeHEdgeDiffraction(ComputeCnossosRays.java:795)

It seems to indicate that, in the routine that calculates the diffraction or the ray tracing (ComputeCnossosRays), it is trying to copy a portion of an array where the start index is greater than the end index, or the expected length is negative (for example, copyOfRange(array, from, to) with from > to).
I thought that the cause was some wrong geometry, but I used QGIS tools to check and correct the geometries, and recalculated, and it still happens the same.
I also modified the density of receptors in Delaunay, to see if some error could be produced by the triangulation, but I have not been able to avoid the error either.

Calculation setup:

// Step 3: Run Calculation
    runScript(connection, "C:/NM_Delaunay_SUPRA01/NoiseModelling_without_gui/scriptrunner/noisemodelling/wps/NoiseModelling/Noise_level_from_source.groovy",
            ["tableBuilding":"Buildings",
             "tableSources":"LW_roads",
             "tableReceivers":"Receivers",
             "tableDEM":"DEM",
             "paramWallAlpha":"0.1",
             "confReflOrder":"1",
             "confMaxReflDist":"50",
             "confMaxSrcDist":"1500",
             "confDiffVertical":"False",
             "confDiffHorizontal":"True",
             "confHumidity":"63",
             "confTemperature":"18",
             "confFavorableOccurrencesDay":"0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5",
             "confFavorableOccurrencesEvening":"0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75, 0.75",
             "confFavorableOccurrencesNight":"1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0",
             ])

Screenshots
Here is the results file for Lday. It can be observed that there comes a point, after the error, in which the calculations fail or simply do not continue.

0001_LDAYIsoContour.zip

And a screenshot:

[nicolas-f]

Hi,

Thanks for the feedback.

We are almost done with the v 5.0 release that contain a almost complete rewrite of the propagation algorithm with extensive cnossos and general tests cases.

I don't know if we will release the fix done on the 4.X branch @pierromond ?

[pierromond]

Thank you @IsotoCedex for this complete feedback of an issue we never saw before (no fast answer...).

I also would like to propose @IsotoCedex to wait the version 5.0 which will bring a lot of new features (and I hope not so much issues :) ) and the rewrite of the propagation algo ... so it should probably resolve this issue. @IsotoCedex can you wait the end of January ?

Kind regards.

[IsotoCedex]

Dear @nicolas-f and @pierromond,
Of course I can wait. For now I have solved it by removing the diffraction calculation in those models.
I think the error has to do with some kind of complex geometry of the buildings, which possibly interacts badly with the digital terrain model. But I have not been able to identify it.
I am looking forward to trying the new version of Noise Modelling.
Thank you very much!