Val3dity 2.0.2 error 303 and 306 issues

Question

Val3dity 2.0.2 error 303 and 306 issues

Closed this issue 7 years ago · 16 comments

Hello
Great Thanks for solving error 104 issue !!!!!
Online validator in this case works perfect, I hope that you will be able to compile version for Windows soon :-)

During analysing validation reports from new online val3ditity 2.0.2. I noticed new issues.
Now errors 303 and 306 appears for cornet models.
Of course models are correct in my opinion.
Please see attached zipped citygml file:
0620_M-34-46-D-a-3-3.zip

I hope that you could to take a look on this soon :-)
Thank you very much for any help.

Answer 1 · 2018-01-17T08:01:03.000Z

hmmm, are you sure it's the correct file?

I get only errors 204 for this model, both on macOS and on Linux, see for yourself:
http://geovalidation.bk.tudelft.nl/val3dity/reports/overview/d8640efa-2ff0-450e-89b6-654dc4f06e65

Answer 2 · 2018-01-17T08:41:53.000Z

Dear Hugo I ignored error 204 and errors 303 and 306 appeared. Please see attached rapport json file. In your control software detect error 204, shell is invalid and remainder controls are held :-( Marcin

…

Dnia 17 styczeń 2018 o 09:01 Hugo Ledoux ***@***.***> napisał(a): hmmm, are you sure it's the correct file? I get only errors 204 for this model, both on macOS and on Linux, see for yourself: http://geovalidation.bk.tudelft.nl/val3dity/reports/overview/d8640efa-2ff0-450e-89b6-654dc4f06e65 — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub #122 (comment) , or mute the thread https://github.com/notifications/unsubscribe-auth/Ah15sxMd4iSCfDn7N4J1nTekinUgN7OEks5tLai_gaJpZM4RflEX .

Answer 3 · 2018-01-17T08:58:00.000Z

I see.

Ignoring 204 is tricky: it could create another error, this is possibly what is happening here. I hesitated to add that option to be honest, exactly because of this.

But I'll check soon to be sure everything is okay here.

Answer 4 · 2018-01-17T09:14:24.000Z

I looked quickly and isolated the 303 case (attached here), but it seems it's not noded properly and thus non-manifold, no? See:
https://www.dropbox.com/s/g6oxugmu9nu80na/2018-01-17%20at%2010.11.png?dl=0

c2.gml.txt

Answer 5 · 2018-01-17T09:17:34.000Z

If I may chime in, @KraMar76 which CGAL version are you using?
I've just tested your file and I get 203,204 with default settings, and 202,303 with --ignore204
I've just compiled val3dity from the master and have cgal 4.7.4
This is on Linux.

Answer 6 · 2018-01-17T10:32:32.000Z

CGAL v4.10+ is now required, with eigen package, see README.

Answer 7 · 2018-01-18T08:48:16.000Z

hmm, but what is incorrect? I'm not able to find the error. This is one shell and any polygon is outside this shell. All polygons are forming shell. All polygons nodes are identical. Please see attached screens.

…

Dnia 17 styczeń 2018 o 10:14 Hugo Ledoux ***@***.***> napisał(a): I looked quickly and isolated the 303 case (attached here), but it seems it's not noded properly and thus non-manifold, no? See: https://www.dropbox.com/s/g6oxugmu9nu80na/2018-01-17%20at%2010.11.png?dl=0 c2.gml.txt https://github.com/tudelft3d/val3dity/files/1638312/c2.gml.txt — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub #122 (comment) , or mute the thread https://github.com/notifications/unsubscribe-auth/Ah15s5BCCnIl-cwNKGa7FxFnvr_cjkPEks5tLbnwgaJpZM4RflEX .

Answer 8 · 2018-01-18T13:59:33.000Z

nothing is attached

Answer 9 · 2018-01-18T14:00:40.000Z

but 204 caused by vertices very close to each others: you might not be able to see the error.

http://geovalidation.bk.tudelft.nl/val3dity/docs/faq/#i-get-many-errors-203-and-204-but-my-planes-look-planar-to-me-why-is-that

Answer 10 · 2018-01-18T14:22:58.000Z

Pictures I mentioned above

Answer 11 · 2018-01-18T14:35:09.000Z

I checked nodes coordinates in gml file and all are the same.

Even if the polygon is not planar - 204 error
but all coincident nodes in neighbouring polygons are the same shell should be manifold

Answer 12 · 2018-01-23T09:12:14.000Z

okay, I tracked the error, and it is indeed an error. An extreme one due to precision: there are nearly collinear points (there is an unnecessary vertex where the arrow points) and because of the triangulation (I think) it reports a wrong 303 error...
https://www.dropbox.com/s/b8agji6toaydz0c/2018-01-23%20at%2010.06.png?dl=0

If you remove that vertex (which doesn't affect the shape of the building) then it's valid (as expected).

How to prevent such errors is not clear to me right now, I have to think about it. It's first time this happens to my knowledge.

Answer 13 · 2018-01-23T10:39:43.000Z

Hi, thanks for the assist.
I understand your explanation but unfortunately I still haven't seen the error 303 here :-(
The vertex inside is maybe unnecessary but this is in both neighbour polygons. Nodes coordinates are identical so polygons vertexes should be identical too.
Have you have a chance to take a look on 306 error?
Haw do you see this ?

Regards

Answer 14 · 2018-01-23T13:24:17.000Z

Not sure I follow you Marcin: you reported a dataset with errors 303 and 306.

I have found the error and fixed it, your dataset above now is 100% free of errors, including no 204 errors. I'll update the code, test it on different OSs, release 2.0.3, and describe the fix.

Again, thanks for spending time reporting these errors, I really appreciate.

Answer 15 · 2018-01-23T13:58:19.000Z

I would like to apologize for my misunderstanding.
I thought that you had found an error in my data and tried to find this too 8-)

Thanks a lot for your help 👍
I will be waiting for new release.

Answer 16 · 2018-01-23T14:52:01.000Z

v2.0.3 fixes this, released and web-app updated.

Windows will wait a bit, need to find an automatic build system (or switch myself to Windows; just joking not gonna happen)