Closed nimafo closed 2 years ago
Thanks, @nimafo, for opening an issue. We need more information though to be able to reproduce and work on the issue.
Hi @clausnagel and thanks foar your quick response, sorry for incomplete issue, I completed the description.
Can you please verify your download link? I am only able to download the errors_3DcityDB.txt
file but not the CityGML input datasets.
@clausnagel I updated the link, I think it should work now. Sorry for this.
Thanks, the download works now. I validated your .gml files against the CityGML schemas and all of them are invalid. You can easily validate the files yourself using the validate
command of citygml-tools (use version 2.0.0 rather than 2.0.0-rc.1). Some errors and findings:
interior-LOD0.gml
<bldg:lod1MultiSurface>
geometry properties where only one is allowed. No wonder that this leads to geometry errors. gml:id
values. The allowed values for gml:id
are strictly defined in the XML world (see xs:ID
and xs:NCName
). For instance, a gml:id
must not start with a number. interior-LOD1.gml
gml:id
values and invalid order of elements.CityDoctor
v3.12.1 for geometry validation rather than val3dity
. So, no wonder that there are also geometry errors in the CityJSON output.interior-LOD2_2.gml
gml:id
values and invalid order of elements.<bldg:lod2Solid>
properties contain gml:MultiSolid
geometries which are not allowed here. Only gml:Solid
or gml:CompositeSolid
are allowed. That's why the geometries of the buildings cannot be parsed by citygml-tools and hence are not converted to CityJSON.interior-LOD2_3.gml
LOD3_0.gml
gml:id
values and invalid order of elements.FZKViewer
and the CityJSON output with the ninja
web viewer but could not spot a difference. Maybe you can point me to the missing geometries using a screenshot? CityDoctor
already reports invalid geometries for the CityGML file. So, no wonder that there are also geometry errors in the CityJSON output. You might check whether a smaller value for snap_tol
makes a difference when validating with val3dity
.To make a long story short: Your CityGML input files have many issues and are invalid. There is no way for citygml-tools to convert them to clean and valid CityJSON files. So, I see no immediate action for citygml-tools. It rather seems that Random3Dcity needs some fixing.
@clausnagel This is great, thanks a lot for your thorough review of the files.
I am trying to convert citygml files generated with
Random3Dcity
. It is a tool developed for procedurally generating 3D city models. It generates also models for interiors in 4 LoDs (0, 1, 2-2, 2-3). the problem is that after conversion to CityJson:InteriorLoD2-2
andInteriorLoD2-3
is not maintained.InteriorLoD0
andInteriorLoD1
contain geometry, according toval3dity
, whileInteriorLoD1
is valid,InteriorLoD1
has the following problems:when I redo the conversion with
--vertex-precision=20
,val3dity
gives:both for InteriorLoD0 and 1. I did it again with
--vertex-precision=10
and--vertex-precision=15
, which resulted in the same error for LoD1 I mentioned above(104, 303, 307).InteriorLoD2-2
andInteriorLoD2-3
no geometries are maintained after conversion (error 609).LoD3
, not only there is missing geometry for some building installations, but also errors 102: Consecutive points same and 201: Intersection rings.Validation of gmls When validating geometries in 3DCityDB, for all of the 5 files I get the following errors:
The complete description here. and
Other
citygml-tools
, I get the following warning:.\citygml-tools to-cityjson --vertex-precision=20 "path\to\gmls"\*.gml
. as mentioned above, I also did test--vertex-precision=3
which I think is the default value,--vertex-precision=10
, and--vertex-precision=15
. Is this the related to the above error about invalid 'NCName'?citygml-tools, version 2.0.0-rc.1