AttributeError: 'dict' object has no attribute 'geometry

[AdrianKriger]

Happy New Year.

With:

cm = [cityjson.load(path=file]]
rds = cm.get_cityobjects(id='1000004')
rds.geometry

Traceback (most recent call last):

  Cell In[131], line 5
    rds.geometry

AttributeError: 'dict' object has no attribute 'geometry'

When I parse the city model through val3dity; it returns a 102 CONSECUTIVE_POINTS_SAME error for a few road objects.
I want to see these repeated points.
How can I see the vertices please?

file -> here

[hugoledoux]

get_cityobjects() returns all the objects, so you need to specify which one (even if only one)

cm = [cityjson.load(path=file]]
rds = cm.get_cityobjects(id='1000004')
rds['1000004'].geometry

[AdrianKriger]

As usual your level of service is fantastic.
Thank you.

geom = rds['1000004'].geometry[0]
geom.boundaries

Gives:

 ...[[[264596.532, 6241575.696, 101.45],
   [264596.13, 6241575.648, 101.45],
   [264596.265, 6241575.657, 101.45]]],
 [[[264595.656, 6241575.666, 101.45],
   [264595.726, 6241575.66, 101.45],
   [264595.593, 6241575.677, 101.45]]],
 [[[264596.265, 6241575.657, 101.45],
   [264596.399, 6241575.673, 101.45],
   [264596.532, 6241575.696, 101.45]]],
 [[[264596.265, 6241575.657, 101.45],
   [264596.336, 6241575.662, 101.45],
   [264596.399, 6241575.673, 101.45]]],
 [[[264595.726, 6241575.66, 101.45],
   [264596.532, 6241575.696, 101.45],
   [264595.593, 6241575.677, 101.45]]],
 [[[264595.529, 6241575.685, 101.45],
   [264595.743, 6241581.233, 101.45],
   [264593.295, 6241576.081, 101.45]]],...

Some of the points are close (these are possible cuvilinear road features) but I don't see duplicates.
Could a tolerance value be producing the error?
Which validation option would I change? --- I'm in a bit restricted with computer / processing or I would run val3dity locally.

How common is this type of error?

The city model has gone from 305 MULTIPLE_CONNECTED_COMPONENTS to 102 CONSECUTIVE_POINTS_SAME.
Is this an improvement?

[hugoledoux]

Could a tolerance value be producing the error?

yes, val3dity has a snap tolerance (https://val3dity.readthedocs.io/en/latest/usage/#snap-tol), you can play with it.

Dunno if improvements, those 2 could be related.

[AdrianKriger]

Ok.
snap tolerance=0.0001 results in 1 invalid 3D geometry -> TINRelief: 305 MULTIPLE_CONNECTED_COMPONENTS.

This is what I had before.
0.0001 seems a bit excessive. If I understand this correctly its saying; some polygons features are 0.0001 meters from the TINRelief.
Is this correct? I can live with a fraction of a mm.
What do downstream applications need? Will this become a challenge if the model is used in later CFD, etc. applications?

[hugoledoux]

no not correct. This says: “if 2 points are closer then X, then we assume that they are the same”. Those points need to be part of the same City Objects, the validation is PER OBJECT, nothing between objects is performed by val3dity.

[AdrianKriger]

Let me see if I understand:
snap_tolerance applies to the 102 CONSECUTIVE_POINTS_SAME error. With a snap_tolerance of 0.0001 we're saying: points, where the mm are different, are not the same.
305 MULTIPLE_CONNECTED_COMPONENTS is concerned with the 'islands' in the TINRelief (ground area in the holes [courtyards] of buildings seperated from the larger terrain).

How would I prevent the 305? Would cjio.triangulate() help?