Invalid geometries in buildings

[maning]

Running the process for the whole LA County, I get this error:

time make chunks

Traceback (most recent call last):
  File "chunk.py", line 73, in <module>
    chunk(argv[1], argv[2], argv[3], argv[4])
  File "chunk.py", line 42, in chunk
    if asShape(features[j]['geometry']).intersects(sectionShape):
  File "/home/penny/venvs/labuildings/local/lib/python2.7/site-packages/shapely/geometry/base.py", line 449, in intersects
    return bool(self.impl['intersects'](self, other))
  File "/home/penny/venvs/labuildings/local/lib/python2.7/site-packages/shapely/predicates.py", line 11, in __call__
    return self.fn(this._geom, other._geom, *args)
  File "/home/penny/venvs/labuildings/local/lib/python2.7/site-packages/shapely/geos.py", line 187, in errcheck_predicate
    raise PredicateError("Failed to evaluate %s" % repr(func))
shapely.geos.PredicateError: Failed to evaluate <_FuncPtr object at 0x2adfc8d12870>
make: *** [chunks] Error 1

From shapely issues, this is related to invalid geometries:

• shapely/shapely#202
• http://nbviewer.jupyter.org/gist/sgillies/546c9250e3bd7ba33519

Will investigate further.

[maning]

Redirecting stdout to file, I got this

Error parsing feature
{'geometry': None,
 'id': '685379',
 'properties': OrderedDict([(u'OBJECTID_1', 692997), (u'OBJECTID', 688434), (u'CODE', u'Building'), (u'BLD_ID', u'2014102489030000'), (u'HEIGHT', 31.2), (u'ELEV', 387.8), (u'AREA',
 7057), (u'LARIAC_SOU', u'LARIAC4'), (u'LARIAC_DAT', u'2014'), (u'AIN', u'4404009001'), (u'STATUS', u'Modified'), (u'CODE_NUM', 1), (u'SOURCE', None), (u'DATE_', None), (u'Shape_Le
ng', 0.0), (u'Shape_Le_1', 0.0), (u'Shape_Area', 0.0), (u'GeneralUse', u'Residential'), (u'SpecificUs', u'Single Family Residence'), (u'YearBuilt', u'0'), (u'Specific_1', None), (u
'Units', u'0')]),
 'type': 'Feature'}
Exported chunks/buildings-060373104003.shp

[maning]

To check for invalid geometries, we created this script: https://github.com/osmlab/labuildings/blob/master/utils/check_shp.py

To run:
python ../utils/check_shp.py buildings.shp > bldg_err.txt

This creates a file of invalid geometries. I removed them from the data using ogr2ogr.

ogr2ogr -overwrite buildings1.shp buildings.shp -where "(BLD_ID NOT IN ('2014073321520000','2014073494560000','480907867665','2014092885640000','2014102489030000','2014102482470000','2014113083630000','527671857199','465354755246','541839844108'))"

The chunk.py is now complete.