datasets including Points breaks computing the topology

[mattijn]

datasets including Points breaks computing the topology

[mattijn]

Thanks for sharing the file. There is nothing wrong with your data. Temporary solution: if you remove the Points from your geojson file it will be parsed correctly.

Bugs that I've to solve:
Multiple things are going wrong here regarding data collections containing Points.

1. If there is a data collection that has a combination of Points and Polygons as such:

polygon_point = [
    {
        "type": "Polygon",
        "coordinates": [[[0, 0], [1, 0], [1, 1], [0, 1], [0, 0]]]
    },
    {
        "type": "Point",
        "coordinates": [0.5, 0.5]
    }
]

Than the Extract class parses the Point type incorrectly into an arcs object containing the coordinate.

import topojson as tp
from topojson.core.extract import Extract

Extract(polygon_point)

Extract(
{'bookkeeping_geoms': [[0]],
 'linestrings': [<shapely.geometry.linestring.LineString object at 0x0000000008B5B588>],
 'objects': {0: {'arcs': [0], 'type': 'Polygon'},
             1: {'arcs': [0.5, 0.5], 'type': 'Point'}},
 'options': TopoOptions(
  {'prequantize': False,
 'presimplify': False,
 'simplify_algorithm': 'dp',
 'simplify_with': 'shapely',
 'topology': True,
 'topoquantize': False,
 'toposimplify': False,
 'winding_order': None}
),
 'type': 'Topology'}
)

The coordinates of type Point or Multipoint should remain as coordinates.

Otherwise you will hit indeed the following error:

tp.Topology(polygon_point)

c:\programdata\miniconda3\lib\site-packages\topojson\core\hashmap.py in resolve_bookkeeping(self, geoms)
    311         arcs = []
    312         for geom in geoms:
--> 313             arcs_in_geom = self.data["bookkeeping_geoms"][geom]
    314             for idx_arc, arc_ref in enumerate(arcs_in_geom):
    315                 arc_ids = self.data["bookkeeping_arcs"][arc_ref]

TypeError: list indices must be integers or slices, not float

2. Moreover, if there is a data object containing only a Point as such:

point = [
    {
        "type": "Point",
        "coordinates": [0.5, 0.5]
    }
]

tp.Topology(point)

It cannot compute a bounding box since topojson currently only considers parsed linestrings:

c:\programdata\miniconda3\lib\site-packages\topojson\core\join.py in joiner(self, data)
    136 
    137             data["linestrings"], data["transform"] = quantize(
--> 138                 data["linestrings"], data["bbox"], quant_factor
    139             )
    140 

c:\programdata\miniconda3\lib\site-packages\topojson\ops.py in quantize(linestrings, bbox, quant_factor)
    403     """
    404 
--> 405     x0, y0, x1, y1 = bbox
    406 
    407     kx = 1 / ((quant_factor - 1) / (x1 - x0))

ValueError: not enough values to unpack (expected 4, got 0)

3. This also means there is currently no delta-encoding of Points within topojson.

[mattijn]

Thanks for filling this issue. Could you also share the GeoJSON that gives the errors?
Than I can have look

[theluxaz]

This is an example json feature collection with 4 features which i could not convert to any way to topojson. It is quite a complex polygon so sorry for hard to read file. Couldn't even convert each feature separately in any way.

I would be really grateful if you could show me a way how it could be acomplished, i'm sure it would be really useful for a lot of people in the future with similar problems :)

geojson show.zip

[theluxaz]

I see, Thank you very much for your detailed answer!!!
I will find a workaround for Points.

Wish you all the best.

[mattijn]

Fixed by #62