shapely.geos:TopologyException: depth mismatch at at 552100 5277600

[bburghaus]

ERROR:shapely.geos:TopologyException: depth mismatch at  at 552100 5277600
Traceback (most recent call last):
  File "maps.py", line 18, in <module>
    layers = plot(
  File "/home/***/.local/lib/python3.8/site-packages/prettymaps/draw.py", line 189, in plot
    layers = {
  File "/home/***/.local/lib/python3.8/site-packages/prettymaps/draw.py", line 190, in <dictcomp>
    layer: get_layer(
  File "/home/***/.local/lib/python3.8/site-packages/prettymaps/fetch.py", line 146, in get_layer
    return get_streets(**kwargs, layer = layer)
  File "/home/***/.local/lib/python3.8/site-packages/prettymaps/fetch.py", line 109, in get_streets
    streets = unary_union([
  File "/home/***/.local/lib/python3.8/site-packages/prettymaps/fetch.py", line 111, in <listcomp>
    MultiLineString(
  File "/home/***/.local/lib/python3.8/site-packages/shapely/geometry/base.py", line 637, in buffer
    return geom_factory(self.impl['buffer_with_params'](self, params, distance))
  File "/home/***/.local/lib/python3.8/site-packages/shapely/geometry/base.py", line 78, in geom_factory
    raise ValueError("No Shapely geometry can be created from null value")
ValueError: No Shapely geometry can be created from null value

To reproduce:

• Use the Barcelona, Spain (square plot with rounded edges) example
• Change the plot coordinates and radius to (47.606292888654316, -122.33203340169379), radius = 5000
• Run script

[jacopofar]

Same with Milan, Italy and radius=4000, but not with radius=3000.

As a workaround, I modified the code to catch the exception, log and continue, but cannot find the root cause.

This is my change, at line 310 of fetch.py:

        multiline_strings = []
        for highway, w in width.items():
            try:
                multiline_strings.append(
                    MultiLineString(
                        streets[
                            [highway in value for value in streets[layer]]
                            & (streets.geometry.type == "LineString")
                        ].geometry.tolist()
                        + list(
                            reduce(
                                lambda x, y: x + y,
                                [
                                    list(lines)
                                    for lines in streets[
                                        [highway in value for value in streets[layer]]
                                        & (streets.geometry.type == "MultiLineString")
                                    ].geometry
                                ],
                                [],
                            )
                        )
                    ).buffer(w)
                )
            except ValueError as e:
                print('Error', e)
                print('Offending object:', highway, w)

        streets = unary_union(multiline_strings)

I simply moved the list comprehension to a normal for where I can use a try...except and print the error and keep running.
In my case it was:

Error No Shapely geometry can be created from null value
Offending object: footway 1

and then the map was generated

[smilebasti]

I also came to this error.

I tried your solution @jacopofar.
It shows me this:

Error No Shapely geometry can be created from null value
Offending object: footway 1

but then the error came, after that, again.

Any thoughts on that?

[jacopofar]

My workaround simply skips the geometry causing the issue, the fact you are seeing that message means the exception was handled. Do you mean that after that you still got a stacktrace and the script was terminated? If so it must be a different error, you can paste the output

[smilebasti]

Washington (1).pdf

It's somehow the same Error

[jacopofar]

It's easier if you paste the error as text rather than a PDF :)

I see the error is in a different piece of code, the one that is dilating the streets. It's a bit hard to fix because with the reduce and list comprehension I cannot tell where exactly the error is coming from. You can try to wrap the whole else at lines 341-355 in a try..catch to ignore the offending data and see whether the result looks good