def clusters(self, maxpoints=1000000, **kwargs):
"""Computes clusters using centroids of geometries.
Parameters:
max_points (int): The maximum number of features that will be used for the cluster computation.
**kwargs: Additional optional arguments for clustering. See clustering.Clustering.
Returns:
(object) A Clustering object.
"""
if not self._has_geometry:
warnings.warn('DataFrame is not spatial.')
return None
if maxpoints is not None and maxpoints < len(self.df):
pois = self.df.sample(n=maxpoints)
else:
pois = self.df.copy()
pois.constructive.centroid(inplace=True)
geom = pois.geometry.to_pygeos().values()
filt = pg.get_coordinate_dimension(geom)
pois.add_column('tmp', filt, dtype=int)
pois = pois[pois.tmp == 2]
pois.drop('tmp', inplace=True)
pois = pois.extract()
return Clustering(pois, **kwargs)
def test_get_coordinate_dimension():
actual = pygeos.get_coordinate_dimension([point, point_z, None]).tolist()
assert actual == [2, 3, -1]
def test_apply_correct_coordinate_dimension():
# ensure that new geometry is 2D with include_z=False
geom = line_string_z
assert pygeos.get_coordinate_dimension(geom) == 3
new_geom = apply(geom, lambda x: x + 1, include_z=False)
assert pygeos.get_coordinate_dimension(new_geom) == 2