dtoolkit.geoaccessor.geoseries.filter_geometry#

dtoolkit.geoaccessor.geoseries.filter_geometry(s: gpd.GeoSeries, /, other: BaseGeometry | gpd.GeoSeries | gpd.GeoDataFrame, predicate: BINARY_PREDICATE, complement: bool = False, **kwargs) → gpd.GeoSeries[source]#

Filter GeoSeries via the spatial relationship between GeoSeries and geometry.

A sugar syntax wraps:

s[s.{predicate}(other, **kwargs)]

Parameters:

otherGeometry, GeoSeries, or GeoDataFrame
predicate{“intersects”, “disjoint”, “crosses”, “overlaps”, “touches”, “covered_by”, “contains”, “within”, “covers”}: Binary predicate.
complementbool, default False: If True, do operation reversely.
**kwargs: See the documentation for GeoSeries.{predicate} for complete details on the keyword arguments.

Returns:

GeoSeries

See also

geopandas.GeoSeries.intersects
geopandas.GeoSeries.covered_by
geopandas.GeoSeries.contains
dtoolkit.geoaccessor.geoseries.filter_geometry
dtoolkit.geoaccessor.geodataframe.filter_geometry

Examples

>>> import dtoolkit.geoaccessor
>>> import geopandas as gpd
>>> from shapely import Polygon, LineString, Point, box
>>> df = gpd.GeoDataFrame(
...     geometry=[
...         Polygon([(0, 0), (1, 1), (0, 1)]),
...         LineString([(0, 0), (0, 2)]),
...         LineString([(0, 0), (0, 1)]),
...         Point(0, 1),
...         Point(-1, -1),
...     ],
... )
>>> df
                                            geometry
0  POLYGON ((0.00000 0.00000, 1.00000 1.00000, 0....
1      LINESTRING (0.00000 0.00000, 0.00000 2.00000)
2      LINESTRING (0.00000 0.00000, 0.00000 1.00000)
3                            POINT (0.00000 1.00000)
4                          POINT (-1.00000 -1.00000)

Filter the geometries out of the bounding box(0, 0, 2, 2).

>>> df.filter_geometry(box(0, 0, 2, 2), "covered_by", complement=True)
                      geometry
4    POINT (-1.00000 -1.00000)

This method is actually faster than the following one:

def is_in_geometry(s: gpd.GeoSeries, other: BaseGeometry) -> pd.Series:
    s_bounds, g_bounds = s.bounds, other.bounds

    return (
        (s_bounds.minx >= g_bounds[0])
        & (s_bounds.miny >= g_bounds[1])
        & (s_bounds.maxx <= g_bounds[2])
        & (s_bounds.maxy <= g_bounds[3])
    )