ssb-sgis 1.0.3__py3-none-any.whl → 1.0.5__py3-none-any.whl

sgis/maps/maps.py

@@ -19,6 +19,7 @@ from shapely import Geometry
 from shapely import box
 from shapely.geometry import Polygon
 
+from ..debug_config import _NoExplore
 from ..geopandas_tools.bounds import get_total_bounds
 from ..geopandas_tools.conversion import to_bbox
 from ..geopandas_tools.conversion import to_gdf
@@ -84,7 +85,8 @@ def explore(
     browser: bool = False,
     smooth_factor: int | float = 1.5,
     size: int | None = None,
-    max_images: int = 15,
+    max_images: int = 10,
+    images_to_gdf: bool = False,
     **kwargs,
 ) -> Explore:
     """Interactive map of GeoDataFrames with layers that can be toggled on/off.
@@ -113,7 +115,9 @@ def explore(
         size: The buffer distance. Only used when center is given. It then defaults to
             1000.
         max_images: Maximum number of images (Image, ImageCollection, Band) to show per
-            map. Defaults to 15.
+            map. Defaults to 10.
+        images_to_gdf: If True (not default), images (Image, ImageCollection, Band)
+            will be converted to GeoDataFrame and added to the map.
         **kwargs: Keyword arguments to pass to geopandas.GeoDataFrame.explore, for
             instance 'cmap' to change the colors, 'scheme' to change how the data
             is grouped. This defaults to 'fisherjenkssampled' for numeric data.
@@ -143,6 +147,9 @@ def explore(
     >>> points["meters"] = points.length
     >>> sg.explore(roads, points, column="meters", cmap="plasma", max_zoom=60, center_4326=(10.7463, 59.92, 500))
     """
+    if isinstance(center, _NoExplore):
+        return
+
     gdfs, column, kwargs = Map._separate_args(gdfs, column, kwargs)
 
     loc_mask, kwargs = _get_location_mask(kwargs | {"size": size}, gdfs)
@@ -186,13 +193,63 @@ def explore(
             mask = to_gdf(center, crs=from_crs)
 
         bounds: Polygon = box(*get_total_bounds(*gdfs, *list(kwargs.values())))
-        if not mask.intersects(bounds).any():
-            mask = mask.set_crs(4326, allow_override=True)
 
-        try:
-            mask = mask.to_crs(to_crs)
-        except ValueError:
-            pass
+        any_intersections: bool = mask.intersects(bounds).any()
+        if not any_intersections and to_crs is None:
+            mask = to_gdf(Polygon(), to_crs)
+        elif not any_intersections:
+            bounds4326 = to_gdf(bounds, to_crs).to_crs(25833).geometry.iloc[0]
+            mask4326 = mask.set_crs(4326, allow_override=True).to_crs(25833)
+
+            if (mask4326.distance(bounds4326) > size).all():
+                # try flipping coordinates
+                x, y = next(iter(mask.geometry.iloc[0].coords))
+                mask4326 = to_gdf([y, x], 4326).to_crs(25833)
+
+            if (mask4326.distance(bounds4326) > size).all():
+                mask = to_gdf(Polygon(), to_crs)
+            else:
+                mask = mask4326.to_crs(to_crs)
+
+        # else:
+        #     mask_flipped = mask
+
+        # # coords = mask.get_coordinates()
+        # if (
+        #     (mask_flipped.distance(bounds) > size).all()
+        #     # and coords["x"].max() < 180
+        #     # and coords["y"].max() < 180
+        #     # and coords["x"].min() > -180
+        #     # and coords["y"].min() > -180
+        # ):
+        #     try:
+        #         bounds4326 = to_gdf(bounds, to_crs).to_crs(4326).geometry.iloc[0]
+        #     except ValueError:
+        #         bounds4326 = to_gdf(bounds, to_crs).set_crs(4326).geometry.iloc[0]
+
+        #     mask4326 = mask.set_crs(4326, allow_override=True)
+
+        #     if (mask4326.distance(bounds4326) > size).all():
+        #         # try flipping coordinates
+        #         x, y = list(mask4326.geometry.iloc[0].coords)[0]
+        #         mask4326 = to_gdf([y, x], 4326)
+
+        #     mask = mask4326
+
+        #     # if mask4326.intersects(bounds4326).any():
+        #     #     mask = mask4326
+        #     # else:
+        #     #     try:
+        #     #         mask = mask.to_crs(to_crs)
+        #     #     except ValueError:
+        #     #         pass
+        # else:
+        #     mask = mask_flipped
+
+        # try:
+        #     mask = mask.to_crs(to_crs)
+        # except ValueError:
+        #     pass
 
         if get_geom_type(mask) in ["point", "line"]:
             mask = mask.buffer(size)
@@ -236,7 +293,7 @@ def samplemap(
     smooth_factor: int = 1.5,
     explore: bool = True,
     browser: bool = False,
-    max_images: int = 15,
+    max_images: int = 10,
     **kwargs,
 ) -> Explore:
     """Shows an interactive map of a random area of GeoDataFrames.
@@ -269,7 +326,7 @@ def samplemap(
         browser: If False (default), the maps will be shown in Jupyter.
             If True the maps will be opened in a browser folder.
         max_images: Maximum number of images (Image, ImageCollection, Band) to show per
-            map. Defaults to 15.
+            map. Defaults to 10.
         **kwargs: Keyword arguments to pass to geopandas.GeoDataFrame.explore, for
             instance 'cmap' to change the colors, 'scheme' to change how the data
             is grouped. This defaults to 'fisherjenkssampled' for numeric data.
@@ -294,6 +351,9 @@ def samplemap(
     >>> samplemap(roads, points, size=5_000, column="meters")
 
     """
+    if isinstance(kwargs.get("center", None), _NoExplore):
+        return
+
     if gdfs and len(gdfs) > 1 and isinstance(gdfs[-1], (float, int)):
         *gdfs, size = gdfs
 
@@ -311,7 +371,7 @@ def samplemap(
 
     if mask is None:
         try:
-            sample = sample.geometry.dropna().sample(1)
+            sample = sample.geometry.loc[lambda x: ~x.is_empty].sample(1)
         except Exception:
             try:
                 sample = sample.sample(1)
@@ -327,11 +387,16 @@ def samplemap(
         except Exception:
             sample = to_gdf(to_shapely(to_bbox(sample))).explode(ignore_index=True)
 
-        sample = sample.clip(mask).sample(1)
+        sample = sample.clip(mask).explode(ignore_index=True).sample(1)
 
+    print(locals())
     random_point = sample.sample_points(size=1)
 
-    center = (random_point.geometry.iloc[0].x, random_point.geometry.iloc[0].y)
+    try:
+        center = (random_point.geometry.iloc[0].x, random_point.geometry.iloc[0].y)
+    except AttributeError as e:
+        raise AttributeError(e, random_point.geometry.iloc[0]) from e
+
     print(f"center={center}, size={size}")
 
     mask = random_point.buffer(size)
@@ -357,7 +422,7 @@ def clipmap(
     max_zoom: int = 40,
     smooth_factor: int | float = 1.5,
     browser: bool = False,
-    max_images: int = 15,
+    max_images: int = 10,
     **kwargs,
 ) -> Explore | Map:
     """Shows an interactive map of a of GeoDataFrames clipped to the mask extent.
@@ -385,7 +450,7 @@ def clipmap(
         browser: If False (default), the maps will be shown in Jupyter.
             If True the maps will be opened in a browser folder.
         max_images: Maximum number of images (Image, ImageCollection, Band) to show per
-            map. Defaults to 15.
+            map. Defaults to 10.
         **kwargs: Keyword arguments to pass to geopandas.GeoDataFrame.explore, for
             instance 'cmap' to change the colors, 'scheme' to change how the data
             is grouped. This defaults to 'fisherjenkssampled' for numeric data.
@@ -395,8 +460,10 @@ def clipmap(
     explore: same functionality, but shows the entire area of the geometries.
     samplemap: same functionality, but shows only a random area of a given size.
     """
-    gdfs, column, kwargs = Map._separate_args(gdfs, column, kwargs)
+    if isinstance(kwargs.get("center", None), _NoExplore):
+        return
 
+    gdfs, column, kwargs = Map._separate_args(gdfs, column, kwargs)
     if mask is None and len(gdfs) > 1:
         mask = gdfs[-1]
         gdfs = gdfs[:-1]
@@ -450,7 +517,9 @@ def clipmap(
         return m
 
 
-def explore_locals(*gdfs: GeoDataFrame, convert: bool = True, **kwargs) -> None:
+def explore_locals(
+    *gdfs: GeoDataFrame, convert: bool = True, crs: Any | None = None, **kwargs
+) -> None:
     """Displays all local variables with geometries (GeoDataFrame etc.).
 
     Local means inside a function or file/notebook.
@@ -459,8 +528,11 @@ def explore_locals(*gdfs: GeoDataFrame, convert: bool = True, **kwargs) -> None:
         *gdfs: Additional GeoDataFrames.
         convert: If True (default), non-GeoDataFrames will be converted
             to GeoDataFrames if possible.
+        crs: Optional crs if no objects have any crs.
         **kwargs: keyword arguments passed to sg.explore.
     """
+    if isinstance(kwargs.get("center", None), _NoExplore):
+        return
 
     def as_dict(obj):
         if hasattr(obj, "__dict__"):
@@ -482,11 +554,19 @@ def explore_locals(*gdfs: GeoDataFrame, convert: bool = True, **kwargs) -> None:
             if not convert:
                 continue
 
+            try:
+                gdf = clean_geoms(to_gdf(value, crs=crs))
+                if len(gdf):
+                    local_gdfs[name] = gdf
+                continue
+            except Exception:
+                pass
+
             if isinstance(value, dict) or hasattr(value, "__dict__"):
                 # add dicts or classes with GeoDataFrames to kwargs
                 for key, val in as_dict(value).items():
                     if isinstance(val, allowed_types):
-                        gdf = clean_geoms(to_gdf(val))
+                        gdf = clean_geoms(to_gdf(val, crs=crs))
                         if len(gdf):
                             local_gdfs[key] = gdf
 
@@ -494,22 +574,13 @@ def explore_locals(*gdfs: GeoDataFrame, convert: bool = True, **kwargs) -> None:
                         try:
                             for k, v in val.items():
                                 if isinstance(v, allowed_types):
-                                    gdf = clean_geoms(to_gdf(v))
+                                    gdf = clean_geoms(to_gdf(v, crs=crs))
                                     if len(gdf):
                                         local_gdfs[k] = gdf
                         except Exception:
                             # no need to raise here
                             pass
 
-                continue
-            try:
-                gdf = clean_geoms(to_gdf(value))
-                if len(gdf):
-                    local_gdfs[name] = gdf
-                continue
-            except Exception:
-                pass
-
         if local_gdfs:
             break
 
@@ -518,7 +589,7 @@ def explore_locals(*gdfs: GeoDataFrame, convert: bool = True, **kwargs) -> None:
         if not frame:
             break
 
-    return explore(*gdfs, **local_gdfs, **kwargs)
+    return explore(*gdfs, **(local_gdfs | kwargs))
 
 
 def qtm(

sgis/maps/thematicmap.py

@@ -68,24 +68,24 @@ class ThematicMap(Map):
         legend_kwargs: dictionary with attributes for the legend. E.g.:
             title: Legend title. Defaults to the column name.
             rounding: If positive number, it will round floats to n decimals.
-                If negative, eg. -2, the number 3429 is rounded to 3400.
-                By default, the rounding depends on the column's maximum value
-                and standard deviation.
+            If negative, eg. -2, the number 3429 is rounded to 3400.
+            By default, the rounding depends on the column's maximum value
+            and standard deviation.
             position: The legend's x and y position in the plot. By default, it's
-                decided dynamically by finding the space with most distance to
-                the geometries. To be specified as a tuple of
-                x and y position between 0 and 1. E.g. position=(0.8, 0.2) for a position
-                in the bottom right corner, (0.2, 0.8) for the upper left corner.
+            decided dynamically by finding the space with most distance to
+            the geometries. To be specified as a tuple of
+            x and y position between 0 and 1. E.g. position=(0.8, 0.2) for a position
+            in the bottom right corner, (0.2, 0.8) for the upper left corner.
             pretty_labels: Whether to capitalize words in text categories.
             label_suffix: For numeric columns. The text to put after each number
-                in the legend labels. Defaults to None.
+            in the legend labels. Defaults to None.
             label_sep: For numeric columns. Text to put in between the two numbers
-                in each color group in the legend. Defaults to '-'.
+            in each color group in the legend. Defaults to '-'.
             thousand_sep: For numeric columns. Separator between each thousand for
-                large numbers. Defaults to None, meaning no separator.
+            large numbers. Defaults to None, meaning no separator.
             decimal_mark: For numeric columns. Text to use as decimal point.
-                Defaults to None, meaning '.' (dot) unless 'thousand_sep' is
-                '.'. In this case, ',' (comma) will be used as decimal mark.
+            Defaults to None, meaning '.' (dot) unless 'thousand_sep' is
+            '.'. In this case, ',' (comma) will be used as decimal mark.
         **kwargs: Additional attributes for the map. E.g.:
             title_color (str): Color of the title font.
             title_fontsize (int): Color of the title font.
@@ -96,7 +96,7 @@ class ThematicMap(Map):
             nan_color: Color for missing data.
 
     Examples:
-    --------
+    ---------
     >>> import sgis as sg
     >>> points = sg.random_points(100, loc=1000).pipe(sg.buff, np.random.rand(100) * 100)
     >>> points2 = sg.random_points(100, loc=1000).pipe(sg.buff, np.random.rand(100) * 100)
@@ -166,6 +166,7 @@ class ThematicMap(Map):
         nan_label: str = "Missing",
         legend_kwargs: dict | None = None,
         title_kwargs: dict | None = None,
+        legend: bool = False,
         **kwargs,
     ) -> None:
         """Initialiser."""
@@ -178,6 +179,9 @@ class ThematicMap(Map):
             nan_label=nan_label,
         )
 
+        if not legend:
+            self.legend = None
+
         self.title = title
         self._size = size
         self._dark = dark
@@ -239,10 +243,11 @@ class ThematicMap(Map):
                 raise TypeError(
                     f"{self.__class__.__name__} legend_kwargs got an unexpected key {key}"
                 )
-            try:
-                setattr(self.legend, key, value)
-            except Exception:
-                setattr(self.legend, f"_{key}", value)
+            if self.legend is not None:
+                try:
+                    setattr(self.legend, key, value)
+                except Exception:
+                    setattr(self.legend, f"_{key}", value)
 
     @property
     def valid_keywords(self) -> set[str]:
@@ -398,7 +403,7 @@ class ThematicMap(Map):
             return kwargs
 
         else:
-            if self.legend.rounding and self.legend.rounding < 0:
+            if self.legend and self.legend.rounding and self.legend.rounding < 0:
                 self.bins = prettify_bins(self.bins, self.legend.rounding)
                 self.bins = list({round(bin_, 5) for bin_ in self.bins})
                 self.bins.sort()
@@ -463,6 +468,8 @@ class ThematicMap(Map):
 
     def _create_legend(self) -> None:
         """Instantiate the Legend class."""
+        if self.legend is None:
+            return
         kwargs = {}
         if self._dark:
             kwargs["facecolor"] = "#0f0f0f"

sgis/networkanalysis/_service_area.py

@@ -5,6 +5,7 @@ from igraph import Graph
 from shapely import force_2d
 from shapely import reverse
 from shapely import unary_union
+from shapely import union_all
 from shapely.geometry import MultiPoint
 from shapely.geometry import Point
 from shapely.ops import nearest_points
@@ -113,7 +114,11 @@ def _service_area(
                 else:
                     snapped_origin: Point = nearest_points(
                         nodes_union,
-                        origins.loc[origins["temp_idx"] == idx, "geometry"].unary_union,
+                        union_all(
+                            origins.loc[
+                                origins["temp_idx"] == idx, "geometry"
+                            ].geometry.values
+                        ),
                     )[0]
 
                     within = sfilter(within, snapped_origin.buffer(0.01))

sgis/networkanalysis/cutting_lines.py

@@ -77,7 +77,7 @@ def split_lines_by_nearest_point(
     """
     PRECISION = 1e-6
 
-    if not len(gdf):
+    if not len(gdf) or not len(points):
         return gdf
 
     if (points.crs is not None and gdf.crs is not None) and not points.crs.equals(
@@ -86,10 +86,14 @@ def split_lines_by_nearest_point(
         raise ValueError("crs mismatch:", points.crs, "and", gdf.crs)
 
     if get_geom_type(gdf) != "line":
-        raise ValueError("'gdf' should only have line geometries.", gdf.geom_type)
+        raise ValueError(
+            f"'gdf' should only have line geometriess. Got {gdf.geom_type.value_counts()}"
+        )
 
     if get_geom_type(points) != "point":
-        raise ValueError("'points' should only have point geometries.")
+        raise ValueError(
+            f"'points' should only have point geometries. Got {points.geom_type.value_counts()}"
+        )
 
     gdf = gdf.copy()
 
@@ -230,9 +234,12 @@ def _change_line_endpoint(
         .values
     )
 
-    relevant_lines_mapped = relevant_lines.groupby(level=0)["geometry"].agg(LineString)
+    is_line = relevant_lines.groupby(level=0).size() > 1
+    relevant_lines_mapped = (
+        relevant_lines.loc[is_line].groupby(level=0)["geometry"].agg(LineString)
+    )
 
-    gdf.loc[is_relevant, "geometry"] = relevant_lines_mapped
+    gdf.loc[relevant_lines_mapped.index, "geometry"] = relevant_lines_mapped
 
     return gdf
 

sgis/networkanalysis/finding_isolated_networks.py

@@ -1,6 +1,7 @@
 """Functions for Finding network components in a GeoDataFrame of lines."""
 
 import networkx as nx
+import pandas as pd
 from geopandas import GeoDataFrame
 
 from .nodes import make_node_ids
@@ -81,7 +82,7 @@ def get_component_size(gdf: GeoDataFrame) -> GeoDataFrame:
     >>> roads = read_parquet_url("https://media.githubusercontent.com/media/statisticsnorway/ssb-sgis/main/tests/testdata/roads_oslo_2022.parquet")
 
     >>> roads = get_component_size(roads)
-    >>> roads.component_size.value_counts().head()
+    >>> roads["component_size"].value_counts().head()
     component_size
     79180    85638
     2         1601
@@ -101,11 +102,17 @@ def get_component_size(gdf: GeoDataFrame) -> GeoDataFrame:
     graph.add_edges_from(edges)
     components = [list(x) for x in nx.connected_components(graph)]
 
-    componentsdict = {
-        idx: len(component) for component in components for idx in component
-    }
-
-    gdf["component_size"] = gdf.source.map(componentsdict)
+    mapper = pd.DataFrame(
+        {
+            idx: [i, len(component)]
+            for i, component in enumerate(components)
+            for idx in component
+        },
+    ).transpose()
+    mapper.columns = ["component_index", "component_size"]
+
+    gdf["component_index"] = gdf["source"].map(mapper["component_index"])
+    gdf["component_size"] = gdf["source"].map(mapper["component_size"])
 
     gdf = gdf.drop(
         ["source_wkt", "target_wkt", "source", "target", "n_source", "n_target"], axis=1

sgis/networkanalysis/networkanalysis.py

@@ -1027,7 +1027,7 @@ class NetworkAnalysis:
 
             if dissolve:
                 results = results.dissolve(by=["origin", self.rules.weight]).loc[
-                    :, ["geometry"]
+                    :, [results.geometry.name]
                 ]
 
             results = results.reset_index()
@@ -1038,7 +1038,7 @@ class NetworkAnalysis:
             ].rename(columns={"temp_idx": "origin"})[["origin"]]
 
             if len(missing):
-                missing["geometry"] = pd.NA
+                missing[results.geometry.name] = pd.NA
                 results = pd.concat([results, missing], ignore_index=True)
 
             results["origin"] = results["origin"].map(self.origins.idx_dict)
@@ -1151,7 +1151,7 @@ class NetworkAnalysis:
         if not all(results.geometry.isna()):
             if dissolve:
                 results = results.dissolve(by=["origin", self.rules.weight]).loc[
-                    :, ["geometry"]
+                    :, [results.geometry.name]
                 ]
             else:
                 results = results.dissolve(
@@ -1166,7 +1166,7 @@ class NetworkAnalysis:
             ].rename(columns={"temp_idx": "origin"})[["origin"]]
 
             if len(missing):
-                missing["geometry"] = pd.NA
+                missing[results.geometry.name] = pd.NA
                 results = pd.concat([results, missing], ignore_index=True)
 
             results["origin"] = results["origin"].map(self.origins.idx_dict)
@@ -1329,7 +1329,7 @@ class NetworkAnalysis:
                 df["cost_std"] = results[self.rules.weight].std()
 
         if fun == "service_area":
-            df["percent_missing"] = results["geometry"].isna().mean() * 100
+            df["percent_missing"] = results[results.geometry.name].isna().mean() * 100
         else:
             df["destinations_count"] = len(self.destinations.gdf)
 
@@ -1456,7 +1456,7 @@ class NetworkAnalysis:
         else:
             points = self.origins.gdf
 
-        points = points.drop_duplicates("geometry")
+        points = points.drop_duplicates(points.geometry.name)
 
         self.network.gdf["meters_"] = self.network.gdf.length
 
@@ -1573,7 +1573,7 @@ class NetworkAnalysis:
         This method is best stored in the NetworkAnalysis class,
         since the point classes are instantiated each time an analysis is run.
         """
-        if self.wkts[what] != [geom.wkt for geom in points.geometry]:
+        if not np.array_equal(self.wkts[what], points.geometry.to_wkt().values):
             return True
 
         if not all(x in self.graph.vs["name"] for x in list(points.temp_idx.values)):
@@ -1590,17 +1590,15 @@ class NetworkAnalysis:
         """
         self.wkts = {}
 
-        self.wkts["network"] = [geom.wkt for geom in self.network.gdf.geometry]
+        self.wkts["network"] = self.network.gdf.geometry.to_wkt().values
 
         if not hasattr(self, "origins"):
             return
 
-        self.wkts["origins"] = [geom.wkt for geom in self.origins.gdf.geometry]
+        self.wkts["origins"] = self.origins.gdf.geometry.to_wkt().values
 
         if self.destinations is not None:
-            self.wkts["destinations"] = [
-                geom.wkt for geom in self.destinations.gdf.geometry
-            ]
+            self.wkts["destinations"] = self.destinations.gdf.geometry.to_wkt().values
 
     @staticmethod
     def _sort_breaks(breaks: str | list | tuple | int | float) -> list[float | int]:

sgis/parallel/parallel.py

@@ -32,7 +32,6 @@ try:
     from ..io.dapla_functions import read_geopandas
     from ..io.dapla_functions import write_geopandas
 
-    # from ..io.write_municipality_data import write_municipality_data
 except ImportError:
     pass
 
@@ -40,11 +39,8 @@ except ImportError:
 try:
     from dapla import read_pandas
     from dapla import write_pandas
-    from dapla.gcs import GCSFileSystem
 except ImportError:
-
-    class GCSFileSystem:
-        """Placeholder."""
+    pass
 
 
 class Parallel:
@@ -806,7 +802,11 @@ def _write_one_muni(
 
     if not len(gdf_muni):
         if write_empty:
-            gdf_muni = gdf_muni.drop(columns="geometry", errors="ignore")
+            try:
+                geom_col = gdf.geometry.name
+            except AttributeError:
+                geom_col = "geometry"
+            gdf_muni = gdf_muni.drop(columns=geom_col, errors="ignore")
             gdf_muni["geometry"] = None
             write_pandas(gdf_muni, out)
         return
@@ -834,7 +834,11 @@ def _write_one_muni_with_neighbors(
 
     if not len(gdf_neighbor):
         if write_empty:
-            gdf_neighbor = gdf_neighbor.drop(columns="geometry", errors="ignore")
+            try:
+                geom_col = gdf.geometry.name
+            except AttributeError:
+                geom_col = "geometry"
+            gdf_neighbor = gdf_neighbor.drop(columns=geom_col, errors="ignore")
             gdf_neighbor["geometry"] = None
             write_pandas(gdf_neighbor, out)
         return
@@ -880,7 +884,9 @@ def _fix_missing_muni_numbers(
         )
 
     try:
-        municipalities = municipalities[[muni_number_col, "geometry"]].to_crs(gdf.crs)
+        municipalities = municipalities[
+            [muni_number_col, municipalities.geometry.name]
+        ].to_crs(gdf.crs)
     except Exception as e:
         raise e.__class__(e, to_print) from e
 
@@ -967,10 +973,21 @@ def parallel_overlay(
 
 
 def _clean_intersection(
-    df1: GeoDataFrame, df2: GeoDataFrame, to_print: str = ""
+    df1: GeoDataFrame, df2: GeoDataFrame, to_print: str | None = None
 ) -> GeoDataFrame:
     print(to_print, "- intersection chunk len:", len(df1))
-    return clean_overlay(df1, df2, how="intersection")
+    cols_to_keep = df1.columns.union(df2.columns.difference({df2.geometry.name}))
+    df1["_range_idx"] = range(len(df1))
+    joined = df1.sjoin(df2, predicate="within", how="left")
+    within = joined.loc[joined["_range_idx"].notna(), cols_to_keep]
+    not_within = joined.loc[joined["_range_idx"].isna(), df1.columns]
+    return pd.concat(
+        [
+            within,
+            clean_overlay(not_within, df2, how="intersection"),
+        ],
+        ignore_index=True,
+    )
 
 
 def chunkwise(