shapely and geopandas improvements (#3)

.github/setup/action.yml

@@ -0,0 +1,16 @@
+name: Setup
+description: Setup Python and install dependencies
+
+runs:
+  using: composite
+  steps:
+    - uses: actions/checkout@v4
+    - name: Set up Python 3.10
+      uses: actions/setup-python@v5
+      with:
+        python-version: "3.10"
+        cache: pip
+        cache-dependency-path: requirements-tests.txt
+    - name: Install dependencies
+      shell: bash
+      run: pip install -r requirements-tests.txt

.github/workflows/tests.yml

@@ -12,31 +12,51 @@ on:
       - ".vscode/**"
       - ".pre-commit-config.yaml"
       - "*.md"
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+env:
+  PIP_DISABLE_PIP_VERSION_CHECK: 1
+  FORCE_COLOR: 1
 
 jobs:
-  build:
-    strategy:
-      matrix:
-        os: [ubuntu-latest]
-        python-version: ["3.10"]
-    runs-on: ${{ matrix.os }}
+  ruff:
+    name: Run ruff check and ruff format
+    runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v5
-        with:
-          python-version: ${{ matrix.python-version }}
-      - name: Install dependencies
-        run: pip install -r requirements-tests.txt
+      - uses: ./.github/setup
       - name: Lint with ruff
         run: python run.py ruff-check
       - name: Format with ruff
         run: python run.py ruff-format
-      - name: Run the test suite
-        run: python run.py pytest
-      - name: Run mypy on the tests and on the stubs
-        run: python run.py mypy
-      - name: Run pyright on the tests and on the stubs
-        run: python run.py pyright
-      - name: Run stubtest on the the stubs
-        run: python run.py stubtest
+  pytest:
+    name: Run the test suite
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ./.github/setup
+      - run: python run.py pytest
+  mypy:
+    name: Run mypy on the tests and on the stubs
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ./.github/setup
+      - run: python run.py mypy
+  stubtest:
+    name: Run stubtest on the the stubs
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ./.github/setup
+      - run: python run.py stubtest
+  pyright:
+    name: Run pyright on the tests and on the stubs
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ./.github/setup
+      - run: python run.py pyright

.gitignore

@@ -2,8 +2,11 @@
 __pycache__/
 /build/
 /dist/
-/.*_cache/
-/.?v?env/
+.*_cache/
+/.venv/
+/.env/
+/venv/
+/env/
 
 # files
 .coverage

.pre-commit-config.yaml

@@ -10,7 +10,7 @@ repos:
       - id: mixed-line-ending
       - id: check-case-conflict
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.1.14 # must match requirements-tests.txt
+    rev: v0.2.1 # must match requirements-tests.txt
     hooks:
       - id: ruff
       - id: ruff-format

requirements-tests.txt

@@ -1,12 +1,16 @@
 # Tools
-ruff==0.1.14  # must match .pre-commit-config.yaml
-pytest>=7.0
+ruff==0.2.1  # must match .pre-commit-config.yaml
+pytest>=8.0
 mypy==1.8.0
-pyright==1.1.347
+pyright==1.1.350
 
 # Runtime
 shapely>=2.0,<2.1
+geopandas>=0.14.3,<1.0
 
 # Transient dependencies
 matplotlib>=3.8.0
 pyproj>=3.6.1
+folium>=0.15.1
+rtree>=1.1.0
+pandas-stubs>=2.1.4.231227

run.py

@@ -7,9 +7,9 @@
 import sys
 
 default_args = {
-    "mypy": ["tests", "stubs/shapely-stubs"],
-    "pyright": ["tests", "stubs/shapely-stubs"],
-    "stubtest": ["--allowlist=stubtest_allowlist.txt", "shapely"],
+    "mypy": ["tests", "stubs/shapely-stubs", "stubs/geopandas-stubs"],
+    "pyright": ["tests", "stubs/shapely-stubs", "stubs/geopandas-stubs"],
+    "stubtest": ["--allowlist=stubtest_allowlist.txt", "shapely", "geopandas"],
     "ruff-check": ["tests", "stubs"],
     "ruff-format": ["tests", "stubs"],
     "pytest": [],

stubs/geopandas-stubs/.ruff.toml

@@ -1,2 +1,4 @@
 extend = "../../pyproject.toml"
+
+[lint]
 isort.known-first-party = ["geopandas"]

stubs/geopandas-stubs/array.pyi

@@ -1,27 +1,28 @@
 import builtins
 from _typeshed import Incomplete, Unused
 from collections.abc import Sequence
-from typing import Any, Literal, NoReturn, overload
+from typing import Any, ClassVar, Literal, NoReturn, SupportsIndex, overload
 from typing_extensions import Self, TypeAlias, deprecated
 
 import numpy as np
 import pandas as pd
 from numpy.typing import ArrayLike, DTypeLike, NDArray
 from pandas.api.extensions import ExtensionArray, ExtensionDtype
 from pyproj import CRS, Transformer
-from shapely import Point
+from shapely import Geometry, Point
 from shapely.geometry.base import BaseGeometry
 
 from geopandas.base import _ConvertibleToCRS
+from geopandas.sindex import PyGEOSSTRTreeIndex
 
-_ArrayOrGeom: TypeAlias = GeometryArray | ArrayLike | BaseGeometry
+_ArrayOrGeom: TypeAlias = GeometryArray | ArrayLike | Geometry
 _Origin: TypeAlias = Literal["center", "centroid"] | Point | tuple[float, float] | tuple[float, float, float]
 
 TransformerFromCRS = Transformer.from_crs
 
 class GeometryDtype(ExtensionDtype):
-    type = BaseGeometry
-    name: str
+    type: ClassVar[type[BaseGeometry]]
+    name: ClassVar[str]
     na_value: float
     @classmethod
     def construct_from_string(cls, string: str) -> Self: ...
@@ -47,19 +48,26 @@ class GeometryArray(ExtensionArray):
     @deprecated("Attribute `.data` is deprecated. Use method `to_numpy()` instead.")
     def data(self) -> NDArray[np.object_]: ...
     @property
-    def sindex(self): ...
+    def sindex(self) -> PyGEOSSTRTreeIndex: ...
     @property
     def has_sindex(self) -> bool: ...
     @property
     def crs(self) -> CRS | None: ...
     @crs.setter
-    def crs(self, value: _ConvertibleToCRS) -> None: ...
+    def crs(self, value: _ConvertibleToCRS | None) -> None: ...
     def check_geographic_crs(self, stacklevel: int) -> None: ...
     @property
     def dtype(self) -> GeometryDtype: ...
     def __len__(self) -> int: ...
-    def __getitem__(self, idx): ...
-    def __setitem__(self, key, value) -> None: ...
+    @overload
+    def __getitem__(self, idx: int | np.integer[Any]) -> BaseGeometry: ...  # Always 1-D, doesn't accept tuple
+    @overload
+    def __getitem__(self, idx: slice | Sequence[int] | NDArray[np.bool_] | NDArray[np.integer[Any]]) -> GeometryArray: ...
+    @overload
+    def __getitem__(
+        self, idx: int | np.integer[Any] | slice | Sequence[int] | NDArray[np.bool_] | NDArray[np.integer[Any]]
+    ) -> BaseGeometry | GeometryArray: ...
+    def __setitem__(self, key, value: _ArrayOrGeom | pd.DataFrame | pd.Series[Any]) -> None: ...
     @property
     def is_valid(self) -> NDArray[np.bool_]: ...
     @property
@@ -82,7 +90,7 @@ class GeometryArray(ExtensionArray):
     def boundary(self) -> GeometryArray: ...
     @property
     def centroid(self) -> GeometryArray: ...
-    def concave_hull(self, ratio: float, allow_holes: bool) -> GeometryArray: ...
+    def concave_hull(self, ratio: float, allow_holes: bool) -> NDArray[np.object_]: ...
     @property
     def convex_hull(self) -> GeometryArray: ...
     def delaunay_triangles(self, tolerance: float, only_edges: bool) -> GeometryArray: ...
@@ -134,7 +142,7 @@ class GeometryArray(ExtensionArray):
     def buffer(self, distance: float | ArrayLike, resolution: int = 16, **kwargs) -> GeometryArray: ...
     def interpolate(self, distance, normalized: bool = False) -> GeometryArray: ...
     def simplify(self, tolerance: float | ArrayLike, preserve_topology: bool = True) -> GeometryArray: ...
-    def project(self, other: _ArrayOrGeom, normalized: bool = False) -> GeometryArray: ...
+    def project(self, other: _ArrayOrGeom, normalized: bool = False) -> NDArray[np.float64]: ...
     def relate(self, other: _ArrayOrGeom) -> NDArray[np.str_]: ...
     def unary_union(self) -> BaseGeometry: ...
     def affine_transform(self, matrix: Incomplete) -> GeometryArray: ...
@@ -157,16 +165,19 @@ class GeometryArray(ExtensionArray):
     @property
     def size(self) -> int: ...
     @property
-    def shape(self) -> tuple[int]: ...
+    def shape(self) -> tuple[int]: ...  # Always 1-D, this is not a mistake
     @property
-    def ndim(self) -> int: ...
+    def ndim(self) -> Literal[1]: ...
     def copy(self, *args: Unused, **kwargs: Unused) -> GeometryArray: ...
     def take(
-        self, indices: Sequence[int] | NDArray[np.int_], allow_fill: bool = False, fill_value: BaseGeometry | None = None
+        self,
+        indices: Sequence[SupportsIndex] | NDArray[np.integer[Any]],
+        allow_fill: bool = False,
+        fill_value: Geometry | None = None,
     ) -> GeometryArray: ...
     def fillna(
         self,
-        value: BaseGeometry | GeometryArray | None = None,
+        value: Geometry | GeometryArray | None = None,
         method: Literal["backfill", "bfill", "pad", "ffill"] | None = None,
         limit: int | None = None,
         copy: bool = True,
@@ -177,7 +188,7 @@ class GeometryArray(ExtensionArray):
     def unique(self) -> GeometryArray: ...
     @property
     def nbytes(self) -> int: ...
-    def shift(self, periods: int = 1, fill_value: BaseGeometry | None = None) -> GeometryArray: ...
+    def shift(self, periods: int = 1, fill_value: Geometry | None = None) -> GeometryArray: ...  # type: ignore[override]
     def argmin(self, skipna: bool = True) -> NoReturn: ...
     def argmax(self, skipna: bool = True) -> NoReturn: ...
     def __array__(self, dtype: Unused | None = None) -> NDArray[np.object_]: ...

stubs/geopandas-stubs/base.pyi

@@ -8,6 +8,7 @@ import pandas as pd
 from numpy.random import BitGenerator, Generator as RandomGenerator, SeedSequence
 from numpy.typing import ArrayLike, NDArray
 from pyproj import CRS
+from shapely import Geometry
 from shapely.geometry.base import BaseGeometry
 
 from geopandas.array import _Origin
@@ -80,34 +81,34 @@ class GeoPandasBase:
     def cascaded_union(self) -> BaseGeometry: ...
     @property
     def unary_union(self) -> BaseGeometry: ...
-    def contains(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
-    def geom_equals(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
+    def contains(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
+    def geom_equals(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
     @deprecated("Method `geom_almost_equals` is deprecated. Use `geom_equals_exact` instead.")
-    def geom_almost_equals(self, other: GeoSeries | BaseGeometry, decimal: int = 6, align: bool = True) -> pd.Series[bool]: ...
+    def geom_almost_equals(self, other: GeoSeries | Geometry, decimal: int = 6, align: bool = True) -> pd.Series[bool]: ...
     def geom_equals_exact(
-        self, other: GeoSeries | BaseGeometry, tolerance: float | ArrayLike, align: bool = True
+        self, other: GeoSeries | Geometry, tolerance: float | ArrayLike, align: bool = True
     ) -> pd.Series[bool]: ...
-    def crosses(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
-    def disjoint(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
-    def intersects(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
-    def overlaps(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
-    def touches(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
-    def within(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
-    def covers(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
-    def covered_by(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[bool]: ...
-    def distance(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[float]: ...
+    def crosses(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
+    def disjoint(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
+    def intersects(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
+    def overlaps(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
+    def touches(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
+    def within(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
+    def covers(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
+    def covered_by(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[bool]: ...
+    def distance(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[float]: ...
     def hausdorff_distance(
-        self, other: GeoSeries | BaseGeometry, align: bool = True, densify: float | ArrayLike | None = None
+        self, other: GeoSeries | Geometry, align: bool = True, densify: float | ArrayLike | None = None
     ) -> pd.Series[float]: ...
     def frechet_distance(
-        self, other: GeoSeries | BaseGeometry, align: bool = True, densify: float | ArrayLike | None = None
+        self, other: GeoSeries | Geometry, align: bool = True, densify: float | ArrayLike | None = None
     ) -> pd.Series[float]: ...
-    def difference(self, other: GeoSeries | BaseGeometry, align: bool = True) -> GeoSeries: ...
-    def symmetric_difference(self, other: GeoSeries | BaseGeometry, align: bool = True) -> GeoSeries: ...
-    def union(self, other: GeoSeries | BaseGeometry, align: bool = True) -> GeoSeries: ...
-    def intersection(self, other: GeoSeries | BaseGeometry, align: bool = True) -> GeoSeries: ...
+    def difference(self, other: GeoSeries | Geometry, align: bool = True) -> GeoSeries: ...
+    def symmetric_difference(self, other: GeoSeries | Geometry, align: bool = True) -> GeoSeries: ...
+    def union(self, other: GeoSeries | Geometry, align: bool = True) -> GeoSeries: ...
+    def intersection(self, other: GeoSeries | Geometry, align: bool = True) -> GeoSeries: ...
     def clip_by_rect(self, xmin: float, ymin: float, xmax: float, ymax: float) -> GeoSeries: ...
-    def shortest_line(self, other: GeoSeries | BaseGeometry, align: bool = True) -> GeoSeries: ...
+    def shortest_line(self, other: GeoSeries | Geometry, align: bool = True) -> GeoSeries: ...
     @property
     def bounds(self) -> pd.DataFrame: ...
     @property
@@ -118,8 +119,8 @@ class GeoPandasBase:
     def has_sindex(self) -> bool: ...
     def buffer(self, distance: float | ArrayLike, resolution: int = 16, **kwargs) -> GeoSeries: ...
     def simplify(self, tolerance: float | ArrayLike, preserve_topology: bool = True) -> GeoSeries: ...
-    def relate(self, other: GeoSeries | BaseGeometry, align: bool = True) -> pd.Series[str]: ...
-    def project(self, other: GeoSeries | BaseGeometry, normalized: bool = False, align: bool = True) -> pd.Series[float]: ...
+    def relate(self, other: GeoSeries | Geometry, align: bool = True) -> pd.Series[str]: ...
+    def project(self, other: GeoSeries | Geometry, normalized: bool = False, align: bool = True) -> pd.Series[float]: ...
     def interpolate(self, distance: float | ArrayLike, normalized: bool = False) -> GeoSeries: ...
     def affine_transform(self, matrix: Incomplete) -> GeoSeries: ...
     def translate(self, xoff: float = 0.0, yoff: float = 0.0, zoff: float = 0.0) -> GeoSeries: ...

stubs/geopandas-stubs/explore.pyi

@@ -1,10 +1,10 @@
 from collections.abc import Callable, Hashable, MutableMapping, Sequence
 from typing import Any
 
-import branca
-import folium
+import branca  # type: ignore[import-untyped]
+import folium  # type: ignore[import-untyped]
 import pandas as pd
-import xyzservices
+import xyzservices  # type: ignore[import-untyped]
 from matplotlib.colors import Colormap
 from numpy.typing import ArrayLike, NDArray