fix: Enable build against geoarrow-c 0.2 (#64)

Author: paleolimbot

geoarrow-pandas/src/geoarrow/pandas/lib.py

@@ -507,18 +507,18 @@ def bounds(self):
         """See :func:`geoarrow.pyarrow.box`"""
         array_or_chunked = _ga.box(self._obj)
         if isinstance(array_or_chunked, _pa.ChunkedArray):
-            flattened = [chunk.flatten() for chunk in array_or_chunked.chunks]
+            flattened = [chunk.storage.flatten() for chunk in array_or_chunked.chunks]
             seriesish = [
                 _pa.chunked_array(item, _pa.float64()) for item in zip(*flattened)
             ]
         else:
-            seriesish = array_or_chunked.flatten()
+            seriesish = array_or_chunked.storage.flatten()
 
         return _pd.DataFrame(
             {
                 "xmin": seriesish[0],
-                "xmax": seriesish[1],
-                "ymin": seriesish[2],
+                "xmax": seriesish[2],
+                "ymin": seriesish[1],
                 "ymax": seriesish[3],
             },
             index=self._obj.index,

geoarrow-pyarrow/src/geoarrow/pyarrow/_array.py

@@ -78,49 +78,18 @@ def __repr__(self):
         return f"{type_name}:{repr(self.type)}[{len(self)}]\n{items_str}".strip()
 
 
-class PointArray(GeometryExtensionArray):
-    pass
-
-
-class LinestringArray(GeometryExtensionArray):
-    pass
-
-
-class PolygonArray(GeometryExtensionArray):
-    pass
-
-
-class MultiPointArray(GeometryExtensionArray):
-    pass
-
-
-class MultiLinestringArray(GeometryExtensionArray):
-    pass
-
-
-class MultiPolygonArray(GeometryExtensionArray):
-    pass
+class BoxArray(GeometryExtensionArray):
+    def __repr__(self):
+        type_name = type(self).__name__
+        items_str = "\n".join(repr(item.bounds) for item in self)
+        return f"{type_name}:{repr(self.type)}[{len(self)}]\n{items_str}".strip()
 
 
 def array_cls_from_name(name):
-    if name == "geoarrow.wkb":
-        return GeometryExtensionArray
-    elif name == "geoarrow.wkt":
-        return GeometryExtensionArray
-    elif name == "geoarrow.point":
-        return PointArray
-    elif name == "geoarrow.linestring":
-        return LinestringArray
-    elif name == "geoarrow.polygon":
-        return PolygonArray
-    elif name == "geoarrow.multipoint":
-        return MultiPointArray
-    elif name == "geoarrow.multilinestring":
-        return MultiLinestringArray
-    elif name == "geoarrow.multipolygon":
-        return MultiPolygonArray
+    if name == "geoarrow.box":
+        return BoxArray
     else:
-        raise ValueError(f'Expected valid extension name but got "{name}"')
+        return GeometryExtensionArray
 
 
 # Inject array_cls_from_name exactly once to avoid circular import
@@ -142,7 +111,7 @@ def array(obj, type_=None, *args, **kwargs) -> GeometryExtensionArray:
     GeometryExtensionArray:WktType(geoarrow.wkt)[1]
     <POINT (0 1)>
     >>> ga.as_geoarrow(["POINT (0 1)"])
-    PointArray:PointType(geoarrow.point)[1]
+    GeometryExtensionArray:PointType(geoarrow.point)[1]
     <POINT (0 1)>
     """
     # Convert GeoPandas to WKB

geoarrow-pyarrow/src/geoarrow/pyarrow/_compute.py

@@ -251,7 +251,7 @@ def as_geoarrow(obj, type=None, coord_type=None, promote_multi=False):
 
     >>> import geoarrow.pyarrow as ga
     >>> ga.as_geoarrow(["POINT (0 1)", "MULTIPOINT Z (0 1 2, 4 5 6)"])
-    MultiPointArray:MultiPointType(geoarrow.multipoint_z)[2]
+    GeometryExtensionArray:MultiPointType(geoarrow.multipoint_z)[2]
     <MULTIPOINT Z (0 1 nan)>
     <MULTIPOINT Z (0 1 2, 4 5 6)>
     """
@@ -307,7 +307,7 @@ def make_point(x, y, z=None, m=None, crs=None):
 
     >>> import geoarrow.pyarrow as ga
     >>> ga.make_point([1, 2, 3], [4, 5, 6])
-    PointArray:PointType(geoarrow.point)[3]
+    GeometryExtensionArray:PointType(geoarrow.point)[3]
     <POINT (1 4)>
     <POINT (2 5)>
     <POINT (3 6)>
@@ -338,10 +338,11 @@ def make_point(x, y, z=None, m=None, crs=None):
 
 def _box_point_struct(storage):
     arrays = storage.flatten()
-    return pa.StructArray.from_arrays(
-        [arrays[0], arrays[0], arrays[1], arrays[1]],
-        names=["xmin", "xmax", "ymin", "ymax"],
+    box_storage = pa.StructArray.from_arrays(
+        [arrays[0], arrays[1], arrays[0], arrays[1]],
+        names=["xmin", "ymin", "xmax", "ymax"],
     )
+    return _type.types.box().to_pyarrow().wrap_array(box_storage)
 
 
 def box(obj):
@@ -350,7 +351,7 @@ def box(obj):
 
     >>> import geoarrow.pyarrow as ga
     >>> ga.box(["LINESTRING (0 10, 34 -1)"]).type
-    StructType(struct<xmin: double, xmax: double, ymin: double, ymax: double>)
+    BoxType(geoarrow.box)
     >>> print(str(ga.box(["LINESTRING (0 10, 34 -1)"])))
     -- is_valid: all not null
     -- child 0 type: double
@@ -359,11 +360,11 @@ def box(obj):
       ]
     -- child 1 type: double
       [
-        34
+        -1
       ]
     -- child 2 type: double
       [
-        -1
+        34
       ]
     -- child 3 type: double
       [
@@ -399,15 +400,15 @@ def _box_agg_point_struct(arrays):
     out = [list(pc.min_max(array).values()) for array in arrays]
     out_dict = {
         "xmin": out[0][0].as_py(),
-        "xmax": out[0][1].as_py(),
         "ymin": out[1][0].as_py(),
+        "xmax": out[0][1].as_py(),
         "ymax": out[1][1].as_py(),
     }
 
     # Apparently pyarrow reorders dict keys when inferring scalar types?
-    return pa.scalar(
-        out_dict, pa.struct([(nm, pa.float64()) for nm in out_dict.keys()])
-    )
+    storage_type = pa.struct([(nm, pa.float64()) for nm in out_dict.keys()])
+    storage_array = pa.array([out_dict], storage_type)
+    return _type.types.box().to_pyarrow().wrap_array(storage_array)[0]
 
 
 def box_agg(obj):
@@ -417,7 +418,7 @@ def box_agg(obj):
 
     >>> import geoarrow.pyarrow as ga
     >>> ga.box_agg(["POINT (0 10)", "POINT (34 -1)"])
-    <pyarrow.StructScalar: [('xmin', 0.0), ('xmax', 34.0), ('ymin', -1.0), ('ymax', 10.0)]>
+    BoxScalar({'xmin': 0.0, 'ymin': -1.0, 'xmax': 34.0, 'ymax': 10.0})
     """
 
     obj = obj_as_array_or_chunked(obj)
@@ -495,7 +496,7 @@ def with_coord_type(obj, coord_type):
 
     >>> import geoarrow.pyarrow as ga
     >>> ga.with_coord_type(["POINT (0 1)"], ga.CoordType.INTERLEAVED)
-    PointArray:PointType(interleaved geoarrow.point)[1]
+    GeometryExtensionArray:PointType(interleaved geoarrow.point)[1]
     <POINT (0 1)>
     """
     return as_geoarrow(obj, coord_type=coord_type)
@@ -537,10 +538,10 @@ def with_dimensions(obj, dimensions):
 
     >>> import geoarrow.pyarrow as ga
     >>> ga.with_dimensions(["POINT (0 1)"], ga.Dimensions.XYZM)
-    PointArray:PointType(geoarrow.point_zm)[1]
+    GeometryExtensionArray:PointType(geoarrow.point_zm)[1]
     <POINT ZM (0 1 nan nan)>
     >>> ga.with_dimensions(["POINT ZM (0 1 2 3)"], ga.Dimensions.XY)
-    PointArray:PointType(geoarrow.point)[1]
+    GeometryExtensionArray:PointType(geoarrow.point)[1]
     <POINT (0 1)>
     """
     obj = as_geoarrow(obj)
@@ -557,13 +558,13 @@ def with_geometry_type(obj, geometry_type):
 
     >>> import geoarrow.pyarrow as ga
     >>> ga.with_geometry_type(["POINT (0 1)"], ga.GeometryType.MULTIPOINT)
-    MultiPointArray:MultiPointType(geoarrow.multipoint)[1]
+    GeometryExtensionArray:MultiPointType(geoarrow.multipoint)[1]
     <MULTIPOINT (0 1)>
     >>> ga.with_geometry_type(["MULTIPOINT (0 1)"], ga.GeometryType.POINT)
-    PointArray:PointType(geoarrow.point)[1]
+    GeometryExtensionArray:PointType(geoarrow.point)[1]
     <POINT (0 1)>
     >>> ga.with_geometry_type(["LINESTRING EMPTY", "POINT (0 1)"], ga.GeometryType.POINT)
-    PointArray:PointType(geoarrow.point)[2]
+    GeometryExtensionArray:PointType(geoarrow.point)[2]
     <POINT (nan nan)>
     <POINT (0 1)>
     >>> ga.with_geometry_type(["MULTIPOINT (0 1, 2 3)"], ga.GeometryType.POINT)

geoarrow-pyarrow/src/geoarrow/pyarrow/_kernel.py

@@ -2,21 +2,29 @@
 
 import pyarrow as pa
 import pyarrow_hotfix as _  # noqa: F401
+from geoarrow.types import box as box_spec
 from geoarrow.pyarrow._type import GeometryExtensionType
 
 
 _lazy_lib = None
+_geoarrow_c_version = None
 
 
 def _geoarrow_c():
-    global _lazy_lib
+    global _lazy_lib, _geoarrow_c_version
     if _lazy_lib is None:
         try:
-            from geoarrow.c import lib
+            import geoarrow.c
+
         except ImportError as e:
             raise ImportError("Requested operation requires geoarrow-c") from e
 
-        _lazy_lib = lib
+        _lazy_lib = geoarrow.c.lib
+        if hasattr(geoarrow.c, "__version_tuple__"):
+            _geoarrow_c_version = geoarrow.c.__version_tuple__
+        else:
+            _geoarrow_c_version = (0, 1, 0)
+
     return _lazy_lib
 
 
@@ -109,11 +117,19 @@ def unique_geometry_types_agg(type_in):
 
     @staticmethod
     def box(type_in):
-        return Kernel("box", type_in)
+        kernel = Kernel("box", type_in)
+        if _geoarrow_c_version <= (0, 1, 3):
+            return BoxKernelCompat(kernel)
+        else:
+            return kernel
 
     @staticmethod
     def box_agg(type_in):
-        return Kernel("box_agg", type_in)
+        kernel = Kernel("box_agg", type_in)
+        if _geoarrow_c_version <= (0, 1, 3):
+            return BoxKernelCompat(kernel)
+        else:
+            return kernel
 
     @staticmethod
     def _pack_options(options):
@@ -132,3 +148,29 @@ def _pack_options(options):
             bytes += v.encode("UTF-8")
 
         return bytes
+
+
+class BoxKernelCompat:
+    """A wrapper around the "box" kernel that works for geoarrow-c 0.1.
+    This is mostly to ease the transition for geoarrow-python CI while
+    all the packages are being updated."""
+
+    def __init__(self, parent: Kernel):
+        self.parent = parent
+        self.type_out = box_spec().to_pyarrow().with_crs(parent._type_in.crs)
+
+    def push(self, arr):
+        parent_result = self.parent.push(arr)
+        return (
+            None if parent_result is None else self._old_box_to_new_box(parent_result)
+        )
+
+    def finish(self):
+        return self._old_box_to_new_box(self.parent.finish())
+
+    def _old_box_to_new_box(self, array):
+        xmin, xmax, ymin, ymax = array.flatten()
+        storage = pa.StructArray.from_arrays(
+            [xmin, ymin, xmax, ymax], names=["xmin", "ymin", "xmax", "ymax"]
+        )
+        return self.type_out.wrap_array(storage)