Merge pull request #19 from ian-r-rose/geoparquet

Geoparquet

conda/meta.yaml

@@ -28,6 +28,7 @@ test:
     - tests
   requires:
     - aiohttp
+    - pyarrow
     - pytest
     - requests
   commands:

intake_geopandas/__init__.py

@@ -2,15 +2,23 @@
 import intake
 
 from ._version import get_versions
-from .geopandas import GeoJSONSource, PostGISSource, ShapefileSource, SpatiaLiteSource
+from .geopandas import (
+    GeoJSONSource,
+    GeoPandasFileSource,
+    GeoParquetSource,
+    PostGISSource,
+    ShapefileSource,
+    SpatiaLiteSource,
+)
 from .regionmask import RegionmaskSource
 
 __version__ = get_versions()['version']
 del get_versions
 
-
 __all__ = [
     'GeoJSONSource',
+    'GeoPandasFileSource',
+    'GeoParquetSource',
     'PostGISSource',
     'ShapefileSource',
     'SpatiaLiteSource',

intake_geopandas/geopandas.py

@@ -57,6 +57,8 @@ def _close(self):
 
 
 class GeoPandasFileSource(GeoPandasSource):
+    name="geopandasfile"
+
     def __init__(
         self,
         urlpath,
@@ -67,6 +69,11 @@ def __init__(
         metadata=None,
     ):
         """
+        A source for a file opened by geopandas. Specializations of this are provided
+        for shapefiles and geojson, but this base class can also be used directly
+        for other file types by providing an OGR driver to `geopandas_kwargs`, e.g.
+        `geopandas_kwargs={"driver": "GPKG"}` to open a geopackage.
+
         Parameters
         ----------
         urlpath : str or iterable, location of data
@@ -149,6 +156,27 @@ def _resolve_single_file(self, filelist):
         )
 
 
+class GeoParquetSource(GeoPandasFileSource):
+    name = "geoparquet"
+
+    def _open_dataset(self):
+        """
+        Open dataset using geopandas.
+        """
+        if self._use_fsspec:
+            with fsspec.open_files(self.urlpath, **self._storage_options) as f:
+                f = self._resolve_single_file(f) if len(f) > 1 else f[0]
+                self._dataframe = geopandas.read_parquet(
+                    f,
+                    **self._geopandas_kwargs,
+                )
+        else:
+            self._dataframe = geopandas.read_parquet(
+                self.urlpath,
+                **self._geopandas_kwargs
+            )
+
+
 class GeoPandasSQLSource(GeoPandasSource):
     def __init__(
         self, uri, sql_expr=None, table=None, geopandas_kwargs=None, metadata=None

requirements.txt

@@ -1,4 +1,2 @@
-intake
 geopandas
-pytest
-fsspec
+intake

setup.py

@@ -12,24 +12,33 @@
     cmdclass=versioneer.get_cmdclass(),
     description='Geopandas plugin for Intake',
     url='https://github.com/intake/intake_geopandas',
-    maintainer='Jacob Tomlinson',
-    maintainer_email='jacob.tomlinson@informaticslab.co.uk',
+    maintainer='Ian Rose',
+    maintainer_email='ian.r.rose@gmail.com',
     license='BSD',
     py_modules=['intake_geopandas'],
     packages=find_packages(),
     package_data={'': ['*.csv', '*.yml', '*.html']},
     entry_points={
         'intake.drivers': [
             'geojson = intake_geopandas.geopandas:GeoJSONSource',
+            'geopandasfile = intake_geopandas.geopandas:GeoPandasFileSource',
+            'geoparquet = intake_geopandas.geopandas:GeoParquetSource',
             'postgis = intake_geopandas.geopandas:PostGISSource',
             'shapefile = intake_geopandas.geopandas:ShapefileSource',
             'spatialite = intake_geopandas.geopandas:SpatiaLiteSource',
             'regionmask = intake_geopandas.regionmask:RegionmaskSource',
         ]
     },
+    classifiers=[
+        "Programming Language :: Python :: 3",
+        "Programming Language :: Python :: 3.6",
+        "Programming Language :: Python :: 3.7",
+        "Programming Language :: Python :: 3.8",
+    ],
+    python_requires=">=3.6",
     include_package_data=True,
     install_requires=requires,
-    extras_require={'':['regionmask']},
+    extras_require={"arrow": ["pyarrow"], "regionmask":["regionmask"]},
     long_description_content_type='text/markdown',
     long_description=open('README.md').read(),
     zip_safe=False,

tests/test_file_source.py

@@ -4,7 +4,8 @@
 
 from pkg_resources import get_distribution, parse_version
 
-from intake_geopandas import GeoJSONSource, ShapefileSource
+from intake_geopandas import GeoJSONSource, GeoPandasFileSource, GeoParquetSource, ShapefileSource
+from geopandas import GeoDataFrame
 
 geom_col_type = (
     "object"
@@ -30,6 +31,18 @@ def geojson_filenames():
     return dict(countries=os.path.join(basedir, "data", "countries.geo.json"))
 
 
+@pytest.fixture
+def gpkg_filename():
+    basedir = os.path.dirname(__file__)
+    return os.path.join(basedir, "data", "countries.gpkg")
+
+
+@pytest.fixture
+def geoparquet_filename():
+    basedir = os.path.dirname(__file__)
+    return os.path.join(basedir, "data", "countries.parquet")
+
+
 @pytest.fixture
 def geojson_datasource(geojson_filenames):
     return GeoJSONSource(geojson_filenames["countries"])
@@ -47,11 +60,26 @@ def test_shape_datasource(shape_datasource):
     }
 
 
-def test_countries_datasource(geojson_datasource):
+def test_geojson_datasource(geojson_datasource):
     info = geojson_datasource.discover()
     geojson_datasource.read()
     assert info["dtype"] == {
         "geometry": geom_col_type,
         "id": "object",
         "name": "object",
     }
+
+
+def test_alternative_ogr_driver(gpkg_filename):
+    gpkg_datasource = GeoPandasFileSource(
+      gpkg_filename,
+      geopandas_kwargs={"driver": "GPKG"},
+    )
+    gdf = gpkg_datasource.read()
+    assert isinstance(gdf, GeoDataFrame)
+
+
+def test_geoparquet_source(geoparquet_filename):
+    datasource = GeoParquetSource(geoparquet_filename)
+    gdf = datasource.read()
+    assert isinstance(gdf, GeoDataFrame)