def test_from_dask_dataframe_with_column_name():
df = pd.DataFrame({"x": [0, 1, 2, 3], "y": [1, 2, 3, 4]})
df["geoms"] = geopandas.points_from_xy(df["x"], df["y"])
dask_obj = dd.from_pandas(df, npartitions=2)
dask_obj = dask_geopandas.from_dask_dataframe(dask_obj, geometry="geoms")
expected = geopandas.GeoDataFrame(df, geometry="geoms")
assert_geodataframe_equal(dask_obj.compute(), expected)
def test_from_dask_dataframe_with_dask_geoseries():
df = pd.DataFrame({"x": [0, 1, 2, 3], "y": [1, 2, 3, 4]})
dask_obj = dd.from_pandas(df, npartitions=2)
dask_obj = dask_geopandas.from_dask_dataframe(
dask_obj, geometry=dask_geopandas.points_from_xy(dask_obj, "x", "y"))
expected = df.set_geometry(geopandas.points_from_xy(df["x"], df["y"]))
assert_geoseries_equal(dask_obj.geometry.compute(), expected.geometry)
def test_points_from_xy():
x = [1, 2, 3, 4, 5]
y = [4, 5, 6, 7, 8]
expected = geopandas.points_from_xy(x, y)
df = pd.DataFrame({"x": x, "y": y})
ddf = dd.from_pandas(df, npartitions=2)
actual = dask_geopandas.points_from_xy(ddf)
assert isinstance(actual, dask_geopandas.GeoSeries)
assert list(actual) == list(expected)
# assign to geometry column and convert to GeoDataFrame
df["geometry"] = expected
expected = geopandas.GeoDataFrame(df)
ddf["geometry"] = actual
ddf = dask_geopandas.from_dask_dataframe(ddf)
result = ddf.compute()
assert_geodataframe_equal(result, expected)
def spatial_join_map_partition(points_filepath,
nuts_filepath,
blocksize=1_000_000):
def spatial_join(gdf_regions):
def compute_spatial_join(df):
df = sjoin(df, gdf_regions, how='left')
return df[['nuts_id', 'amenity', 'osm_id']]
return compute_spatial_join
gdf_nuts = gpd.read_file(nuts_filepath,
ignore_fields=[
'levl_code', 'cntr_code', 'name_latn',
'nuts_name', 'population'
],
driver='GPKG')
ddf_amenity = dd.read_csv(points_filepath, blocksize=blocksize)
ddf_amenity = dask_geopandas.from_dask_dataframe(ddf_amenity)
ddf_amenity = ddf_amenity.set_geometry(
dask_geopandas.points_from_xy(ddf_amenity, 'lon', 'lat'))
ddf_amenity = ddf_amenity.set_crs(pyproj.CRS(4326))
print(ddf_amenity.npartitions)
ddf_amenity = ddf_amenity.map_partitions(spatial_join(gdf_nuts.copy()),
meta={
'nuts_id': object,
'amenity': object,
'osm_id': object
})
s = ddf_amenity.groupby(['nuts_id', 'amenity'])['osm_id'] \
.count() \
.compute()
s.name = 'counts'
df = s.reset_index()
return df