def test_spatial_knn_query_correctness(self):
rectangle_rdd = RectangleRDD(self.sc, inputLocation, offset, splitter,
True)
result_no_index = KNNQuery.SpatialKnnQuery(rectangle_rdd,
self.query_point,
self.top_k, False)
rectangle_rdd.buildIndex(IndexType.RTREE, False)
result_with_index = KNNQuery.SpatialKnnQuery(rectangle_rdd,
self.query_point,
self.top_k, True)
sorted_result_no_index = sorted(
result_no_index,
key=lambda geo_data: distance_sorting_functions(
geo_data, self.query_point))
sorted_result_with_index = sorted(
result_with_index,
key=lambda geo_data: distance_sorting_functions(
geo_data, self.query_point))
difference = 0
for x in range(self.top_k):
difference += sorted_result_no_index[x].geom.distance(
sorted_result_with_index[x].geom)
assert difference == 0
def MinimumBoundingRectangle(self):
from geospark.core.SpatialRDD import RectangleRDD
rectangle_rdd = RectangleRDD()
srdd = self._srdd.MinimumBoundingRectangle()
rectangle_rdd.set_srdd(srdd)
return rectangle_rdd
def test_spatial_knn_query_using_index(self):
rectangle_rdd = RectangleRDD(self.sc, inputLocation, offset, splitter,
True)
rectangle_rdd.buildIndex(IndexType.RTREE, False)
for i in range(self.loop_times):
result = KNNQuery.SpatialKnnQuery(rectangle_rdd, self.query_point,
self.top_k, False)
assert result.__len__() > -1
assert result[0].getUserData() is not None
def test_spatial_range_query_using_index(self):
spatial_rdd = RectangleRDD(self.sc, inputLocation, offset, splitter,
True, StorageLevel.MEMORY_ONLY)
spatial_rdd.buildIndex(IndexType.RTREE, False)
for i in range(self.loop_times):
result_size = RangeQuery.SpatialRangeQuery(spatial_rdd,
self.query_envelope,
False, True).count()
assert result_size == 193
assert RangeQuery.SpatialRangeQuery(spatial_rdd, self.query_envelope, False, True).take(10)[1].getUserData()\
is not None
def test_spatial_knn_using_polygon(self):
rectangle_rdd = RectangleRDD(self.sc, inputLocation, offset, splitter,
True)
result_no_index = KNNQuery.SpatialKnnQuery(rectangle_rdd,
self.query_polygon,
self.top_k, False)
print(result_no_index)
def test_voronoi_spatial_partitioning(self):
spatial_rdd = RectangleRDD(sparkContext=self.sc,
InputLocation=inputLocation,
Offset=offset,
splitter=splitter,
carryInputData=True,
partitions=10,
newLevel=StorageLevel.MEMORY_ONLY)
spatial_rdd.analyze()
spatial_rdd.spatialPartitioning(GridType.VORONOI)
for envelope in spatial_rdd.grids:
print(envelope)
assert spatial_rdd.countWithoutDuplicates(
) == spatial_rdd.countWithoutDuplicatesSPRDD()
def test_build_index_without_set_grid(self):
spatial_rdd = RectangleRDD(sparkContext=self.sc,
InputLocation=inputLocation,
Offset=offset,
splitter=splitter,
carryInputData=True,
partitions=numPartitions)
spatial_rdd.analyze()
spatial_rdd.buildIndex(IndexType.RTREE, False)
def test_constructor(self):
spatial_rdd = RectangleRDD(sparkContext=self.sc,
InputLocation=inputLocation,
Offset=offset,
splitter=splitter,
carryInputData=True,
partitions=numPartitions,
newLevel=StorageLevel.MEMORY_ONLY)
spatial_rdd.analyze()
assert inputCount == spatial_rdd.approximateTotalCount
assert inputBoundary == spatial_rdd.boundaryEnvelope
spatial_rdd = RectangleRDD(self.sc, inputLocation, offset, splitter,
True, numPartitions,
StorageLevel.MEMORY_ONLY)
spatial_rdd.analyze()
assert inputCount == spatial_rdd.approximateTotalCount
assert inputBoundary == spatial_rdd.boundaryEnvelope
def getCenterRectangleRDDAsSpatialRDD(self) -> 'RectangleRDD':
from geospark.core.SpatialRDD import RectangleRDD
srdd = self._srdd.getCenterLineStringRDDAsSpatialRDD()
rectangle_rdd = RectangleRDD()
rectangle_rdd.set_srdd(srdd)
return rectangle_rdd
def test_empty_constructor(self):
spatial_rdd = RectangleRDD(sparkContext=self.sc,
InputLocation=inputLocation,
Offset=offset,
splitter=splitter,
carryInputData=True,
partitions=numPartitions,
newLevel=StorageLevel.MEMORY_ONLY)
spatial_rdd.analyze()
spatial_rdd.buildIndex(IndexType.RTREE, False)
spatial_rdd_copy = RectangleRDD()
spatial_rdd_copy.rawJvmSpatialRDD = spatial_rdd.rawJvmSpatialRDD
spatial_rdd_copy.analyze()
def create_rectangle_rdd(self, location, splitter, num_partitions):
rdd = RectangleRDD(self.sc, location, splitter, True, num_partitions)
return RectangleRDD(rdd.rawJvmSpatialRDD, StorageLevel.MEMORY_ONLY)