def test_cache_working(self):
f1 = 'f1.shp'
f1 = create_dummy_shp(f1)
cf1 = utils.cached_shapefile_path(f1)
self.assertFalse(os.path.exists(cf1))
_ = read_shapefile(f1)
self.assertFalse(os.path.exists(cf1))
_ = read_shapefile(f1, cached=True)
self.assertTrue(os.path.exists(cf1))
# nested calls
self.assertTrue(cf1 == utils.cached_shapefile_path(cf1))
# wait a bit
time.sleep(0.1)
f1 = create_dummy_shp(f1)
cf2 = utils.cached_shapefile_path(f1)
self.assertFalse(os.path.exists(cf1))
_ = read_shapefile(f1, cached=True)
self.assertFalse(os.path.exists(cf1))
self.assertTrue(os.path.exists(cf2))
df = read_shapefile(f1, cached=True)
np.testing.assert_allclose(df.min_x, [1., 2.])
np.testing.assert_allclose(df.max_x, [2., 3.])
np.testing.assert_allclose(df.min_y, [1., 1.3])
np.testing.assert_allclose(df.max_y, [2., 2.3])
self.assertRaises(ValueError, read_shapefile, 'f1.sph')
self.assertRaises(ValueError, utils.cached_shapefile_path, 'f1.splash')
def test_cache_working(self):
f1 = "f1.shp"
f1 = create_dummy_shp(f1)
cf1 = utils.cached_shapefile_path(f1)
self.assertFalse(os.path.exists(cf1))
_ = read_shapefile(f1)
self.assertFalse(os.path.exists(cf1))
_ = read_shapefile(f1, cached=True)
self.assertTrue(os.path.exists(cf1))
# nested calls
self.assertTrue(cf1 == utils.cached_shapefile_path(cf1))
# wait a bit
time.sleep(0.1)
f1 = create_dummy_shp(f1)
cf2 = utils.cached_shapefile_path(f1)
self.assertFalse(os.path.exists(cf1))
_ = read_shapefile(f1, cached=True)
self.assertFalse(os.path.exists(cf1))
self.assertTrue(os.path.exists(cf2))
df = read_shapefile(f1, cached=True)
np.testing.assert_allclose(df.min_x, [1.0, 2.0])
np.testing.assert_allclose(df.max_x, [2.0, 3.0])
np.testing.assert_allclose(df.min_y, [1.0, 1.3])
np.testing.assert_allclose(df.max_y, [2.0, 2.3])
self.assertRaises(ValueError, read_shapefile, "f1.sph")
self.assertRaises(ValueError, utils.cached_shapefile_path, "f1.splash")
def read_shapefile_to_grid(fpath, grid):
"""Same as read_shapefile but directly transformed to a grid.
The whole thing is cached so that the second call will
will be much faster.
Parameters
----------
fpath: path to the file
grid: the arrival grid
"""
# ensure it is a cached pickle (copy code smell)
shape_cpath = cached_shapefile_path(fpath)
if not os.path.exists(shape_cpath):
out = read_shapefile(fpath, cached=False)
with open(shape_cpath, 'wb') as f:
pickle.dump(out, f)
return _memory_shapefile_to_grid(shape_cpath, grid=grid, **grid.to_dict())
def read_shapefile_to_grid(fpath, grid):
"""Same as read_shapefile but directly transformed to a grid.
The whole thing is cached so that the second call will
will be much faster.
Parameters
----------
fpath: path to the file
grid: the arrival grid
"""
# ensure it is a cached pickle (copy code smell)
shape_cpath = cached_shapefile_path(fpath)
if not os.path.exists(shape_cpath):
out = read_shapefile(fpath, cached=False)
with open(shape_cpath, 'wb') as f:
pickle.dump(out, f)
return _memory_transform(shape_cpath, grid=grid, grid_str=str(grid))
def read_shapefile(fpath, cached=False):
"""Reads a shapefile using geopandas.
For convenience, it adds four columns to the dataframe:
[min_x, max_x, min_y, max_y]
Because reading a shapefile can take a long time, Salem provides a
caching utility (cached=True). This will save a pickle of the shapefile
in the cache directory.
"""
import geopandas as gpd
_, ext = os.path.splitext(fpath)
if ext.lower() in ['.shp', '.p']:
if cached:
cpath = cached_shapefile_path(fpath)
# unpickle if cached, read and pickle if not
if os.path.exists(cpath):
with open(cpath, 'rb') as f:
out = pickle.load(f)
else:
out = read_shapefile(fpath, cached=False)
with open(cpath, 'wb') as f:
pickle.dump(out, f)
else:
out = gpd.read_file(fpath)
out['min_x'] = [g.bounds[0] for g in out.geometry]
out['max_x'] = [g.bounds[2] for g in out.geometry]
out['min_y'] = [g.bounds[1] for g in out.geometry]
out['max_y'] = [g.bounds[3] for g in out.geometry]
else:
raise ValueError('File extension not recognised: {}'.format(ext))
return out
def read_shapefile(fpath, cached=False):
"""Reads a shapefile using geopandas.
For convenience, it adds four columns to the dataframe:
[min_x, max_x, min_y, max_y]
Because reading a shapefile can take a long time, Salem provides a
caching utility (cached=True). This will save a pickle of the shapefile
in the cache directory.
"""
import geopandas as gpd
_, ext = os.path.splitext(fpath)
if ext.lower() in ['.shp', '.p']:
if cached:
cpath = cached_shapefile_path(fpath)
# unpickle if cached, read and pickle if not
if os.path.exists(cpath):
with open(cpath, 'rb') as f:
out = pickle.load(f)
else:
out = read_shapefile(fpath, cached=False)
with open(cpath, 'wb') as f:
pickle.dump(out, f)
else:
out = gpd.read_file(fpath)
out['min_x'] = [g.bounds[0] for g in out.geometry]
out['max_x'] = [g.bounds[2] for g in out.geometry]
out['min_y'] = [g.bounds[1] for g in out.geometry]
out['max_y'] = [g.bounds[3] for g in out.geometry]
else:
raise ValueError('File extension not recognised: {}'.format(ext))
return out
