def test_open_as_container(self):
# Create a container for testing
pfio.set_root("posix")
zip_file_name = "test"
zip_file_path = zip_file_name + ".zip"
# in the zip, the leading slash will be removed
# TODO(tianqi): related to issue #61
dirname_zip = self.tmpdir.name.lstrip('/') + '/'
file_name_zip = self.tmpfile_path.lstrip('/')
first_level_dir = dirname_zip.split('/')[0]
shutil.make_archive(zip_file_name, "zip", base_dir=self.tmpdir.name)
with pfio.open_as_container(zip_file_path) as container:
file_generator = container.list()
file_list = list(file_generator)
self.assertIn(first_level_dir, file_list)
self.assertNotIn(file_name_zip, file_list)
self.assertNotIn("", file_list)
file_generator = container.list(dirname_zip)
file_list = list(file_generator)
self.assertNotIn(first_level_dir, file_list)
self.assertIn(os.path.basename(file_name_zip), file_list)
self.assertNotIn("", file_list)
self.assertTrue(container.isdir(dirname_zip))
self.assertFalse(container.isdir(file_name_zip))
self.assertIsInstance(container.info(), str)
with container.open(file_name_zip, "r") as f:
self.assertEqual(f.read(), self.test_string_str)
pfio.remove(zip_file_path)
def test_root_local_override(self):
pfio.set_root('file://' + self.tmpdir.name)
with pfio.open(self.tmpfile_name, "r") as fp:
self.assertEqual(fp.read(), self.test_string_str)
# override with full URI
with open(__file__, "r") as my_script:
with pfio.open('file://' + __file__) as fp:
self.assertEqual(fp.read(), my_script.read().encode("utf-8"))
def test_set_root(self):
# Set default context globally in this process
pfio.set_root('posix')
# Using the context to open local file
with pfio.open(self.tmpfile_path, "r") as fp:
self.assertEqual(fp.read(), self.test_string_str)
pfio.set_root('file://' + self.tmpdir.name)
with pfio.open(self.tmpfile_name, "r") as fp:
self.assertEqual(fp.read(), self.test_string_str)
def test_root_fs_override(self):
from pyarrow import hdfs
hdfs_tmpfile = "tmpfile_hdfs"
hdfs_file_string = "this is a test string for hdfs"
conn = hdfs.connect()
with conn.open(hdfs_tmpfile, "wb") as f:
f.write(hdfs_file_string.encode('utf-8'))
pfio.set_root("hdfs")
with pfio.open(hdfs_tmpfile, "r") as fp:
self.assertEqual(fp.read(), hdfs_file_string)
# override with full URI
with open(__file__, "r") as my_script:
with pfio.open("file://" + __file__, "r") as fp:
self.assertEqual(fp.read(), my_script.read())
with pfio.open(hdfs_tmpfile, "r") as fp:
self.assertEqual(fp.read(), hdfs_file_string)
conn.delete(hdfs_tmpfile)
conn.close()