def execute(self, context):
gcs_hook = GCSHook(gcp_conn_id=self.gcp_conn_id,
delegate_to=self.delegate_to)
sftp_hook = SFTPHook(self.sftp_conn_id)
if WILDCARD in self.source_path:
total_wildcards = self.source_path.count(WILDCARD)
if total_wildcards > 1:
raise AirflowException(
"Only one wildcard '*' is allowed in source_path parameter. "
"Found {} in {}.".format(total_wildcards,
self.source_path))
prefix, delimiter = self.source_path.split(WILDCARD, 1)
base_path = os.path.dirname(prefix)
files, _, _ = sftp_hook.get_tree_map(base_path,
prefix=prefix,
delimiter=delimiter)
for file in files:
destination_path = file.replace(base_path,
self.destination_path, 1)
self._copy_single_object(gcs_hook, sftp_hook, file,
destination_path)
else:
destination_object = (self.destination_path
if self.destination_path else
self.source_path.rsplit("/", 1)[1])
self._copy_single_object(gcs_hook, sftp_hook, self.source_path,
destination_object)
class TestSFTPHook(unittest.TestCase):
@provide_session
def update_connection(self, login, session=None):
connection = (session.query(Connection).filter(
Connection.conn_id == "sftp_default").first())
old_login = connection.login
connection.login = login
session.commit()
return old_login
def setUp(self):
self.old_login = self.update_connection(SFTP_CONNECTION_USER)
self.hook = SFTPHook()
os.makedirs(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, SUB_DIR))
with open(os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS), 'a') as file:
file.write('Test file')
with open(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, SUB_DIR,
TMP_FILE_FOR_TESTS), 'a') as file:
file.write('Test file')
def test_get_conn(self):
output = self.hook.get_conn()
self.assertEqual(type(output), pysftp.Connection)
def test_close_conn(self):
self.hook.conn = self.hook.get_conn()
self.assertTrue(self.hook.conn is not None)
self.hook.close_conn()
self.assertTrue(self.hook.conn is None)
def test_describe_directory(self):
output = self.hook.describe_directory(TMP_PATH)
self.assertTrue(TMP_DIR_FOR_TESTS in output)
def test_list_directory(self):
output = self.hook.list_directory(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
self.assertEqual(output, [SUB_DIR])
def test_create_and_delete_directory(self):
new_dir_name = 'new_dir'
self.hook.create_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, new_dir_name))
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
self.assertTrue(new_dir_name in output)
self.hook.delete_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, new_dir_name))
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
self.assertTrue(new_dir_name not in output)
def test_create_and_delete_directories(self):
base_dir = "base_dir"
sub_dir = "sub_dir"
new_dir_path = os.path.join(base_dir, sub_dir)
self.hook.create_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, new_dir_path))
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
self.assertTrue(base_dir in output)
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, base_dir))
self.assertTrue(sub_dir in output)
self.hook.delete_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, new_dir_path))
self.hook.delete_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, base_dir))
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
self.assertTrue(new_dir_path not in output)
self.assertTrue(base_dir not in output)
def test_store_retrieve_and_delete_file(self):
self.hook.store_file(
remote_full_path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS,
TMP_FILE_FOR_TESTS),
local_full_path=os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS))
output = self.hook.list_directory(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
self.assertEqual(output, [SUB_DIR, TMP_FILE_FOR_TESTS])
retrieved_file_name = 'retrieved.txt'
self.hook.retrieve_file(
remote_full_path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS,
TMP_FILE_FOR_TESTS),
local_full_path=os.path.join(TMP_PATH, retrieved_file_name))
self.assertTrue(retrieved_file_name in os.listdir(TMP_PATH))
os.remove(os.path.join(TMP_PATH, retrieved_file_name))
self.hook.delete_file(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, TMP_FILE_FOR_TESTS))
output = self.hook.list_directory(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
self.assertEqual(output, [SUB_DIR])
def test_get_mod_time(self):
self.hook.store_file(
remote_full_path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS,
TMP_FILE_FOR_TESTS),
local_full_path=os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS))
output = self.hook.get_mod_time(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, TMP_FILE_FOR_TESTS))
self.assertEqual(len(output), 14)
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_default(self, get_connection):
connection = Connection(login='******', host='host')
get_connection.return_value = connection
hook = SFTPHook()
self.assertEqual(hook.no_host_key_check, False)
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_enabled(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"no_host_key_check": true}')
get_connection.return_value = connection
hook = SFTPHook()
self.assertEqual(hook.no_host_key_check, True)
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_disabled(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"no_host_key_check": false}')
get_connection.return_value = connection
hook = SFTPHook()
self.assertEqual(hook.no_host_key_check, False)
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_disabled_for_all_but_true(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"no_host_key_check": "foo"}')
get_connection.return_value = connection
hook = SFTPHook()
self.assertEqual(hook.no_host_key_check, False)
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_ignore(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"ignore_hostkey_verification": true}')
get_connection.return_value = connection
hook = SFTPHook()
self.assertEqual(hook.no_host_key_check, True)
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_no_ignore(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"ignore_hostkey_verification": false}')
get_connection.return_value = connection
hook = SFTPHook()
self.assertEqual(hook.no_host_key_check, False)
@parameterized.expand([
(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS), True),
(os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS), True),
(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS + "abc"), False),
(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, "abc"), False),
])
def test_path_exists(self, path, exists):
result = self.hook.path_exists(path)
self.assertEqual(result, exists)
@parameterized.expand([
("test/path/file.bin", None, None, True),
("test/path/file.bin", "test", None, True),
("test/path/file.bin", "test/", None, True),
("test/path/file.bin", None, "bin", True),
("test/path/file.bin", "test", "bin", True),
("test/path/file.bin", "test/", "file.bin", True),
("test/path/file.bin", None, "file.bin", True),
("test/path/file.bin", "diff", None, False),
("test/path/file.bin", "test//", None, False),
("test/path/file.bin", None, ".txt", False),
("test/path/file.bin", "diff", ".txt", False),
])
def test_path_match(self, path, prefix, delimiter, match):
result = self.hook._is_path_match(path=path,
prefix=prefix,
delimiter=delimiter)
self.assertEqual(result, match)
def test_get_tree_map(self):
tree_map = self.hook.get_tree_map(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
files, dirs, unknowns = tree_map
self.assertEqual(files, [
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, SUB_DIR,
TMP_FILE_FOR_TESTS)
])
self.assertEqual(dirs,
[os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, SUB_DIR)])
self.assertEqual(unknowns, [])
def tearDown(self):
shutil.rmtree(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
os.remove(os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS))
self.update_connection(self.old_login)
class TestSFTPHook(unittest.TestCase):
@provide_session
def update_connection(self, login, session=None):
connection = session.query(Connection).filter(
Connection.conn_id == "sftp_default").first()
old_login = connection.login
connection.login = login
session.commit()
return old_login
def setUp(self):
self.old_login = self.update_connection(SFTP_CONNECTION_USER)
self.hook = SFTPHook()
os.makedirs(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, SUB_DIR))
with open(os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS), 'a') as file:
file.write('Test file')
with open(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, SUB_DIR,
TMP_FILE_FOR_TESTS), 'a') as file:
file.write('Test file')
def test_get_conn(self):
output = self.hook.get_conn()
assert isinstance(output, pysftp.Connection)
def test_close_conn(self):
self.hook.conn = self.hook.get_conn()
assert self.hook.conn is not None
self.hook.close_conn()
assert self.hook.conn is None
def test_describe_directory(self):
output = self.hook.describe_directory(TMP_PATH)
assert TMP_DIR_FOR_TESTS in output
def test_list_directory(self):
output = self.hook.list_directory(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
assert output == [SUB_DIR]
def test_create_and_delete_directory(self):
new_dir_name = 'new_dir'
self.hook.create_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, new_dir_name))
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
assert new_dir_name in output
self.hook.delete_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, new_dir_name))
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
assert new_dir_name not in output
def test_create_and_delete_directories(self):
base_dir = "base_dir"
sub_dir = "sub_dir"
new_dir_path = os.path.join(base_dir, sub_dir)
self.hook.create_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, new_dir_path))
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
assert base_dir in output
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, base_dir))
assert sub_dir in output
self.hook.delete_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, new_dir_path))
self.hook.delete_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, base_dir))
output = self.hook.describe_directory(
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
assert new_dir_path not in output
assert base_dir not in output
def test_store_retrieve_and_delete_file(self):
self.hook.store_file(
remote_full_path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS,
TMP_FILE_FOR_TESTS),
local_full_path=os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS),
)
output = self.hook.list_directory(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
assert output == [SUB_DIR, TMP_FILE_FOR_TESTS]
retrieved_file_name = 'retrieved.txt'
self.hook.retrieve_file(
remote_full_path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS,
TMP_FILE_FOR_TESTS),
local_full_path=os.path.join(TMP_PATH, retrieved_file_name),
)
assert retrieved_file_name in os.listdir(TMP_PATH)
os.remove(os.path.join(TMP_PATH, retrieved_file_name))
self.hook.delete_file(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, TMP_FILE_FOR_TESTS))
output = self.hook.list_directory(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
assert output == [SUB_DIR]
def test_get_mod_time(self):
self.hook.store_file(
remote_full_path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS,
TMP_FILE_FOR_TESTS),
local_full_path=os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS),
)
output = self.hook.get_mod_time(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, TMP_FILE_FOR_TESTS))
assert len(output) == 14
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_default(self, get_connection):
connection = Connection(login='******', host='host')
get_connection.return_value = connection
hook = SFTPHook()
assert hook.no_host_key_check is False
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_enabled(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"no_host_key_check": true}')
get_connection.return_value = connection
hook = SFTPHook()
assert hook.no_host_key_check is True
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_disabled(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"no_host_key_check": false}')
get_connection.return_value = connection
hook = SFTPHook()
assert hook.no_host_key_check is False
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_ciphers(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"ciphers": ["A", "B", "C"]}')
get_connection.return_value = connection
hook = SFTPHook()
assert hook.ciphers == ["A", "B", "C"]
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_disabled_for_all_but_true(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"no_host_key_check": "foo"}')
get_connection.return_value = connection
hook = SFTPHook()
assert hook.no_host_key_check is False
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_ignore(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"ignore_hostkey_verification": true}')
get_connection.return_value = connection
hook = SFTPHook()
assert hook.no_host_key_check is True
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_no_host_key_check_no_ignore(self, get_connection):
connection = Connection(login='******',
host='host',
extra='{"ignore_hostkey_verification": false}')
get_connection.return_value = connection
hook = SFTPHook()
assert hook.no_host_key_check is False
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_host_key_default(self, get_connection):
connection = Connection(login='******', host='host')
get_connection.return_value = connection
hook = SFTPHook()
assert hook.host_key is None
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_host_key(self, get_connection):
connection = Connection(
login='******',
host='host',
extra=json.dumps({
"host_key": TEST_HOST_KEY,
"no_host_key_check": False
}),
)
get_connection.return_value = connection
hook = SFTPHook()
assert hook.host_key.get_base64() == TEST_HOST_KEY
@mock.patch('airflow.providers.sftp.hooks.sftp.SFTPHook.get_connection')
def test_host_key_with_no_host_key_check(self, get_connection):
connection = Connection(login='******',
host='host',
extra=json.dumps({"host_key": TEST_HOST_KEY}))
get_connection.return_value = connection
hook = SFTPHook()
assert hook.host_key is None
@parameterized.expand([
(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS), True),
(os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS), True),
(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS + "abc"), False),
(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, "abc"), False),
])
def test_path_exists(self, path, exists):
result = self.hook.path_exists(path)
assert result == exists
@parameterized.expand([
("test/path/file.bin", None, None, True),
("test/path/file.bin", "test", None, True),
("test/path/file.bin", "test/", None, True),
("test/path/file.bin", None, "bin", True),
("test/path/file.bin", "test", "bin", True),
("test/path/file.bin", "test/", "file.bin", True),
("test/path/file.bin", None, "file.bin", True),
("test/path/file.bin", "diff", None, False),
("test/path/file.bin", "test//", None, False),
("test/path/file.bin", None, ".txt", False),
("test/path/file.bin", "diff", ".txt", False),
])
def test_path_match(self, path, prefix, delimiter, match):
result = self.hook._is_path_match(path=path,
prefix=prefix,
delimiter=delimiter)
assert result == match
def test_get_tree_map(self):
tree_map = self.hook.get_tree_map(
path=os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
files, dirs, unknowns = tree_map
assert files == [
os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, SUB_DIR,
TMP_FILE_FOR_TESTS)
]
assert dirs == [os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS, SUB_DIR)]
assert unknowns == []
def tearDown(self):
shutil.rmtree(os.path.join(TMP_PATH, TMP_DIR_FOR_TESTS))
os.remove(os.path.join(TMP_PATH, TMP_FILE_FOR_TESTS))
self.update_connection(self.old_login)
