set columns with list (#299)

* set columns with list

* lock

* fix test

* Check path existence

gokart/s3_zip_client.py

@@ -17,6 +17,9 @@ def exists(self) -> bool:
 
     def make_archive(self) -> None:
         extension = os.path.splitext(self._file_path)[1]
+        if not os.path.exists(self._temporary_directory):
+            # Check path existence since shutil.make_archive() of python 3.10+ does not check it.
+            raise FileNotFoundError(f'Temporary directory {self._temporary_directory} is not found.')
         shutil.make_archive(base_name=self._temporary_directory, format=extension[1:], root_dir=self._temporary_directory)
         self._client.put(self._temporary_file_path(), self._file_path)
 

gokart/task.py

@@ -270,7 +270,7 @@ def _flatten_recursively(dfs):
 
         required_columns = required_columns or set()
         if data.empty and len(data.index) == 0 and len(required_columns - set(data.columns)) > 0:
-            return pd.DataFrame(columns=required_columns)
+            return pd.DataFrame(columns=list(required_columns))
         assert required_columns.issubset(set(data.columns)), f'data must have columns {required_columns}, but actually have only {data.columns}.'
         if drop_columns:
             data = data[list(required_columns)]

gokart/testing/check_if_run_with_empty_data_frame.py

@@ -66,7 +66,7 @@ def _test_run_with_empty_data_frame(cmdline_args: List[str], test_run_params: te
     if test_run_params.namespace is not None:
         all_tasks = [t for t in all_tasks if t.task_namespace == test_run_params.namespace]
 
-    with patch('gokart.TaskOnKart.load_data_frame', new=lambda *args, required_columns=None, **kwargs: pd.DataFrame(columns=required_columns)):
+    with patch('gokart.TaskOnKart.load_data_frame', new=lambda *args, required_columns=None, **kwargs: pd.DataFrame(columns=list(required_columns))):
         with patch('gokart.TaskOnKart.dump', new=lambda *args, **kwargs: None):
             test_status_list = [_run_with_test_status(t) for t in all_tasks]