Add procedure for uploading gzipped file

* gn3/file_utils.py (extract_uploaded_file): New procedure.
* tests/unit/test_file_utils.py: Test cases for ^^.
* tests/unit/upload-data.tar.gz: New test data.

3 files changed, 75 insertions, 1 deletions

diff --git a/gn3/file_utils.py b/gn3/file_utils.py
index a6929e1..d9301bb 100644
--- a/gn3/file_utils.py
+++ b/gn3/file_utils.py
@@ -2,9 +2,12 @@
 import hashlib
 import json
 import os
+import shutil
+import tarfile
 
 from functools import partial
 from typing import Dict
+from werkzeug.utils import secure_filename
 from gn3.settings import APP_DEFAULTS
 
 
@@ -46,3 +49,29 @@ def jsonfile_to_dict(json_file: str) -> Dict:
         data = json.load(_file)
         return data
     raise FileNotFoundError
+
+
+def extract_uploaded_file(gzipped_file, target_dir: str) -> Dict:
+    """Get the (directory) hash of extracted contents of GZIPPED_FILE; and move
+contents to TARGET_DIR/<dir-hash>.
+
+    """
+    tar_target_loc = os.path.join(target_dir,
+                                  secure_filename(gzipped_file.filename))
+    gzipped_file.save(tar_target_loc)
+    try:
+        # Extract to "tar_target_loc/tempdir"
+        tar = tarfile.open(tar_target_loc)
+        tar.extractall(
+            path=os.path.join(target_dir, "tempdir"))
+        tar.close()
+    # pylint: disable=W0703
+    except Exception:
+        return {"status": 128, "error": "gzip failed to unpack file"}
+    dir_hash = get_dir_hash(tar_target_loc)
+    if os.path.exists(os.path.join(target_dir, dir_hash)):
+        shutil.rmtree(os.path.join(target_dir, 'tempdir'))
+    else:
+        os.rename(os.path.join(target_dir, "tempdir"),
+                  os.path.join(target_dir, dir_hash))
+    return {"status": 0, "token": dir_hash}
diff --git a/tests/unit/test_file_utils.py b/tests/unit/test_file_utils.py
index f5efa02..166e576 100644
--- a/tests/unit/test_file_utils.py
+++ b/tests/unit/test_file_utils.py
@@ -1,11 +1,22 @@
 """Test cases for procedures defined in file_utils.py"""
 import os
+import shutil
 import unittest
 
+from dataclasses import dataclass
+from typing import Callable
 from unittest import mock
+from gn3.file_utils import extract_uploaded_file
+from gn3.file_utils import get_dir_hash
 from gn3.file_utils import jsonfile_to_dict
 from gn3.file_utils import lookup_file
-from gn3.file_utils import get_dir_hash
+
+
+@dataclass
+class MockFile:
+    """Mock File object returned by request.files"""
+    filename: str
+    save: Callable
 
 
 class TestFileUtils(unittest.TestCase):
@@ -63,3 +74,37 @@ non-existent"""
         self.assertRaises(FileNotFoundError,
                           jsonfile_to_dict,
                           "/non-existent-dir")
+
+    @mock.patch("gn3.file_utils.secure_filename")
+    def test_extract_uploaded_file(self, mock_file):
+        """Test that the gzip file is extracted to the right location"""
+        file_loc = os.path.join(
+            os.path.dirname(__file__),
+            "upload-data.tar.gz")
+        mock_file.return_value = file_loc
+        mock_fileobj = MockFile(save=mock.MagicMock(),
+                                filename="upload-data.tar.gz")
+        result = extract_uploaded_file(mock_fileobj, "/tmp")
+        mock_fileobj.save.assert_called_once_with(file_loc)
+        mock_file.assert_called_once_with("upload-data.tar.gz")
+        # Clean up!
+        shutil.rmtree(os.path.join("/tmp",
+                                   "d41d8cd98f00b204e9800998ecf8427e"))
+        self.assertEqual(result,
+                         {"status": 0,
+                          "token": "d41d8cd98f00b204e9800998ecf8427e"})
+
+    @mock.patch("gn3.file_utils.secure_filename")
+    def test_extract_uploaded_file_non_existent_gzip(self, mock_file):
+        """Test that the right error message is returned when there is a problem
+extracting the file"""
+        file_loc = os.path.join(
+            os.path.dirname(__file__),
+            "CTtyodSTh5")  # Does not exist!
+        mock_file.return_value = file_loc
+        mock_fileobj = MockFile(save=mock.MagicMock(),
+                                filename="")
+        result = extract_uploaded_file(mock_fileobj, "/tmp")
+        self.assertEqual(result,
+                         {"status": 128,
+                          "error": "gzip failed to unpack file"})
diff --git a/tests/unit/upload-data.tar.gz b/tests/unit/upload-data.tar.gz
new file mode 100644
index 0000000..86077d9
--- /dev/null
+++ b/tests/unit/upload-data.tar.gz