Feature: Allow a data only export/import cycle (#6871)

src/documents/tests/test_management_exporter.py

@@ -37,10 +37,16 @@ from documents.sanity_checker import check_sanity
 from documents.settings import EXPORTER_FILE_NAME
 from documents.tests.utils import DirectoriesMixin
 from documents.tests.utils import FileSystemAssertsMixin
+from documents.tests.utils import SampleDirMixin
 from documents.tests.utils import paperless_environment
 
 
-class TestExportImport(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
+class TestExportImport(
+    DirectoriesMixin,
+    FileSystemAssertsMixin,
+    SampleDirMixin,
+    TestCase,
+):
     def setUp(self) -> None:
         self.target = Path(tempfile.mkdtemp())
         self.addCleanup(shutil.rmtree, self.target)
@@ -139,6 +145,7 @@ class TestExportImport(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
     @override_settings(PASSPHRASE="test")
     def _do_export(
         self,
+        *,
         use_filename_format=False,
         compare_checksums=False,
         delete=False,
@@ -146,6 +153,7 @@ class TestExportImport(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
         no_thumbnail=False,
         split_manifest=False,
         use_folder_prefix=False,
+        data_only=False,
     ):
         args = ["document_exporter", self.target]
         if use_filename_format:
@@ -162,6 +170,8 @@ class TestExportImport(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
             args += ["--split-manifest"]
         if use_folder_prefix:
             args += ["--use-folder-prefix"]
+        if data_only:
+            args += ["--data-only"]
 
         call_command(*args)
 
@@ -794,3 +804,39 @@ class TestExportImport(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
             manifest = self._do_export(use_filename_format=True)
             for obj in manifest:
                 self.assertNotEqual(obj["model"], "auditlog.logentry")
+
+    def test_export_data_only(self):
+        """
+        GIVEN:
+            - Request to export documents with data only
+        WHEN:
+            - Export command is called
+        THEN:
+            - No document files are exported
+            - Manifest and version are exported
+        """
+
+        shutil.rmtree(self.dirs.media_dir / "documents")
+        shutil.copytree(
+            self.SAMPLE_DIR / "documents",
+            self.dirs.media_dir / "documents",
+        )
+
+        _ = self._do_export(data_only=True)
+
+        # Manifest and version files only should be present in the exported directory
+        self.assertFileCountInDir(self.target, 2)
+        self.assertIsFile(self.target / "manifest.json")
+        self.assertIsFile(self.target / "version.json")
+
+        shutil.rmtree(self.dirs.media_dir / "documents")
+        Document.objects.all().delete()
+
+        call_command(
+            "document_importer",
+            "--no-progress-bar",
+            "--data-only",
+            self.target,
+        )
+
+        self.assertEqual(Document.objects.all().count(), 4)

src/documents/tests/test_management_importer.py

@@ -14,9 +14,15 @@ from documents.settings import EXPORTER_ARCHIVE_NAME
 from documents.settings import EXPORTER_FILE_NAME
 from documents.tests.utils import DirectoriesMixin
 from documents.tests.utils import FileSystemAssertsMixin
+from documents.tests.utils import SampleDirMixin
 
 
-class TestCommandImport(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
+class TestCommandImport(
+    DirectoriesMixin,
+    FileSystemAssertsMixin,
+    SampleDirMixin,
+    TestCase,
+):
     def test_check_manifest_exists(self):
         """
         GIVEN:
@@ -120,14 +126,14 @@ class TestCommandImport(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
                 },
             ]
             with self.assertRaises(CommandError) as cm:
-                cmd._check_manifest_valid()
+                cmd._check_manifest_files_valid()
                 self.assertInt("Failed to read from original file", str(cm.exception))
 
             original_path.chmod(0o444)
             archive_path.chmod(0o222)
 
             with self.assertRaises(CommandError) as cm:
-                cmd._check_manifest_valid()
+                cmd._check_manifest_files_valid()
                 self.assertInt("Failed to read from archive file", str(cm.exception))
 
     def test_import_source_not_existing(self):

src/documents/tests/utils.py

@@ -156,10 +156,6 @@ class DirectoriesMixin:
     they are cleaned up on exit
     """
 
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        self.dirs = None
-
     def setUp(self) -> None:
         self.dirs = setup_directories()
         super().setUp()
@@ -200,6 +196,16 @@ class FileSystemAssertsMixin:
 
         self.assertEqual(hash1, hash2, "File SHA256 mismatch")
 
+    def assertFileCountInDir(self, path: Union[PathLike, str], count: int):
+        path = Path(path).resolve()
+        self.assertTrue(path.is_dir(), f"Path {path} is not a directory")
+        files = [x for x in path.iterdir() if x.is_file()]
+        self.assertEqual(
+            len(files),
+            count,
+            f"Path {path} contains {len(files)} files instead of {count} files",
+        )
+
 
 class ConsumerProgressMixin:
     """