Chore: Materialise duplicated media-tree samples from the shared copies

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

src/conftest.py

@@ -181,3 +181,9 @@ def with_form_pdf_file() -> Path:
 def multi_page_images_pdf_file() -> Path:
     """Multi-page PDF of scanned images, no text layer."""
     return samples.MULTI_PAGE_IMAGES_PDF
+
+
+@pytest.fixture(scope="session")
+def thumbnail_webp_file() -> Path:
+    """Small WebP thumbnail."""
+    return samples.THUMBNAIL_WEBP

src/documents/tests/conftest.py

@@ -12,29 +12,25 @@ if TYPE_CHECKING:
     from paperless_testing.dirs import PaperlessDirs
 
 
-@pytest.fixture(scope="session")
-def document_samples_dir() -> Path:
-    """Path to the shared test sample documents."""
-    return Path(__file__).parent / "samples" / "documents"
-
-
 @pytest.fixture()
 def sample_doc(
     paperless_dirs: "PaperlessDirs",
-    document_samples_dir: Path,
+    simple_digital_pdf_file: Path,
+    multi_page_images_pdf_file: Path,
+    thumbnail_webp_file: Path,
 ) -> "Document":
     """Create a document with valid files and matching checksums."""
     with filelock.FileLock(paperless_dirs.media_lock):
         shutil.copy(
-            document_samples_dir / "originals" / "0000001.pdf",
+            simple_digital_pdf_file,
             paperless_dirs.originals_dir / "0000001.pdf",
         )
         shutil.copy(
-            document_samples_dir / "archive" / "0000001.pdf",
+            multi_page_images_pdf_file,
             paperless_dirs.archive_dir / "0000001.pdf",
         )
         shutil.copy(
-            document_samples_dir / "thumbnails" / "0000001.webp",
+            thumbnail_webp_file,
             paperless_dirs.thumbnail_dir / "0000001.webp",
         )
 

src/documents/tests/test_api_documents.py

@@ -57,6 +57,7 @@ from paperless_testing.permissions import grant_all_global
 from paperless_testing.permissions import grant_global
 from paperless_testing.permissions import grant_object
 from paperless_testing.samples import SIMPLE_DIGITAL_PDF
+from paperless_testing.samples import THUMBNAIL_WEBP
 
 
 class TestDocumentApi(DirectoriesMixin, ConsumeTaskMixin, APITestCase):
@@ -2531,13 +2532,7 @@ class TestDocumentApi(DirectoriesMixin, ConsumeTaskMixin, APITestCase):
             archive_filename="archive.pdf",
         )
 
-        source_file: Path = (
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "thumbnails"
-            / "0000001.webp"
-        )
+        source_file: Path = THUMBNAIL_WEBP
         archive_file: Path = SIMPLE_DIGITAL_PDF
 
         shutil.copy(source_file, doc.source_path)

src/documents/tests/test_api_email.py

@@ -12,6 +12,9 @@ from documents.tests.utils import SampleDirMixin
 from paperless_testing.dirs import DirectoriesMixin
 from paperless_testing.factories import UserFactory
 from paperless_testing.permissions import grant_global
+from paperless_testing.samples import MULTI_PAGE_DIGITAL_PDF
+from paperless_testing.samples import SIMPLE_DIGITAL_PDF
+from paperless_testing.samples import WITH_FORM_PDF
 
 
 class TestEmail(DirectoriesMixin, SampleDirMixin, APITestCase):
@@ -42,15 +45,15 @@ class TestEmail(DirectoriesMixin, SampleDirMixin, APITestCase):
 
         # Copy sample files to document paths (using different files to distinguish versions)
         shutil.copy(
-            self.SAMPLE_DIR / "documents" / "originals" / "0000001.pdf",
+            SIMPLE_DIGITAL_PDF,
             self.doc1.archive_path,
         )
         shutil.copy(
-            self.SAMPLE_DIR / "documents" / "originals" / "0000002.pdf",
+            MULTI_PAGE_DIGITAL_PDF,
             self.doc1.source_path,
         )
         shutil.copy(
-            self.SAMPLE_DIR / "documents" / "originals" / "0000003.pdf",
+            WITH_FORM_PDF,
             self.doc2.source_path,
         )
 

src/documents/tests/test_bulk_edit.py

@@ -27,6 +27,9 @@ from documents.permissions import set_permissions_for_objects
 from paperless_testing.dirs import DirectoriesMixin
 from paperless_testing.factories import DocumentFactory
 from paperless_testing.permissions import grant_object
+from paperless_testing.samples import MULTI_PAGE_DIGITAL_PDF
+from paperless_testing.samples import SIMPLE_DIGITAL_PDF
+from paperless_testing.samples import WITH_FORM_PDF
 
 
 class TestBulkEdit(DirectoriesMixin, TestCase):
@@ -704,47 +707,27 @@ class TestPDFActions(DirectoriesMixin, TestCase):
         super().setUp()
         sample1 = self.dirs.scratch_dir / "sample.pdf"
         shutil.copy(
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000001.pdf",
+            SIMPLE_DIGITAL_PDF,
             sample1,
         )
         sample1_archive = self.dirs.archive_dir / "sample_archive.pdf"
         shutil.copy(
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000001.pdf",
+            SIMPLE_DIGITAL_PDF,
             sample1_archive,
         )
         sample2 = self.dirs.scratch_dir / "sample2.pdf"
         shutil.copy(
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000002.pdf",
+            MULTI_PAGE_DIGITAL_PDF,
             sample2,
         )
         sample2_archive = self.dirs.archive_dir / "sample2_archive.pdf"
         shutil.copy(
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000002.pdf",
+            MULTI_PAGE_DIGITAL_PDF,
             sample2_archive,
         )
         sample3 = self.dirs.scratch_dir / "sample3.pdf"
         shutil.copy(
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000003.pdf",
+            WITH_FORM_PDF,
             sample3,
         )
         self.doc1 = Document.objects.create(
@@ -778,11 +761,7 @@ class TestPDFActions(DirectoriesMixin, TestCase):
         )
         img_doc_archive = self.dirs.archive_dir / "sample_image.pdf"
         shutil.copy(
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000001.pdf",
+            SIMPLE_DIGITAL_PDF,
             img_doc_archive,
         )
         self.img_doc = Document.objects.create(

src/documents/tests/test_consumer.py

@@ -36,6 +36,8 @@ from paperless_testing.assertions import FileSystemAssertsMixin
 from paperless_testing.dirs import DirectoriesMixin
 from paperless_testing.factories import UserFactory
 from paperless_testing.fakes.progress import FakeProgressManager
+from paperless_testing.samples import MULTI_PAGE_DIGITAL_PDF
+from paperless_testing.samples import MULTI_PAGE_IMAGES_PDF
 from paperless_testing.samples import SIMPLE_DIGITAL_PDF
 
 
@@ -113,9 +115,7 @@ class _BaseNewStyleParser:
 
 
 class DummyParser(_BaseNewStyleParser):
-    _ARCHIVE_SRC = (
-        Path(__file__).parent / "samples" / "documents" / "archive" / "0000001.pdf"
-    )
+    _ARCHIVE_SRC = MULTI_PAGE_IMAGES_PDF
 
     def parse(self, document_path, mime_type, *, produce_archive: bool = True) -> None:
         self._text = "The Text"
@@ -201,33 +201,19 @@ class TestConsumer(
         self.addCleanup(patcher.stop)
 
     def get_test_file(self):
-        src = (
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000001.pdf"
-        )
+        src = SIMPLE_DIGITAL_PDF
         dst = self.dirs.scratch_dir / "sample.pdf"
         shutil.copy(src, dst)
         return dst
 
     def get_test_file2(self):
-        src = (
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000002.pdf"
-        )
+        src = MULTI_PAGE_DIGITAL_PDF
         dst = self.dirs.scratch_dir / "sample2.pdf"
         shutil.copy(src, dst)
         return dst
 
     def get_test_archive_file(self):
-        src = (
-            Path(__file__).parent / "samples" / "documents" / "archive" / "0000001.pdf"
-        )
+        src = MULTI_PAGE_IMAGES_PDF
         dst = self.dirs.scratch_dir / "sample_archive.pdf"
         shutil.copy(src, dst)
         return dst
@@ -1604,13 +1590,7 @@ class TestConsumerRemoteOCR(
         self.addCleanup(patcher.stop)
 
     def _consume(self, *, overrides: DocumentMetadataOverrides | None = None) -> bool:
-        src = (
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000001.pdf"
-        )
+        src = SIMPLE_DIGITAL_PDF
         dst = self.dirs.scratch_dir / "sample.pdf"
         shutil.copy(src, dst)
 

src/documents/tests/test_management_exporter.py

@@ -53,6 +53,7 @@ from paperless_testing.dirs import DirectoriesMixin
 from paperless_testing.dirs import paperless_environment
 from paperless_testing.factories import DocumentBarcodeFactory
 from paperless_testing.permissions import grant_object
+from paperless_testing.samples import install_document_samples
 
 
 @pytest.mark.management
@@ -198,10 +199,7 @@ class TestExportImport(
 
     def test_exporter(self, *, use_filename_format=False) -> None:
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         num_permission_objects = Permission.objects.count()
 
@@ -305,10 +303,7 @@ class TestExportImport(
 
     def test_exporter_with_filename_format(self) -> None:
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         with override_settings(
             FILENAME_FORMAT="{created_year}/{correspondent}/{title}",
@@ -317,10 +312,7 @@ class TestExportImport(
 
     def test_exporter_includes_share_links_and_bundles(self) -> None:
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         share_link = ShareLink.objects.create(
             slug="share-link-slug",
@@ -419,10 +411,7 @@ class TestExportImport(
 
     def test_update_export_changed_time(self) -> None:
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         self._do_export()
         self.assertIsFile(self.target / "manifest.json")
@@ -458,10 +447,7 @@ class TestExportImport(
 
     def test_update_export_changed_checksum(self) -> None:
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         self._do_export()
 
@@ -488,10 +474,7 @@ class TestExportImport(
 
     def test_update_export_deleted_document(self) -> None:
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         manifest = self._do_export()
 
@@ -523,10 +506,7 @@ class TestExportImport(
     @override_settings(FILENAME_FORMAT="{title}/{correspondent}")
     def test_update_export_changed_location(self) -> None:
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         self._do_export(use_filename_format=True)
         self.assertIsFile(self.target / "wow1" / "c.pdf")
@@ -568,10 +548,7 @@ class TestExportImport(
             - Zipfile contains exported files
         """
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         args = ["document_exporter", self.target, "--zip"]
 
@@ -600,10 +577,7 @@ class TestExportImport(
             - Zipfile contains exported files
         """
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         args = ["document_exporter", self.target, "--zip", "--use-filename-format"]
 
@@ -639,10 +613,7 @@ class TestExportImport(
             - The existing file and directory in target are removed
         """
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         # Create stuff in target directory
         existing_file = self.target / "test.txt"
@@ -738,10 +709,7 @@ class TestExportImport(
             - Documents can be imported again
         """
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         manifest = self._do_export()
         has_archive = False
@@ -784,10 +752,7 @@ class TestExportImport(
             - Documents can be imported again
         """
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         manifest = self._do_export()
         has_thumbnail = False
@@ -832,10 +797,7 @@ class TestExportImport(
             - Documents can be imported again
         """
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         manifest = self._do_export(split_manifest=True)
         has_document = False
@@ -899,10 +861,7 @@ class TestExportImport(
             - Documents can be imported again
         """
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         self._do_export(use_folder_prefix=True)
 
@@ -929,10 +888,7 @@ class TestExportImport(
             - Documents can be imported again
         """
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         self._do_export(use_folder_prefix=True, split_manifest=True)
 
@@ -959,10 +915,7 @@ class TestExportImport(
         """
 
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         num_content_type_objects = ContentType.objects.count()
         num_permission_objects = Permission.objects.count()
@@ -999,10 +952,7 @@ class TestExportImport(
 
     def test_exporter_with_auditlog_disabled(self) -> None:
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         with override_settings(
             AUDIT_LOG_ENABLED=False,
@@ -1022,10 +972,7 @@ class TestExportImport(
               survive the round-trip with deleted_at preserved
         """
         shutil.rmtree(Path(self.dirs.media_dir) / "documents")
-        shutil.copytree(
-            Path(__file__).parent / "samples" / "documents",
-            Path(self.dirs.media_dir) / "documents",
-        )
+        install_document_samples(Path(self.dirs.media_dir) / "documents")
 
         # d1 has self.note and self.cfi1 attached via setUp
         self.d1.delete()
@@ -1069,10 +1016,7 @@ class TestExportImport(
         """
 
         shutil.rmtree(self.dirs.media_dir / "documents")
-        shutil.copytree(
-            self.SAMPLE_DIR / "documents",
-            self.dirs.media_dir / "documents",
-        )
+        install_document_samples(self.dirs.media_dir / "documents")
 
         _ = self._do_export(data_only=True)
 

src/documents/tests/test_sample_files.py

@@ -0,0 +1,34 @@
+import hashlib
+from pathlib import Path
+
+from paperless_testing.samples import SIMPLE_DIGITAL_PDF
+from paperless_testing.samples import install_document_samples
+
+EXPECTED_MEDIA_FILES = [
+    "originals/0000001.pdf",
+    "originals/0000002.pdf",
+    "originals/0000003.pdf",
+    "originals/0000004.pdf",
+    "originals/0000005.pdf",
+    "originals/0000006.pdf",
+    "archive/0000001.pdf",
+    "thumbnails/0000001.webp",
+    "thumbnails/0000002.webp",
+    "thumbnails/0000003.webp",
+    "thumbnails/0000004.webp",
+]
+
+
+def _sha(path: Path) -> str:
+    return hashlib.sha256(path.read_bytes()).hexdigest()
+
+
+def test_install_recreates_every_historical_opaque_name(tmp_path: Path) -> None:
+    install_document_samples(tmp_path)
+    for relpath in EXPECTED_MEDIA_FILES:
+        assert (tmp_path / relpath).is_file(), relpath
+
+
+def test_install_uses_the_shared_bytes(tmp_path: Path) -> None:
+    install_document_samples(tmp_path)
+    assert _sha(tmp_path / "originals" / "0000001.pdf") == _sha(SIMPLE_DIGITAL_PDF)

src/documents/tests/test_tasks.py

@@ -21,6 +21,7 @@ from documents.tests.helpers import dummy_preprocess
 from paperless_testing.assertions import FileSystemAssertsMixin
 from paperless_testing.dirs import DirectoriesMixin
 from paperless_testing.factories import DocumentFactory
+from paperless_testing.samples import SIMPLE_DIGITAL_PDF
 
 
 @pytest.mark.django_db
@@ -229,20 +230,12 @@ class TestUpdateContent(DirectoriesMixin, TestCase):
         """
         sample1 = self.dirs.scratch_dir / "sample.pdf"
         shutil.copy(
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000001.pdf",
+            SIMPLE_DIGITAL_PDF,
             sample1,
         )
         sample1_archive = self.dirs.archive_dir / "sample_archive.pdf"
         shutil.copy(
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000001.pdf",
+            SIMPLE_DIGITAL_PDF,
             sample1_archive,
         )
         doc = Document.objects.create(
@@ -270,11 +263,7 @@ class TestUpdateContent(DirectoriesMixin, TestCase):
         """
         sample1 = self.dirs.scratch_dir / "sample.pdf"
         shutil.copy(
-            Path(__file__).parent
-            / "samples"
-            / "documents"
-            / "originals"
-            / "0000001.pdf",
+            SIMPLE_DIGITAL_PDF,
             sample1,
         )
         doc = Document.objects.create(

src/paperless_testing/samples.py

@@ -4,6 +4,7 @@ A sample used by a single app stays in that app's ``tests/samples`` tree.
 Only a file that two or more apps need lives here, exactly once.
 """
 
+import shutil
 from pathlib import Path
 
 SHARED_SAMPLES_DIR = (Path(__file__).parent / "sample_files").resolve()
@@ -12,3 +13,35 @@ SIMPLE_DIGITAL_PDF = SHARED_SAMPLES_DIR / "simple-digital.pdf"
 MULTI_PAGE_DIGITAL_PDF = SHARED_SAMPLES_DIR / "multi-page-digital.pdf"
 WITH_FORM_PDF = SHARED_SAMPLES_DIR / "with-form.pdf"
 MULTI_PAGE_IMAGES_PDF = SHARED_SAMPLES_DIR / "multi-page-images.pdf"
+THUMBNAIL_WEBP = SHARED_SAMPLES_DIR / "thumbnail.webp"
+
+# The fake media tree under documents/tests/samples/documents keeps only the
+# files unique to it. The files that duplicate a shared sample are laid back
+# down under their historical opaque names (DB rows and exporter manifests
+# refer to them) by install_document_samples().
+_DOCUMENT_SAMPLES_DIR = (
+    Path(__file__).parent.parent / "documents" / "tests" / "samples" / "documents"
+).resolve()
+
+_SHARED_AS_MEDIA = (
+    ("originals/0000001.pdf", SIMPLE_DIGITAL_PDF),
+    ("originals/0000002.pdf", MULTI_PAGE_DIGITAL_PDF),
+    ("originals/0000003.pdf", WITH_FORM_PDF),
+    ("archive/0000001.pdf", MULTI_PAGE_IMAGES_PDF),
+    ("thumbnails/0000001.webp", THUMBNAIL_WEBP),
+    ("thumbnails/0000002.webp", THUMBNAIL_WEBP),
+    ("thumbnails/0000003.webp", THUMBNAIL_WEBP),
+    ("thumbnails/0000004.webp", THUMBNAIL_WEBP),
+)
+
+
+def install_document_samples(dest: Path) -> None:
+    """Populate ``dest`` with the full fake media tree.
+
+    Replaces ``shutil.copytree(<samples>/documents, dest)`` in tests.
+    """
+    shutil.copytree(_DOCUMENT_SAMPLES_DIR, dest, dirs_exist_ok=True)
+    for relpath, source in _SHARED_AS_MEDIA:
+        target = dest / relpath
+        target.parent.mkdir(parents=True, exist_ok=True)
+        shutil.copy2(source, target)