Fix: run pre-flight ASN check after barcode detection

src/documents/barcodes.py

@@ -13,10 +13,10 @@ from pikepdf import Page
 from pikepdf import PasswordError
 from pikepdf import Pdf
 
+from documents.consumer import ConsumerPreflightPlugin
 from documents.converters import convert_from_tiff_to_pdf
 from documents.data_models import ConsumableDocument
 from documents.data_models import DocumentMetadataOverrides
-from documents.models import Document
 from documents.models import Tag
 from documents.plugins.base import ConsumeTaskPlugin
 from documents.plugins.base import StopConsumeTaskError
@@ -116,24 +116,6 @@ class BarcodePlugin(ConsumeTaskPlugin):
         self._tiff_conversion_done = False
         self.barcodes: list[Barcode] = []
 
-    def _apply_detected_asn(self, detected_asn: int) -> None:
-        """
-        Apply a detected ASN to metadata if allowed.
-        """
-        if (
-            self.metadata.skip_asn_if_exists
-            and Document.global_objects.filter(
-                archive_serial_number=detected_asn,
-            ).exists()
-        ):
-            logger.info(
-                f"Found ASN in barcode {detected_asn} but skipping because it already exists.",
-            )
-            return
-
-        logger.info(f"Found ASN in barcode: {detected_asn}")
-        self.metadata.asn = detected_asn
-
     def run(self) -> None:
         # Some operations may use PIL, override pixel setting if needed
         maybe_override_pixel_limit()
@@ -205,8 +187,22 @@ class BarcodePlugin(ConsumeTaskPlugin):
 
         # Update/overwrite an ASN if possible
         # After splitting, as otherwise each split document gets the same ASN
-        if self.settings.barcode_enable_asn and (located_asn := self.asn) is not None:
-            self._apply_detected_asn(located_asn)
+        if (
+            self.settings.barcode_enable_asn
+            and not self.metadata.skip_asn
+            and (located_asn := self.asn) is not None
+        ):
+            logger.info(f"Found ASN in barcode: {located_asn}")
+            self.metadata.asn = located_asn
+            # (Re-)run the preflight ASN check
+            preflight_plugin = ConsumerPreflightPlugin(
+                input_doc=self.input_doc,
+                metadata=self.metadata,
+                status_mgr=self.status_mgr,
+                base_tmp_dir=self.base_tmp_dir,
+                task_id=self.task_id,
+            )
+            preflight_plugin.pre_check_asn_value()
 
     def cleanup(self) -> None:
         self.temp_dir.cleanup()

src/documents/bulk_edit.py

@@ -7,6 +7,7 @@ from pathlib import Path
 from typing import TYPE_CHECKING
 from typing import Literal
 
+from celery import chain
 from celery import chord
 from celery import group
 from celery import shared_task
@@ -37,42 +38,6 @@ if TYPE_CHECKING:
 logger: logging.Logger = logging.getLogger("paperless.bulk_edit")
 
 
-@shared_task(bind=True)
-def restore_archive_serial_numbers_task(
-    self,
-    backup: dict[int, int],
-    *args,
-    **kwargs,
-) -> None:
-    restore_archive_serial_numbers(backup)
-
-
-def release_archive_serial_numbers(doc_ids: list[int]) -> dict[int, int]:
-    """
-    Clears ASNs on documents that are about to be replaced so new documents
-    can be assigned ASNs without uniqueness collisions. Returns a backup map
-    of doc_id -> previous ASN for potential restoration.
-    """
-    qs = Document.objects.filter(
-        id__in=doc_ids,
-        archive_serial_number__isnull=False,
-    ).only("pk", "archive_serial_number")
-    backup = dict(qs.values_list("pk", "archive_serial_number"))
-    qs.update(archive_serial_number=None)
-    logger.info(f"Released archive serial numbers for documents {list(backup.keys())}")
-    return backup
-
-
-def restore_archive_serial_numbers(backup: dict[int, int]) -> None:
-    """
-    Restores ASNs using the provided backup map, intended for
-    rollback when replacement consumption fails.
-    """
-    for doc_id, asn in backup.items():
-        Document.objects.filter(pk=doc_id).update(archive_serial_number=asn)
-    logger.info(f"Restored archive serial numbers for documents {list(backup.keys())}")
-
-
 def set_correspondent(
     doc_ids: list[int],
     correspondent: Correspondent,
@@ -421,7 +386,6 @@ def merge(
 
     merged_pdf = pikepdf.new()
     version: str = merged_pdf.pdf_version
-    handoff_asn: int | None = None
     # use doc_ids to preserve order
     for doc_id in doc_ids:
         doc = qs.get(id=doc_id)
@@ -437,8 +401,6 @@ def merge(
                 version = max(version, pdf.pdf_version)
                 merged_pdf.pages.extend(pdf.pages)
             affected_docs.append(doc.id)
-            if handoff_asn is None and doc.archive_serial_number is not None:
-                handoff_asn = doc.archive_serial_number
         except Exception as e:
             logger.exception(
                 f"Error merging document {doc.id}, it will not be included in the merge: {e}",
@@ -464,8 +426,6 @@ def merge(
                 DocumentMetadataOverrides.from_document(metadata_document)
             )
             overrides.title = metadata_document.title + " (merged)"
-            if metadata_document.archive_serial_number is not None:
-                handoff_asn = metadata_document.archive_serial_number
         else:
             overrides = DocumentMetadataOverrides()
     else:
@@ -473,11 +433,8 @@ def merge(
 
     if user is not None:
         overrides.owner_id = user.id
-    if not delete_originals:
-        overrides.skip_asn_if_exists = True
-
-    if delete_originals and handoff_asn is not None:
-        overrides.asn = handoff_asn
+    # Avoid copying or detecting ASN from merged PDFs to prevent collision
+    overrides.skip_asn = True
 
     logger.info("Adding merged document to the task queue.")
 
@@ -490,20 +447,12 @@ def merge(
     )
 
     if delete_originals:
-        backup = release_archive_serial_numbers(affected_docs)
         logger.info(
             "Queueing removal of original documents after consumption of merged document",
         )
-        try:
-            consume_task.apply_async(
-                link=[delete.si(affected_docs)],
-                link_error=[restore_archive_serial_numbers_task.s(backup)],
-            )
-        except Exception:
-            restore_archive_serial_numbers(backup)
-            raise
-        else:
-            consume_task.delay()
+        chain(consume_task, delete.si(affected_docs)).delay()
+    else:
+        consume_task.delay()
 
     return "OK"
 
@@ -545,8 +494,6 @@ def split(
                 overrides.title = f"{doc.title} (split {idx + 1})"
                 if user is not None:
                     overrides.owner_id = user.id
-                if not delete_originals:
-                    overrides.skip_asn_if_exists = True
                 logger.info(
                     f"Adding split document with pages {split_doc} to the task queue.",
                 )
@@ -561,20 +508,10 @@ def split(
                 )
 
             if delete_originals:
-                backup = release_archive_serial_numbers([doc.id])
                 logger.info(
                     "Queueing removal of original document after consumption of the split documents",
                 )
-                try:
-                    chord(
-                        header=consume_tasks,
-                        body=delete.si([doc.id]),
-                    ).apply_async(
-                        link_error=[restore_archive_serial_numbers_task.s(backup)],
-                    )
-                except Exception:
-                    restore_archive_serial_numbers(backup)
-                    raise
+                chord(header=consume_tasks, body=delete.si([doc.id])).delay()
             else:
                 group(consume_tasks).delay()
 
@@ -677,10 +614,7 @@ def edit_pdf(
             )
             if user is not None:
                 overrides.owner_id = user.id
-            if not delete_original:
-                overrides.skip_asn_if_exists = True
-            if delete_original and len(pdf_docs) == 1:
-                overrides.asn = doc.archive_serial_number
+
             for idx, pdf in enumerate(pdf_docs, start=1):
                 filepath: Path = (
                     Path(tempfile.mkdtemp(dir=settings.SCRATCH_DIR))
@@ -699,17 +633,7 @@ def edit_pdf(
                 )
 
             if delete_original:
-                backup = release_archive_serial_numbers([doc.id])
-                try:
-                    chord(
-                        header=consume_tasks,
-                        body=delete.si([doc.id]),
-                    ).apply_async(
-                        link_error=[restore_archive_serial_numbers_task.s(backup)],
-                    )
-                except Exception:
-                    restore_archive_serial_numbers(backup)
-                    raise
+                chord(header=consume_tasks, body=delete.si([doc.id])).delay()
             else:
                 group(consume_tasks).delay()
 

src/documents/consumer.py

@@ -696,7 +696,7 @@ class ConsumerPlugin(
                 pk=self.metadata.storage_path_id,
             )
 
-        if self.metadata.asn is not None:
+        if self.metadata.asn is not None and not self.metadata.skip_asn:
             document.archive_serial_number = self.metadata.asn
 
         if self.metadata.owner_id:
@@ -812,8 +812,8 @@ class ConsumerPreflightPlugin(
         """
         Check that if override_asn is given, it is unique and within a valid range
         """
-        if self.metadata.asn is None:
-            # if ASN is None
+        if self.metadata.skip_asn or self.metadata.asn is None:
+            # if skip is set or ASN is None
             return
         # Validate the range is above zero and less than uint32_t max
         # otherwise, Whoosh can't handle it in the index

src/documents/data_models.py

@@ -30,7 +30,7 @@ class DocumentMetadataOverrides:
     change_users: list[int] | None = None
     change_groups: list[int] | None = None
     custom_fields: dict | None = None
-    skip_asn_if_exists: bool = False
+    skip_asn: bool = False
 
     def update(self, other: "DocumentMetadataOverrides") -> "DocumentMetadataOverrides":
         """
@@ -50,8 +50,8 @@ class DocumentMetadataOverrides:
             self.storage_path_id = other.storage_path_id
         if other.owner_id is not None:
             self.owner_id = other.owner_id
-        if other.skip_asn_if_exists:
-            self.skip_asn_if_exists = True
+        if other.skip_asn:
+            self.skip_asn = True
 
         # merge
         if self.tag_ids is None:

src/documents/tests/test_barcodes.py

@@ -11,6 +11,7 @@ from django.test import override_settings
 
 from documents import tasks
 from documents.barcodes import BarcodePlugin
+from documents.consumer import ConsumerError
 from documents.data_models import ConsumableDocument
 from documents.data_models import DocumentMetadataOverrides
 from documents.data_models import DocumentSource
@@ -93,6 +94,41 @@ class TestBarcode(
 
             self.assertDictEqual(separator_page_numbers, {1: False})
 
+    @override_settings(CONSUMER_ENABLE_ASN_BARCODE=True)
+    def test_asn_barcode_duplicate_in_trash_fails(self):
+        """
+        GIVEN:
+            - A document with ASN barcode 123 is in the trash
+        WHEN:
+            - A file with the same barcode ASN is consumed
+        THEN:
+            - The ASN check is re-run and consumption fails
+        """
+        test_file = self.BARCODE_SAMPLE_DIR / "barcode-39-asn-123.pdf"
+
+        first_doc = Document.objects.create(
+            title="First ASN 123",
+            content="",
+            checksum="asn123first",
+            mime_type="application/pdf",
+            archive_serial_number=123,
+        )
+
+        first_doc.delete()
+
+        dupe_asn = settings.SCRATCH_DIR / "barcode-39-asn-123-second.pdf"
+        shutil.copy(test_file, dupe_asn)
+
+        with mock.patch("documents.tasks.ProgressManager", DummyProgressManager):
+            with self.assertRaisesRegex(ConsumerError, r"ASN 123.*trash"):
+                tasks.consume_file(
+                    ConsumableDocument(
+                        source=DocumentSource.ConsumeFolder,
+                        original_file=dupe_asn,
+                    ),
+                    None,
+                )
+
     @override_settings(
         CONSUMER_BARCODE_TIFF_SUPPORT=True,
     )

src/documents/tests/test_bulk_edit.py

@@ -602,21 +602,23 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             expected_filename,
         )
         self.assertEqual(consume_file_args[1].title, None)
-        # No metadata_document_id, delete_originals False, so ASN should be None
-        self.assertIsNone(consume_file_args[1].asn)
+        self.assertTrue(consume_file_args[1].skip_asn)
 
         # With metadata_document_id overrides
         result = bulk_edit.merge(doc_ids, metadata_document_id=metadata_document_id)
         consume_file_args, _ = mock_consume_file.call_args
         self.assertEqual(consume_file_args[1].title, "B (merged)")
         self.assertEqual(consume_file_args[1].created, self.doc2.created)
+        self.assertTrue(consume_file_args[1].skip_asn)
 
         self.assertEqual(result, "OK")
 
     @mock.patch("documents.bulk_edit.delete.si")
     @mock.patch("documents.tasks.consume_file.s")
+    @mock.patch("documents.bulk_edit.chain")
     def test_merge_and_delete_originals(
         self,
+        mock_chain,
         mock_consume_file,
         mock_delete_documents,
     ):
@@ -630,12 +632,6 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             - Document deletion task should be called
         """
         doc_ids = [self.doc1.id, self.doc2.id, self.doc3.id]
-        self.doc1.archive_serial_number = 101
-        self.doc2.archive_serial_number = 102
-        self.doc3.archive_serial_number = 103
-        self.doc1.save()
-        self.doc2.save()
-        self.doc3.save()
 
         result = bulk_edit.merge(doc_ids, delete_originals=True)
         self.assertEqual(result, "OK")
@@ -646,8 +642,7 @@ class TestPDFActions(DirectoriesMixin, TestCase):
 
         mock_consume_file.assert_called()
         mock_delete_documents.assert_called()
-        consume_sig = mock_consume_file.return_value
-        consume_sig.apply_async.assert_called_once()
+        mock_chain.assert_called_once()
 
         consume_file_args, _ = mock_consume_file.call_args
         self.assertEqual(
@@ -655,7 +650,7 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             expected_filename,
         )
         self.assertEqual(consume_file_args[1].title, None)
-        self.assertEqual(consume_file_args[1].asn, 101)
+        self.assertTrue(consume_file_args[1].skip_asn)
 
         delete_documents_args, _ = mock_delete_documents.call_args
         self.assertEqual(
@@ -663,92 +658,6 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             doc_ids,
         )
 
-        self.doc1.refresh_from_db()
-        self.doc2.refresh_from_db()
-        self.doc3.refresh_from_db()
-        self.assertIsNone(self.doc1.archive_serial_number)
-        self.assertIsNone(self.doc2.archive_serial_number)
-        self.assertIsNone(self.doc3.archive_serial_number)
-
-    @mock.patch("documents.bulk_edit.delete.si")
-    @mock.patch("documents.tasks.consume_file.s")
-    def test_merge_and_delete_originals_restore_on_failure(
-        self,
-        mock_consume_file,
-        mock_delete_documents,
-    ):
-        """
-        GIVEN:
-            - Existing documents
-        WHEN:
-            - Merge action with deleting documents is called with 1 document
-            - Error occurs when queuing consume file task
-        THEN:
-            - Archive serial numbers are restored
-        """
-        doc_ids = [self.doc1.id]
-        self.doc1.archive_serial_number = 111
-        self.doc1.save()
-        sig = mock.Mock()
-        sig.apply_async.side_effect = Exception("boom")
-        mock_consume_file.return_value = sig
-
-        with self.assertRaises(Exception):
-            bulk_edit.merge(doc_ids, delete_originals=True)
-
-        self.doc1.refresh_from_db()
-        self.assertEqual(self.doc1.archive_serial_number, 111)
-
-    @mock.patch("documents.bulk_edit.delete.si")
-    @mock.patch("documents.tasks.consume_file.s")
-    def test_merge_and_delete_originals_metadata_handoff(
-        self,
-        mock_consume_file,
-        mock_delete_documents,
-    ):
-        """
-        GIVEN:
-            - Existing documents with ASNs
-        WHEN:
-            - Merge with delete_originals=True and metadata_document_id set
-        THEN:
-            - Handoff ASN uses metadata document ASN
-        """
-        doc_ids = [self.doc1.id, self.doc2.id]
-        self.doc1.archive_serial_number = 101
-        self.doc2.archive_serial_number = 202
-        self.doc1.save()
-        self.doc2.save()
-
-        result = bulk_edit.merge(
-            doc_ids,
-            metadata_document_id=self.doc2.id,
-            delete_originals=True,
-        )
-        self.assertEqual(result, "OK")
-
-        consume_file_args, _ = mock_consume_file.call_args
-        self.assertEqual(consume_file_args[1].asn, 202)
-
-    def test_restore_archive_serial_numbers_task(self):
-        """
-        GIVEN:
-            - Existing document with no archive serial number
-        WHEN:
-            - Restore archive serial number task is called with backup data
-        THEN:
-            - Document archive serial number is restored
-        """
-        self.doc1.archive_serial_number = 444
-        self.doc1.save()
-        Document.objects.filter(pk=self.doc1.id).update(archive_serial_number=None)
-
-        backup = {self.doc1.id: 444}
-        bulk_edit.restore_archive_serial_numbers_task(backup)
-
-        self.doc1.refresh_from_db()
-        self.assertEqual(self.doc1.archive_serial_number, 444)
-
     @mock.patch("documents.tasks.consume_file.s")
     def test_merge_with_archive_fallback(self, mock_consume_file):
         """
@@ -817,7 +726,6 @@ class TestPDFActions(DirectoriesMixin, TestCase):
         self.assertEqual(mock_consume_file.call_count, 2)
         consume_file_args, _ = mock_consume_file.call_args
         self.assertEqual(consume_file_args[1].title, "B (split 2)")
-        self.assertIsNone(consume_file_args[1].asn)
 
         self.assertEqual(result, "OK")
 
@@ -842,8 +750,6 @@ class TestPDFActions(DirectoriesMixin, TestCase):
         """
         doc_ids = [self.doc2.id]
         pages = [[1, 2], [3]]
-        self.doc2.archive_serial_number = 200
-        self.doc2.save()
 
         result = bulk_edit.split(doc_ids, pages, delete_originals=True)
         self.assertEqual(result, "OK")
@@ -861,42 +767,6 @@ class TestPDFActions(DirectoriesMixin, TestCase):
             doc_ids,
         )
 
-        self.doc2.refresh_from_db()
-        self.assertIsNone(self.doc2.archive_serial_number)
-
-    @mock.patch("documents.bulk_edit.delete.si")
-    @mock.patch("documents.tasks.consume_file.s")
-    @mock.patch("documents.bulk_edit.chord")
-    def test_split_restore_on_failure(
-        self,
-        mock_chord,
-        mock_consume_file,
-        mock_delete_documents,
-    ):
-        """
-        GIVEN:
-            - Existing documents
-        WHEN:
-            - Split action with deleting documents is called with 1 document and 2 page groups
-            - Error occurs when queuing chord task
-        THEN:
-            - Archive serial numbers are restored
-        """
-        doc_ids = [self.doc2.id]
-        pages = [[1, 2]]
-        self.doc2.archive_serial_number = 222
-        self.doc2.save()
-
-        sig = mock.Mock()
-        sig.apply_async.side_effect = Exception("boom")
-        mock_chord.return_value = sig
-
-        result = bulk_edit.split(doc_ids, pages, delete_originals=True)
-        self.assertEqual(result, "OK")
-
-        self.doc2.refresh_from_db()
-        self.assertEqual(self.doc2.archive_serial_number, 222)
-
     @mock.patch("documents.tasks.consume_file.delay")
     @mock.patch("pikepdf.Pdf.save")
     def test_split_with_errors(self, mock_save_pdf, mock_consume_file):
@@ -1097,49 +967,10 @@ class TestPDFActions(DirectoriesMixin, TestCase):
         mock_chord.return_value.delay.return_value = None
         doc_ids = [self.doc2.id]
         operations = [{"page": 1}, {"page": 2}]
-        self.doc2.archive_serial_number = 250
-        self.doc2.save()
 
         result = bulk_edit.edit_pdf(doc_ids, operations, delete_original=True)
         self.assertEqual(result, "OK")
         mock_chord.assert_called_once()
-        consume_file_args, _ = mock_consume_file.call_args
-        self.assertEqual(consume_file_args[1].asn, 250)
-        self.doc2.refresh_from_db()
-        self.assertIsNone(self.doc2.archive_serial_number)
-
-    @mock.patch("documents.bulk_edit.delete.si")
-    @mock.patch("documents.tasks.consume_file.s")
-    @mock.patch("documents.bulk_edit.chord")
-    def test_edit_pdf_restore_on_failure(
-        self,
-        mock_chord,
-        mock_consume_file,
-        mock_delete_documents,
-    ):
-        """
-        GIVEN:
-            - Existing document
-        WHEN:
-            - edit_pdf is called with delete_original=True
-            - Error occurs when queuing chord task
-        THEN:
-            - Archive serial numbers are restored
-        """
-        doc_ids = [self.doc2.id]
-        operations = [{"page": 1}]
-        self.doc2.archive_serial_number = 333
-        self.doc2.save()
-
-        sig = mock.Mock()
-        sig.apply_async.side_effect = Exception("boom")
-        mock_chord.return_value = sig
-
-        with self.assertRaises(Exception):
-            bulk_edit.edit_pdf(doc_ids, operations, delete_original=True)
-
-        self.doc2.refresh_from_db()
-        self.assertEqual(self.doc2.archive_serial_number, 333)
 
     @mock.patch("documents.tasks.update_document_content_maybe_archive_file.delay")
     def test_edit_pdf_with_update_document(self, mock_update_document):

src/documents/tests/test_consumer.py

@@ -14,7 +14,6 @@ from django.test import override_settings
 from django.utils import timezone
 from guardian.core import ObjectPermissionChecker
 
-from documents.barcodes import BarcodePlugin
 from documents.consumer import ConsumerError
 from documents.data_models import DocumentMetadataOverrides
 from documents.data_models import DocumentSource
@@ -413,6 +412,14 @@ class TestConsumer(
         self.assertEqual(document.archive_serial_number, 123)
         self._assert_first_last_send_progress()
 
+    def testMetadataOverridesSkipAsnPropagation(self):
+        overrides = DocumentMetadataOverrides()
+        incoming = DocumentMetadataOverrides(skip_asn=True)
+
+        overrides.update(incoming)
+
+        self.assertTrue(overrides.skip_asn)
+
     def testOverrideTitlePlaceholders(self):
         c = Correspondent.objects.create(name="Correspondent Name")
         dt = DocumentType.objects.create(name="DocType Name")
@@ -1233,46 +1240,3 @@ class PostConsumeTestCase(DirectoriesMixin, GetConsumerMixin, TestCase):
                         r"sample\.pdf: Error while executing post-consume script: Command '\[.*\]' returned non-zero exit status \d+\.",
                     ):
                         consumer.run_post_consume_script(doc)
-
-
-class TestMetadataOverrides(TestCase):
-    def test_update_skip_asn_if_exists(self):
-        base = DocumentMetadataOverrides()
-        incoming = DocumentMetadataOverrides(skip_asn_if_exists=True)
-        base.update(incoming)
-        self.assertTrue(base.skip_asn_if_exists)
-
-
-class TestBarcodeApplyDetectedASN(TestCase):
-    """
-    GIVEN:
-        - Existing Documents with ASN 123
-    WHEN:
-        - A BarcodePlugin which detected an ASN
-    THEN:
-        - If skip_asn_if_exists is set, and ASN exists, do not set ASN
-        - If skip_asn_if_exists is set, and ASN does not exist, set ASN
-    """
-
-    def test_apply_detected_asn_skips_existing_when_flag_set(self):
-        doc = Document.objects.create(
-            checksum="X1",
-            title="D1",
-            archive_serial_number=123,
-        )
-        metadata = DocumentMetadataOverrides(skip_asn_if_exists=True)
-        plugin = BarcodePlugin(
-            input_doc=mock.Mock(),
-            metadata=metadata,
-            status_mgr=mock.Mock(),
-            base_tmp_dir=tempfile.gettempdir(),
-            task_id="test-task",
-        )
-
-        plugin._apply_detected_asn(123)
-        self.assertIsNone(plugin.metadata.asn)
-
-        doc.hard_delete()
-
-        plugin._apply_detected_asn(123)
-        self.assertEqual(plugin.metadata.asn, 123)