Enhancement: log when pre-check fails for documents in trash (#7355)

src/documents/consumer.py

@@ -231,7 +231,9 @@ class ConsumerError(Exception):
 
 class ConsumerStatusShortMessage(str, Enum):
     DOCUMENT_ALREADY_EXISTS = "document_already_exists"
+    DOCUMENT_ALREADY_EXISTS_IN_TRASH = "document_already_exists_in_trash"
     ASN_ALREADY_EXISTS = "asn_already_exists"
+    ASN_ALREADY_EXISTS_IN_TRASH = "asn_already_exists_in_trash"
     ASN_RANGE = "asn_value_out_of_range"
     FILE_NOT_FOUND = "file_not_found"
     PRE_CONSUME_SCRIPT_NOT_FOUND = "pre_consume_script_not_found"
@@ -321,12 +323,18 @@ class ConsumerPlugin(
             Q(checksum=checksum) | Q(archive_checksum=checksum),
         )
         if existing_doc.exists():
+            msg = ConsumerStatusShortMessage.DOCUMENT_ALREADY_EXISTS
+            log_msg = f"Not consuming {self.filename}: It is a duplicate of {existing_doc.get().title} (#{existing_doc.get().pk})."
+
+            if existing_doc.first().deleted_at is not None:
+                msg = ConsumerStatusShortMessage.DOCUMENT_ALREADY_EXISTS_IN_TRASH
+                log_msg += " Note: existing document is in the trash."
+
             if settings.CONSUMER_DELETE_DUPLICATES:
                 os.unlink(self.input_doc.original_file)
             self._fail(
-                ConsumerStatusShortMessage.DOCUMENT_ALREADY_EXISTS,
-                f"Not consuming {self.filename}: It is a duplicate of"
-                f" {existing_doc.get().title} (#{existing_doc.get().pk})",
+                msg,
+                log_msg,
             )
 
     def pre_check_directories(self):
@@ -358,12 +366,20 @@ class ConsumerPlugin(
                 f"[{Document.ARCHIVE_SERIAL_NUMBER_MIN:,}, "
                 f"{Document.ARCHIVE_SERIAL_NUMBER_MAX:,}]",
             )
-        if Document.global_objects.filter(
+        existing_asn_doc = Document.global_objects.filter(
             archive_serial_number=self.metadata.asn,
-        ).exists():
+        )
+        if existing_asn_doc.exists():
+            msg = ConsumerStatusShortMessage.ASN_ALREADY_EXISTS
+            log_msg = f"Not consuming {self.filename}: Given ASN {self.metadata.asn} already exists!"
+
+            if existing_asn_doc.first().deleted_at is not None:
+                msg = ConsumerStatusShortMessage.ASN_ALREADY_EXISTS_IN_TRASH
+                log_msg += " Note: existing document is in the trash."
+
             self._fail(
-                ConsumerStatusShortMessage.ASN_ALREADY_EXISTS,
-                f"Not consuming {self.filename}: Given ASN {self.metadata.asn} already exists!",
+                msg,
+                log_msg,
             )
 
     def run_pre_consume_script(self):

src/documents/tests/test_consumer.py

@@ -322,6 +322,18 @@ class TestConsumer(
         shutil.copy(src, dst)
         return dst
 
+    def get_test_file2(self):
+        src = (
+            Path(__file__).parent
+            / "samples"
+            / "documents"
+            / "originals"
+            / "0000002.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"
@@ -642,6 +654,47 @@ class TestConsumer(
         with self.get_consumer(self.get_test_file()) as consumer:
             consumer.run()
 
+    def testDuplicateInTrash(self):
+        with self.get_consumer(self.get_test_file()) as consumer:
+            consumer.run()
+
+        Document.objects.all().delete()
+
+        with self.get_consumer(self.get_test_file()) as consumer:
+            with self.assertRaisesMessage(ConsumerError, "document is in the trash"):
+                consumer.run()
+
+    def testAsnExists(self):
+        with self.get_consumer(
+            self.get_test_file(),
+            DocumentMetadataOverrides(asn=123),
+        ) as consumer:
+            consumer.run()
+
+        with self.get_consumer(
+            self.get_test_file2(),
+            DocumentMetadataOverrides(asn=123),
+        ) as consumer:
+            with self.assertRaisesMessage(ConsumerError, "ASN 123 already exists"):
+                consumer.run()
+
+    def testAsnExistsInTrash(self):
+        with self.get_consumer(
+            self.get_test_file(),
+            DocumentMetadataOverrides(asn=123),
+        ) as consumer:
+            consumer.run()
+
+            document = Document.objects.first()
+            document.delete()
+
+        with self.get_consumer(
+            self.get_test_file2(),
+            DocumentMetadataOverrides(asn=123),
+        ) as consumer:
+            with self.assertRaisesMessage(ConsumerError, "document is in the trash"):
+                consumer.run()
+
     @mock.patch("documents.parsers.document_consumer_declaration.send")
     def testNoParsers(self, m):
         m.return_value = []