mime type handling

src/documents/tests/test_api.py

@@ -45,7 +45,7 @@ class DocumentApiTest(APITestCase):
         dt = DocumentType.objects.create(name="dt", pk=63)
         tag = Tag.objects.create(name="t", pk=85)
 
-        doc = Document.objects.create(title="WOW", content="the content", correspondent=c, document_type=dt, checksum="123")
+        doc = Document.objects.create(title="WOW", content="the content", correspondent=c, document_type=dt, checksum="123", mime_type="application/pdf")
 
         doc.tags.add(tag)
 
@@ -95,7 +95,7 @@ class DocumentApiTest(APITestCase):
         with open(filename, "wb") as f:
             f.write(content)
 
-        doc = Document.objects.create(title="none", filename=os.path.basename(filename), file_type="pdf")
+        doc = Document.objects.create(title="none", filename=os.path.basename(filename), mime_type="application/pdf")
 
         with open(os.path.join(self.thumbnail_dir, "{:07d}.png".format(doc.pk)), "wb") as f:
             f.write(content_thumbnail)
@@ -117,7 +117,7 @@ class DocumentApiTest(APITestCase):
 
     def test_document_actions_not_existing_file(self):
 
-        doc = Document.objects.create(title="none", filename=os.path.basename("asd"), file_type="pdf")
+        doc = Document.objects.create(title="none", filename=os.path.basename("asd"), mime_type="application/pdf")
 
         response = self.client.get('/api/documents/{}/download/'.format(doc.pk))
         self.assertEqual(response.status_code, 404)
@@ -130,9 +130,9 @@ class DocumentApiTest(APITestCase):
 
     def test_document_filters(self):
 
-        doc1 = Document.objects.create(title="none1", checksum="A")
-        doc2 = Document.objects.create(title="none2", checksum="B")
-        doc3 = Document.objects.create(title="none3", checksum="C")
+        doc1 = Document.objects.create(title="none1", checksum="A", mime_type="application/pdf")
+        doc2 = Document.objects.create(title="none2", checksum="B", mime_type="application/pdf")
+        doc3 = Document.objects.create(title="none3", checksum="C", mime_type="application/pdf")
 
         tag_inbox = Tag.objects.create(name="t1", is_inbox_tag=True)
         tag_2 = Tag.objects.create(name="t2")

src/documents/tests/test_consumer.py

@@ -437,6 +437,18 @@ class FaultyParser(DocumentParser):
         raise ParseError("Does not compute.")
 
 
+def fake_magic_from_file(file, mime=False):
+
+    if mime:
+        if os.path.splitext(file)[1] == ".pdf":
+            return "application/pdf"
+        else:
+            return "unknown"
+    else:
+        return "A verbose string that describes the contents of the file"
+
+
+@mock.patch("documents.consumer.magic.from_file", fake_magic_from_file)
 class TestConsumer(TestCase):
 
     def make_dummy_parser(self, path, logging_group):
@@ -462,7 +474,7 @@ class TestConsumer(TestCase):
         m = patcher.start()
         m.return_value = [(None, {
             "parser": self.make_dummy_parser,
-            "test": lambda _: True,
+            "mime_types": ["application/pdf"],
             "weight": 0
         })]
 
@@ -592,7 +604,7 @@ class TestConsumer(TestCase):
     def testFaultyParser(self, m):
         m.return_value = [(None, {
             "parser": self.make_faulty_parser,
-            "test": lambda _: True,
+            "mime_types": ["application/pdf"],
             "weight": 0
         })]
 

src/documents/tests/test_document_model.py

@@ -13,9 +13,12 @@ class TestDocument(TestCase):
             title="Title",
             content="content",
             checksum="checksum",
+            mime_type="application/pdf"
         )
+
         file_path = document.source_path
         thumb_path = document.thumbnail_path
+
         with mock.patch("documents.signals.handlers.os.unlink") as mock_unlink:
             document.delete()
             mock_unlink.assert_any_call(file_path)

src/documents/tests/test_file_handling.py

@@ -31,7 +31,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="")
     def test_generate_source_filename(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -44,7 +44,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{correspondent}/{correspondent}")
     def test_file_renaming(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -81,7 +81,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{correspondent}/{correspondent}")
     def test_file_renaming_missing_permissions(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -111,10 +111,10 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{correspondent}/{correspondent}")
     def test_file_renaming_database_error(self):
 
-        document1 = Document.objects.create(file_type="pdf", storage_type=Document.STORAGE_TYPE_UNENCRYPTED, checksum="AAAAA")
+        document1 = Document.objects.create(mime_type="application/pdf", storage_type=Document.STORAGE_TYPE_UNENCRYPTED, checksum="AAAAA")
 
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.checksum = "BBBBB"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
@@ -149,7 +149,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{correspondent}/{correspondent}")
     def test_document_delete(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -170,7 +170,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{correspondent}/{correspondent}")
     def test_document_delete_nofile(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -179,7 +179,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{correspondent}/{correspondent}")
     def test_directory_not_empty(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -206,7 +206,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{tags[type]}")
     def test_tags_with_underscore(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -222,7 +222,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{tags[type]}")
     def test_tags_with_dash(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -238,7 +238,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{tags[type]}")
     def test_tags_malformed(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -254,7 +254,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{tags[0]}")
     def test_tags_all(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -269,7 +269,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{tags[1]}")
     def test_tags_out_of_bounds(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -284,7 +284,7 @@ class TestDate(TestCase):
     @override_settings(PAPERLESS_FILENAME_FORMAT="{correspondent}/{correspondent}/{correspondent}")
     def test_nested_directory_cleanup(self):
         document = Document()
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
         document.save()
 
@@ -309,7 +309,7 @@ class TestDate(TestCase):
     def test_format_none(self):
         document = Document()
         document.pk = 1
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
 
         self.assertEqual(generate_filename(document), "0000001.pdf")
@@ -335,7 +335,7 @@ class TestDate(TestCase):
     def test_invalid_format(self):
         document = Document()
         document.pk = 1
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
 
         self.assertEqual(generate_filename(document), "0000001.pdf")
@@ -344,7 +344,7 @@ class TestDate(TestCase):
     def test_invalid_format_key(self):
         document = Document()
         document.pk = 1
-        document.file_type = "pdf"
+        document.mime_type = "application/pdf"
         document.storage_type = Document.STORAGE_TYPE_UNENCRYPTED
 
         self.assertEqual(generate_filename(document), "0000001.pdf")

src/documents/tests/test_matchables.py

@@ -213,7 +213,7 @@ class TestDocumentConsumptionFinishedSignal(TestCase):
         TestCase.setUp(self)
         User.objects.create_user(username='test_consumer', password='12345')
         self.doc_contains = Document.objects.create(
-            content="I contain the keyword.", file_type="pdf")
+            content="I contain the keyword.", mime_type="application/pdf")
 
     def test_tag_applied_any(self):
         t1 = Tag.objects.create(

src/documents/tests/test_parsers.py

@@ -1,3 +1,4 @@
+import os
 from tempfile import TemporaryDirectory
 from unittest import mock
 
@@ -5,7 +6,18 @@ from django.test import TestCase
 
 from documents.parsers import get_parser_class
 
+def fake_magic_from_file(file, mime=False):
 
+    if mime:
+        if os.path.splitext(file)[1] == ".pdf":
+            return "application/pdf"
+        else:
+            return "unknown"
+    else:
+        return "A verbose string that describes the contents of the file"
+
+
+@mock.patch("documents.parsers.magic.from_file", fake_magic_from_file)
 class TestParserDiscovery(TestCase):
 
     @mock.patch("documents.parsers.document_consumer_declaration.send")
@@ -14,7 +26,7 @@ class TestParserDiscovery(TestCase):
             pass
 
         m.return_value = (
-            (None, {"weight": 0, "parser": DummyParser, "test": lambda _: True}),
+            (None, {"weight": 0, "parser": DummyParser, "mime_types": ["application/pdf"]}),
         )
 
         self.assertEqual(
@@ -32,8 +44,8 @@ class TestParserDiscovery(TestCase):
             pass
 
         m.return_value = (
-            (None, {"weight": 0, "parser": DummyParser1, "test": lambda _: True}),
-            (None, {"weight": 1, "parser": DummyParser2, "test": lambda _: True}),
+            (None, {"weight": 0, "parser": DummyParser1, "mime_types": ["application/pdf"]}),
+            (None, {"weight": 1, "parser": DummyParser2, "mime_types": ["application/pdf"]}),
         )
 
         self.assertEqual(