Changed the way parsers are discovered. This also prepares for upcoming changes regarding content types and file types: parsers should declare what they support, and actual file extensions should not be hardcoded everywhere.

src/documents/parsers.py

@@ -41,15 +41,16 @@ def get_parser_class(doc):
     Determine the appropriate parser class based on the file
     """
 
-    parsers = []
-    for response in document_consumer_declaration.send(None):
-        parsers.append(response[1])
-
     options = []
-    for parser in parsers:
-        result = parser(doc)
-        if result:
-            options.append(result)
+
+    # Sein letzter Befehl war: KOMMT! Und sie kamen. Alle. Sogar die Parser.
+
+    for response in document_consumer_declaration.send(None):
+        parser_declaration = response[1]
+        parser_test = parser_declaration["test"]
+
+        if parser_test(doc):
+            options.append(parser_declaration)
 
     if not options:
         return None

src/documents/tests/test_parsers.py

@@ -14,7 +14,7 @@ class TestParserDiscovery(TestCase):
             pass
 
         m.return_value = (
-            (None, lambda _: {"weight": 0, "parser": DummyParser}),
+            (None, {"weight": 0, "parser": DummyParser, "test": lambda _: True}),
         )
 
         self.assertEqual(
@@ -32,8 +32,8 @@ class TestParserDiscovery(TestCase):
             pass
 
         m.return_value = (
-            (None, lambda _: {"weight": 0, "parser": DummyParser1}),
-            (None, lambda _: {"weight": 1, "parser": DummyParser2}),
+            (None, {"weight": 0, "parser": DummyParser1, "test": lambda _: True}),
+            (None, {"weight": 1, "parser": DummyParser2, "test": lambda _: True}),
         )
 
         self.assertEqual(
@@ -43,7 +43,7 @@ class TestParserDiscovery(TestCase):
 
     @mock.patch("documents.parsers.document_consumer_declaration.send")
     def test__get_parser_class_0_parsers(self, m, *args):
-        m.return_value = ((None, lambda _: None),)
+        m.return_value = []
         with TemporaryDirectory() as tmpdir:
             self.assertIsNone(
                 get_parser_class("doc.pdf")