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/paperless_tesseract/apps.py

@@ -1,5 +1,7 @@
 from django.apps import AppConfig
 
+from paperless_tesseract.signals import tesseract_consumer_declaration
+
 
 class PaperlessTesseractConfig(AppConfig):
 
@@ -9,8 +11,6 @@ class PaperlessTesseractConfig(AppConfig):
 
         from documents.signals import document_consumer_declaration
 
-        from .signals import ConsumerDeclaration
-
-        document_consumer_declaration.connect(ConsumerDeclaration.handle)
+        document_consumer_declaration.connect(tesseract_consumer_declaration)
 
         AppConfig.ready(self)

src/paperless_tesseract/signals.py

@@ -3,21 +3,16 @@ import re
 from .parsers import RasterisedDocumentParser
 
 
-class ConsumerDeclaration:
+def tesseract_consumer_declaration(sender, **kwargs):
+    return {
+        "parser": RasterisedDocumentParser,
+        "weight": 0,
+        "test": tesseract_consumer_test
+    }
 
-    MATCHING_FILES = re.compile(r"^.*\.(pdf|jpe?g|gif|png|tiff?|pnm|bmp)$")
 
-    @classmethod
-    def handle(cls, sender, **kwargs):
-        return cls.test
+MATCHING_FILES = re.compile(r"^.*\.(pdf|jpe?g|gif|png|tiff?|pnm|bmp)$")
 
-    @classmethod
-    def test(cls, doc):
 
-        if cls.MATCHING_FILES.match(doc.lower()):
-            return {
-                "parser": RasterisedDocumentParser,
-                "weight": 0
-            }
-
-        return None
+def tesseract_consumer_test(doc):
+    return MATCHING_FILES.match(doc.lower())

src/paperless_tesseract/tests/test_signals.py

@@ -1,6 +1,6 @@
 from django.test import TestCase
 
-from ..signals import ConsumerDeclaration
+from paperless_tesseract.signals import tesseract_consumer_test
 
 
 class SignalsTestCase(TestCase):
@@ -20,7 +20,7 @@ class SignalsTestCase(TestCase):
         for prefix in prefixes:
             for suffix in suffixes:
                 name = "{}.{}".format(prefix, suffix)
-                self.assertTrue(ConsumerDeclaration.test(name))
+                self.assertTrue(tesseract_consumer_test(name))
 
     def test_test_handles_various_file_names_false(self):
 
@@ -30,7 +30,7 @@ class SignalsTestCase(TestCase):
         for prefix in prefixes:
             for suffix in suffixes:
                 name = "{}.{}".format(prefix, suffix)
-                self.assertFalse(ConsumerDeclaration.test(name))
+                self.assertFalse(tesseract_consumer_test(name))
 
-        self.assertFalse(ConsumerDeclaration.test(""))
-        self.assertFalse(ConsumerDeclaration.test("doc"))
+        self.assertFalse(tesseract_consumer_test(""))
+        self.assertFalse(tesseract_consumer_test("doc"))