Feature: Allow user to control PIL image pixel limit (#5997)

src/paperless_tesseract/parsers.py

@@ -12,6 +12,7 @@ from PIL import Image
 from documents.parsers import DocumentParser
 from documents.parsers import ParseError
 from documents.parsers import make_thumbnail_from_pdf
+from documents.utils import maybe_override_pixel_limit
 from paperless.config import OcrConfig
 from paperless.models import ArchiveFileChoices
 from paperless.models import CleanChoices
@@ -255,6 +256,9 @@ class RasterisedDocumentParser(DocumentParser):
             ocrmypdf_args["sidecar"] = sidecar_file
 
         if self.is_image(mime_type):
+            # This may be required, depending on the known imformation
+            maybe_override_pixel_limit()
+
             dpi = self.get_dpi(input_file)
             a4_dpi = self.calculate_a4_dpi(input_file)
 

src/paperless_tesseract/tests/test_parser.py

@@ -246,7 +246,7 @@ class TestParser(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
 
         self.assertRaises(ParseError, f)
 
-    @override_settings(OCR_IMAGE_DPI=72)
+    @override_settings(OCR_IMAGE_DPI=72, MAX_IMAGE_PIXELS=0)
     def test_image_no_dpi_default(self):
         parser = RasterisedDocumentParser(None)