From 93009c1eeda1cc339dbf3acb0882dc0313c8a12f Mon Sep 17 00:00:00 2001
From: Dennis Brakhane <brakhane@gmail.com>
Date: Tue, 11 Jul 2023 16:41:31 +0200
Subject: [PATCH] Don't consider better OCR as failing
Tesseract 5.3.0 does a better job at OCR, and correctly
reads "a webp" instead of "awebp", this is good, so we
don't want the test to fail.
---
src/paperless_tesseract/tests/test_parser.py | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/src/paperless_tesseract/tests/test_parser.py b/src/paperless_tesseract/tests/test_parser.py
index 7850ad4ef..8b3de5615 100644
--- a/src/paperless_tesseract/tests/test_parser.py
+++ b/src/paperless_tesseract/tests/test_parser.py
@@ -861,8 +861,9 @@ class TestParserFileTypes(DirectoriesMixin, FileSystemAssertsMixin, TestCase):
parser = RasterisedDocumentParser(None)
parser.parse(os.path.join(self.SAMPLE_FILES, "document.webp"), "image/webp")
self.assertIsFile(parser.archive_path)
- # OCR consistent mangles this space, oh well
- self.assertIn(
- "this is awebp document, created 11/14/2022.",
+ # Older tesseracts consistently mangle the space between "a webp",
+ # tesseract 5.3.0 seems to do a better job, so we're accepting both
+ self.assertRegex(
parser.get_text().lower(),
+ r"this is a ?webp document, created 11/14/2022.",
)