Return None When There Isn't Meaningful Letterboxing

perception/hashers/tools.py

@@ -1064,10 +1064,7 @@ def unletterbox(
 
         if len(set(corners)) == 4:
             LOGGER.debug("No common corner color detected, skipping content detection.")
-            return (
-                (0, w),
-                (0, h),
-            )  # Return full image bounds instead of None to maintain backwards compatibility
+            return None
         # Use the most common corner value as the background intensity.
         counts = Counter(corners)
         bg_gray = counts.most_common(1)[0][0]
@@ -1079,10 +1076,7 @@ def unletterbox(
     # If every pixel is classified as content, there is no border to remove.
     if content_mask.all():
         LOGGER.debug("All pixels differ from background; no letterbox detected.")
-        return (
-            (0, w),
-            (0, h),
-        )  # Return full image bounds instead of None to maintain backwards compatibility
+        return None
 
     # Find the content bounding box by projecting the mask onto rows and
     # columns. cv2.reduce is used instead of np.sum for performance.
@@ -1116,10 +1110,7 @@ def unletterbox(
             "Crop would not reduce either dimension by %.0f%%; skipping.",
             min_reduction * 100,
         )
-        return (
-            (0, w),
-            (0, h),
-        )  # Return full image bounds instead of None to maintain backwards compatibility
+        return None
     # Reject if the remaining content region is too small to be useful.
     if width < min_side_length or height < min_side_length:
         LOGGER.debug(
@@ -1156,7 +1147,7 @@ def unletterbox_crop(
         min_reduction: The minimum fraction (0–1) of the original width
             or height that must be removed for the crop to be worthwhile.
             If the crop removes less than this from both dimensions,
-            the original image is returned. Defaults to 0.02 (2%).
+            ``None`` is returned. Defaults to 0.02 (2%).
     Returns:
         The cropped image or None if the image is mostly blank space.
     """

perception/local_descriptor_deduplication.py

@@ -157,7 +157,7 @@ def validate_match(
             descriptorB["descriptors"].astype("float32"), 2
         )
         good_A2B, good_B2A = map(
-            lambda distances: (distances[:, 0] < distances[:, 1] * self.ratio),
+            lambda distances: distances[:, 0] < distances[:, 1] * self.ratio,
             [distances_A2B, distances_B2A],
         )
         match = min(
@@ -289,10 +289,9 @@ def load_and_preprocess(filepath, max_size=DEFAULT_MAX_SIZE, grayscale=True):
         LOGGER.warning("Failed to load image %s", filepath)
         return None
     res = pht.unletterbox(image)
-    if res is None:
-        return None
-    (x1, x2), (y1, y2) = res
-    image = np.ascontiguousarray(image[y1:y2, x1:x2])
+    if res is not None:
+        (x1, x2), (y1, y2) = res
+        image = np.ascontiguousarray(image[y1:y2, x1:x2])
     if grayscale:
         image = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
 

tests/test_tools.py

@@ -106,7 +106,14 @@ def test_compute_euclidean_pairwise_duplicates():
     # Use grouped files.
     counts = np.array([3, 3, 2, 2])
     expected = np.array(
-        [[2 / 3, 2 / 3], [0, 0], [0, 0], [1 / 3, 1 / 2], [0, 0], [0, 0]]
+        [
+            [2 / 3, 2 / 3],
+            [0, 0],
+            [0, 0],
+            [1 / 3, 1 / 2],
+            [0, 0],
+            [0, 0],
+        ]
     )
     actual = tools.extensions.compute_euclidean_pairwise_duplicates(
         X=X.astype("int32"),
@@ -206,12 +213,7 @@ def test_unletterbox_color():
     padded[50 : 50 + image.shape[0], 25 : 25 + image.shape[1]] = image
     # Should not unletterbox since not black.
     results = hashers.tools.unletterbox(padded, only_remove_black=True)
-    assert results is not None
-    (x1, x2), (y1, y2) = results
-    assert y1 == 0
-    assert y2 == padded.shape[0]
-    assert x1 == 0
-    assert x2 == padded.shape[1]
+    assert results is None
 
     # Should  unletterbox color:
     results = hashers.tools.unletterbox(padded, only_remove_black=False)
@@ -252,12 +254,7 @@ def test_unletterbox_noblackbars():
     image = hashers.tools.read(testing.DEFAULT_TEST_IMAGES[0])
 
     results = hashers.tools.unletterbox(image)
-    assert results is not None
-    (x1, x2), (y1, y2) = results
-    assert x1 == 0
-    assert y1 == 0
-    assert x2 == image.shape[1]
-    assert y2 == image.shape[0]
+    assert results is None
 
 
 def test_ffmpeg_video():
@@ -278,8 +275,10 @@ def test_ffmpeg_video():
             diff = np.abs(frame1.astype("int32") - frame2.astype("int32")).flatten()
             assert index1 == index2, f"Index mismatch for {filename}"
             np.testing.assert_allclose(
-                timestamp1, timestamp2
-            ), f"Timestamp mismatch for {filename}"
+                timestamp1,
+                timestamp2,
+                err_msg=f"Timestamp mismatch for {filename}",
+            )
             assert np.percentile(diff, 75) < 25, f"Frame mismatch for {filename}"