[scene_manager] Fix save_images not working with UTF-8 paths #450

Author: Breakthrough

scenedetect/scene_manager.py

@@ -86,6 +86,7 @@ def on_new_scene(frame_img: numpy.ndarray, frame_num: int):
 import sys
 import threading
 from enum import Enum
+from pathlib import Path
 from typing import Callable, Dict, Iterable, List, Optional, TextIO, Tuple, Union
 
 import cv2
@@ -587,8 +588,13 @@ def save_images(
                     frame_im = cv2.resize(
                         frame_im, (0, 0), fx=scale, fy=scale, interpolation=interpolation.value
                     )
-
-                cv2.imwrite(get_and_create_path(file_path, output_dir), frame_im, imwrite_param)
+                path = Path(get_and_create_path(file_path, output_dir))
+                (is_ok, encoded) = cv2.imencode(f".{image_extension}", frame_im, imwrite_param)
+                if is_ok:
+                    encoded.tofile(path)
+                else:
+                    logger.error(f"Failed to encode image for {file_path}")
+            #
             else:
                 completed = False
                 break

tests/test_cli.py

@@ -17,6 +17,7 @@
 from pathlib import Path
 
 import cv2
+import numpy as np
 import pytest
 
 from scenedetect.video_splitter import is_ffmpeg_available, is_mkvmerge_available
@@ -455,17 +456,35 @@ def test_cli_save_images(tmp_path: Path):
         )
         == 0
     )
+    images = [image for image in tmp_path.glob("*.jpg")]
+    # Should detect two scenes and generate 3 images per scene with above params.
+    assert len(images) == 6
     # Open one of the created images and make sure it has the correct resolution.
-    # TODO: Also need to test that the right number of images was generated, and compare with
-    # expected frames from the actual video.
-    images = glob.glob(os.path.join(tmp_path, "*.jpg"))
-    assert images
     image = cv2.imread(images[0])
     assert image.shape == (544, 1280, 3)
 
 
+def test_cli_save_images_path_handling(tmp_path: Path):
+    """Test `save-images` ability to handle UTF-8 paths."""
+    assert (
+        invoke_scenedetect(
+            "-i {VIDEO} -s {STATS} time {TIME} {DETECTOR} save-images -f %s"
+            % ("電腦檔案-$SCENE_NUMBER-$IMAGE_NUMBER"),
+            output_dir=tmp_path,
+        )
+        == 0
+    )
+    images = [image for image in tmp_path.glob("電腦檔案-*.jpg")]
+    # Should detect two scenes and generate 3 images per scene with above params.
+    assert len(images) == 6
+    # Check the created images can be read and have the correct size.
+    # We can't use `cv2.imread` here since it doesn't seem to work correctly with UTF-8 paths.
+    image = cv2.imdecode(np.fromfile(images[0], dtype=np.uint8), cv2.IMREAD_UNCHANGED)
+    assert image.shape == (544, 1280, 3)
+
+
 # TODO(#134): This works fine with OpenCV currently, but needs to be supported for PyAV and MoviePy.
-def test_cli_save_images_rotation(rotated_video_file, tmp_path):
+def test_cli_save_images_rotation(rotated_video_file, tmp_path: Path):
     """Test that `save-images` command rotates images correctly with the default backend."""
     assert (
         invoke_scenedetect(
@@ -475,8 +494,9 @@ def test_cli_save_images_rotation(rotated_video_file, tmp_path):
         )
         == 0
     )
-    images = glob.glob(os.path.join(tmp_path, "*.jpg"))
-    assert images
+    images = [image for image in tmp_path.glob("*.jpg")]
+    # Should detect two scenes and generate 3 images per scene with above params.
+    assert len(images) == 6
     image = cv2.imread(images[0])
     # Note same resolution as in test_cli_save_images but rotated 90 degrees.
     assert image.shape == (1280, 544, 3)

website/pages/changelog.md

@@ -583,19 +583,20 @@ Development
 
 ## PySceneDetect 0.6.5 (TBD)
 
- - [bugfix] Fix new detectors not working with `default-detector` config option
- - [bugfix] Fix crash when using `save-images`/`save_images()` with OpenCV backend [#455](https://github.com/Breakthrough/PySceneDetect/issues/455)
+ - [bugfix] Fix `SyntaxWarning` due to incorrect escaping [#400](https://github.com/Breakthrough/PySceneDetect/issues/400)
+ - [bugfix] Fix `ContentDetector` crash when using callbacks [#416](https://github.com/Breakthrough/PySceneDetect/issues/416) [#420](https://github.com/Breakthrough/PySceneDetect/issues/420)
+ - [feature] Add ability to configure CSV separators for rows/columns in config file [#423](https://github.com/Breakthrough/PySceneDetect/issues/423)
+ - [feature] Add new `--show` flag to `export-html` command to launch browser after processing [#442](https://github.com/Breakthrough/PySceneDetect/issues/442)
  - [general] Timecodes of the form `MM:SS[.nnn]` are now processed correctly [#443](https://github.com/Breakthrough/PySceneDetect/issues/443)
- - [feature] Add new `--show` flag to `export-html` command to launch browser after processing (#442)
+ - [bugfix] Fix `save-images`/`save_images()` not working correctly with UTF-8 paths [#450](https://github.com/Breakthrough/PySceneDetect/issues/455)
+ - [bugfix] Fix crash when using `save-images`/`save_images()` with OpenCV backend [#455](https://github.com/Breakthrough/PySceneDetect/issues/455)
+ - [bugfix] Fix new detectors not working with `default-detector` config option
  - [improvement] The `export-html` command now implicitly invokes `save-images` with default parameters
      - The output of the `export-html` command will always use the result of the `save-images` command that *precedes* it
  - [general] Updates to Windows distributions:
     - The MoviePy backend is now included with Windows distributions
     - Bundled Python interpreter is now Python 3.13
     - Updated PyAV 10 -> 13.1.0 and OpenCV 4.10.0.82 -> 4.10.0.84
  - [improvement] `save_to_csv` now works with paths from `pathlib`
- - [bugfix] Fix `SyntaxWarning` due to incorrect escaping [#400](https://github.com/Breakthrough/PySceneDetect/issues/400)
- - [bugfix] Fix `ContentDetector` crash when using callbacks [#416](https://github.com/Breakthrough/PySceneDetect/issues/416) [#420](https://github.com/Breakthrough/PySceneDetect/issues/420)
  - [api] The `save_to_csv` function now works correctly with paths from the `pathlib` module
  - [api] Add `col_separator` and `row_separator` args to `write_scene_list` function in `scenedetect.scene_manager`
- - [feature] Add ability to configure CSV separators for rows/columns in config file [#423](https://github.com/Breakthrough/PySceneDetect/issues/423)