Switch to moviepy 2.1 which fixes several video problems, like not ke…

…eping it veritcal

e2e-tests/cmd_test.py

@@ -85,7 +85,7 @@ def test_translations_with_tts(self, tts_engine):
                 ), "Utterance end check failed"
 
                 assert "- Bon dia. - Bé." == text_array[0], "translated text 0"
-                assert "El meu nom és Jordi Mas." == text_array[1], "translated text 1"
+                assert "Em dic Jordi Mas." == text_array[1], "translated text 1"
                 assert "Sóc de Barcelona." == text_array[2], "translated text 2"
                 assert (
                     "I m'encanta aquesta ciutat." == text_array[3]

open_dubbing/video_processing.py

@@ -17,7 +17,7 @@
 
 from typing import Final
 
-from moviepy.editor import AudioFileClip, VideoFileClip, concatenate_videoclips
+from moviepy import AudioFileClip, VideoFileClip, concatenate_videoclips
 
 _DEFAULT_FPS: Final[int] = 30
 _DEFAULT_DUBBED_VIDEO_FILE: Final[str] = "dubbed_video"
@@ -36,13 +36,13 @@ def split_audio_video(*, video_file: str, output_directory: str) -> tuple[str, s
             warnings.filterwarnings("ignore", category=UserWarning)
             audio_clip = video_clip.audio
             audio_output_file = os.path.join(output_directory, filename + "_audio.mp3")
-            audio_clip.write_audiofile(audio_output_file, verbose=False, logger=None)
-            video_clip_without_audio = video_clip.set_audio(None)
+            audio_clip.write_audiofile(audio_output_file, logger=None)
+            video_clip_without_audio = video_clip.with_audio(None)
             fps = video_clip.fps or _DEFAULT_FPS
 
             video_output_file = os.path.join(output_directory, filename + "_video.mp4")
             video_clip_without_audio.write_videofile(
-                video_output_file, codec="libx264", fps=fps, verbose=False, logger=None
+                video_output_file, codec="libx264", fps=fps, logger=None
             )
         return video_output_file, audio_output_file
 
@@ -69,8 +69,8 @@ def combine_audio_video(
             )
             audio = concatenate_videoclips([audio, silence])
         elif duration_difference < 0:
-            audio = audio.subclip(0, video.duration)
-        final_clip = video.set_audio(audio)
+            audio = audio.subclipped(0, video.duration)
+        final_clip = video.with_audio(audio)
         target_language_suffix = "_" + target_language.replace("-", "_").lower()
         dubbed_video_file = os.path.join(
             output_directory,
@@ -84,7 +84,6 @@ def combine_audio_video(
             audio_codec="aac",
             temp_audiofile="temp-audio.m4a",
             remove_temp=True,
-            verbose=False,
             logger=None,
         )
         return dubbed_video_file

requirements.txt

@@ -1,5 +1,5 @@
 numpy >= 1.17.3
-moviepy == 1.0.3
+moviepy == 2.1.1
 demucs == 4.0.1
 psutil == 6.0.0
 torch >= 2.0.0,< 2.5

tests/video_processing_test.py

@@ -20,8 +20,8 @@
 import numpy as np
 
 from moviepy.audio.AudioClip import AudioArrayClip
-from moviepy.editor import ColorClip
 from moviepy.video.compositing.CompositeVideoClip import clips_array
+from moviepy.video.VideoClip import ColorClip
 
 from open_dubbing.video_processing import VideoProcessing
 
@@ -39,15 +39,15 @@ def _create_mock_video(self, directory: str, video_duration: int = 5) -> str:
             The full path to the saved video file.
         """
         filename = os.path.join(directory, "mock_video.mp4")
-        red = ColorClip((256, 200), color=(255, 0, 0)).set_duration(video_duration)
-        green = ColorClip((256, 200), color=(0, 255, 0)).set_duration(video_duration)
-        blue = ColorClip((256, 200), color=(0, 0, 255)).set_duration(video_duration)
+        red = ColorClip((256, 200), color=(255, 0, 0), duration=video_duration)
+        green = ColorClip((256, 200), color=(0, 255, 0), duration=video_duration)
+        blue = ColorClip((256, 200), color=(0, 0, 255), duration=video_duration)
         combined_arrays = clips_array([[red, green, blue]])
         combined_arrays.fps = 30
         samples = int(44100 * video_duration)
         audio_data = np.zeros((samples, 2), dtype=np.int16)
         audio_clip = AudioArrayClip(audio_data, fps=44100)
-        final_clip = combined_arrays.set_audio(audio_clip)
+        final_clip = combined_arrays.with_audio(audio_clip)
         final_clip.write_videofile(filename, logger=None)
         return filename
 
@@ -80,11 +80,9 @@ def test_combine_audio_video(self):
             video_duration = 5
             directory = temporary_directory
             video_path = os.path.join(directory, "video.mp4")
-            red = ColorClip((256, 200), color=(255, 0, 0)).set_duration(video_duration)
-            green = ColorClip((256, 200), color=(0, 255, 0)).set_duration(
-                video_duration
-            )
-            blue = ColorClip((256, 200), color=(0, 0, 255)).set_duration(video_duration)
+            red = ColorClip((256, 200), color=(255, 0, 0), duration=video_duration)
+            green = ColorClip((256, 200), color=(0, 255, 0), duration=video_duration)
+            blue = ColorClip((256, 200), color=(0, 0, 255), duration=video_duration)
             combined_arrays = clips_array([[red, green, blue]])
             combined_arrays.fps = 30
             combined_arrays.write_videofile(video_path)