ref without moviepy

app.py

@@ -1,5 +1,4 @@
 import logging
-# Отключаем спам от httpx
 logging.getLogger("httpx").setLevel(logging.WARNING)
 
 import gradio as gr
@@ -15,7 +14,6 @@ from sklearn.cluster import AgglomerativeClustering
 import tempfile
 import os
 import subprocess
-from moviepy.editor import VideoFileClip
 
 # Initialize Whisper with flash attention
 pipe = pipeline(
@@ -58,33 +56,37 @@ def format_time(seconds):
         return f"{minutes:02}:{seconds:02}"
 
 def extract_audio_from_video(video_path):
-    """Extract audio from video file using ffmpeg directly."""
+    """Extract audio from video file using ffmpeg."""
     try:
         # Create temporary file for audio
         temp_audio = tempfile.NamedTemporaryFile(suffix='.wav', delete=False)
         temp_audio_path = temp_audio.name
         temp_audio.close()
 
-        # Используем ffmpeg напрямую
+        # FFmpeg command to extract audio
         command = [
             'ffmpeg',
             '-i', video_path,
-            '-vn',  # Отключаем видео
-            '-acodec', 'pcm_s16le',  # Используем простой аудиокодек
-            '-ar', '16000',  # Частота дискретизации
-            '-ac', '1',  # Моно
-            '-y',  # Перезаписываем файл если существует
+            '-vn',
+            '-acodec', 'pcm_s16le',
+            '-ar', '16000',
+            '-ac', '1',
+            '-y',
             temp_audio_path
         ]
 
-        # Запускаем ffmpeg и перехватываем вывод
-        result = subprocess.run(command, capture_output=True, text=True)
+        # Run ffmpeg
+        process = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+        stdout, stderr = process.communicate()
 
-        if result.returncode != 0:
-            raise Exception(f"FFmpeg error: {result.stderr}")
+        if process.returncode != 0:
+            raise Exception(f"FFmpeg error: {stderr.decode()}")
 
-        if not os.path.exists(temp_audio_path) or os.path.getsize(temp_audio_path) == 0:
-            raise Exception("Failed to extract audio: output file is empty")
+        return temp_audio_path
+    except Exception as e:
+        if os.path.exists(temp_audio_path):
+            os.unlink(temp_audio_path)
+        raise gr.Error(f"Error extracting audio from video: {str(e)}")
 
         return temp_audio_path
     except Exception as e:

requirements.txt

@@ -1,8 +1,6 @@
 transformers[torch]
 speechbrain
 scikit-learn
-moviepy
-imageio-ffmpeg
 gradio
 numpy
 torchaudio