initial

backend/tasks/youtube_tasks.py

@@ -0,0 +1,120 @@
+import os
+import yt_dlp
+from pydub import AudioSegment
+from datetime import datetime
+
+from backend.celery_app import celery
+from backend.models import db, Question, DownloadJob, DownloadJobStatus
+from backend.services.audio_service import generate_audio_filename, get_audio_path
+
+
+@celery.task(bind=True)
+def download_youtube_audio(self, question_id, youtube_url, start_time, end_time):
+    """
+    Download and trim YouTube audio clip
+
+    Args:
+        question_id: Question ID to update
+        youtube_url: YouTube video URL
+        start_time: Start time in seconds
+        end_time: End time in seconds
+    """
+    from backend.app import create_app
+    app = create_app()
+
+    with app.app_context():
+        job = DownloadJob.query.filter_by(celery_task_id=self.request.id).first()
+        question = Question.query.get(question_id)
+
+        if not job or not question:
+            return {'success': False, 'error': 'Job or question not found'}
+
+        try:
+            # Update status to processing
+            job.status = DownloadJobStatus.PROCESSING
+            job.progress = 10
+            db.session.commit()
+
+            # Create temp and final directories
+            audio_folder = app.config['AUDIO_FOLDER']
+            os.makedirs(audio_folder, exist_ok=True)
+            temp_dir = os.path.join(audio_folder, 'temp')
+            os.makedirs(temp_dir, exist_ok=True)
+
+            # Download full audio
+            temp_filename = f"{self.request.id}_full"
+            temp_path = os.path.join(temp_dir, temp_filename)
+
+            def progress_hook(d):
+                """Update progress during download"""
+                if d['status'] == 'downloading':
+                    try:
+                        # Extract percentage from string like "50.5%"
+                        percent_str = d.get('_percent_str', '0%').strip('%')
+                        percent = float(percent_str)
+                        # Map download progress to 10-60% range
+                        progress = 10 + int(percent * 0.5)
+                        job.progress = progress
+                        db.session.commit()
+                    except:
+                        pass
+
+            ydl_opts = {
+                'format': app.config['YTDLP_FORMAT'],
+                'outtmpl': temp_path + '.%(ext)s',
+                'postprocessors': [{
+                    'key': 'FFmpegExtractAudio',
+                    'preferredcodec': 'mp3',
+                    'preferredquality': app.config['YTDLP_QUALITY'],
+                }],
+                'progress_hooks': [progress_hook],
+                'quiet': True,
+                'no_warnings': True,
+            }
+
+            with yt_dlp.YoutubeDL(ydl_opts) as ydl:
+                ydl.download([youtube_url])
+
+            job.progress = 60
+            db.session.commit()
+
+            # Find downloaded file
+            full_audio_path = temp_path + '.mp3'
+
+            # Trim audio using pydub
+            audio = AudioSegment.from_mp3(full_audio_path)
+            clip = audio[start_time * 1000:end_time * 1000]  # pydub uses milliseconds
+
+            job.progress = 80
+            db.session.commit()
+
+            # Save trimmed clip
+            final_filename = generate_audio_filename('mp3')
+            final_path = os.path.join(audio_folder, final_filename)
+            clip.export(final_path, format='mp3', bitrate='192k')
+
+            # Clean up temp file
+            if os.path.exists(full_audio_path):
+                os.remove(full_audio_path)
+
+            # Update question with audio path
+            audio_url = get_audio_path(final_filename)
+            question.audio_path = audio_url
+
+            # Update job status
+            job.status = DownloadJobStatus.COMPLETED
+            job.progress = 100
+            job.completed_at = datetime.utcnow()
+            db.session.commit()
+
+            return {
+                'success': True,
+                'audio_path': audio_url,
+                'question_id': question_id
+            }
+
+        except Exception as e:
+            job.status = DownloadJobStatus.FAILED
+            job.error_message = str(e)
+            db.session.commit()
+            return {'success': False, 'error': str(e)}