4

core/video_engine.py

@@ -11,11 +11,26 @@ TEMP_AUDIO_DIR = OUTPUTS_DIR / "temp_audio"
 AUDIO_DIR = OUTPUTS_DIR / "audio"
 BACKGROUNDS_DIR = WORKSPACE_ROOT / "backgrounds"
 
+# Droid Sans Fallback often supports Arabic/CJK
+FONT_ARABIC = "/usr/share/fonts/truetype/droid/DroidSansFallbackFull.ttf"
+# Fallback for English text
+FONT_SANS = "/usr/share/fonts/truetype/liberation/LiberationSans-Bold.ttf"
+
 def escape_ffmpeg_text(text):
-    # Escape for ffmpeg drawtext filter
+    if not text:
-    # Single quotes need special handling: ' becomes '\''
+        return ""
-    # Also colons and other special chars might need escaping depending on context
+    # Use ord() to check for backslash to avoid syntax errors with literals
-    return text.replace("'", "'\\''").replace(":", "\\:")
+    result = ""
+    for char in text:
+        if ord(char) == 92: # Backslash
+            result += "\\"
+        elif char == "'":
+            result += "'\\''"
+        elif char == ":":
+            result += ":"
+        else:
+            result += char
+    return result
 
 def ensure_bg_video(bg_path):
     """Ensures a background video exists and is not empty. Creates a placeholder if needed."""
@@ -33,6 +48,13 @@ def ensure_bg_video(bg_path):
             '-pix_fmt', 'yuv420p', str(bg_path)
         ], check=True)
 
+def get_audio_duration(file_path):
+    result = subprocess.run([
+        'ffprobe', '-v', 'error', '-show_entries', 'format=duration',
+        '-of', 'default=noprint_wrappers=1:nokey=1', str(file_path)
+    ], capture_output=True, text=True)
+    return float(result.stdout.strip())
+
 def generate_video(task_id):
     from core.models import VideoTask
     task = VideoTask.objects.get(id=task_id)
@@ -56,7 +78,7 @@ def generate_video(task_id):
         all_ayahs = data['ayahs']
         selected_ayahs = [a for a in all_ayahs if task.verse_start <= a['numberInSurah'] <= task.verse_end]
         
-        # Download audio files
+        # Download audio files and collect text
         for i, ayah in enumerate(selected_ayahs):
             audio_url = ayah['audio']
             audio_path = TEMP_AUDIO_DIR / f"task_{task.id}_ayah_{i}.mp3"
@@ -68,35 +90,64 @@ def generate_video(task_id):
             audio_files.append(str(audio_path))
             verses_data.append({
                 'text': ayah['text'],
-                'audio': str(audio_path)
+                'duration': get_audio_duration(audio_path)
             })
 
         # 2. Combine Audio
         combined_audio = AUDIO_DIR / f"task_{task.id}_full.mp3"
-        # We use a filter_complex to concat audio as it's more reliable than concat protocol for different mp3s
+        filter_complex_audio = "".join([f"[{i}:a]" for i in range(len(audio_files))]) + f"concat=n={len(audio_files)}:v=0:a=1[a]"
-        filter_complex = "".join([f"[{i}:a]" for i in range(len(audio_files))]) + f"concat=n={len(audio_files)}:v=0:a=1[a]"
         cmd = ['ffmpeg', '-y']
         for f in audio_files:
             cmd.extend(['-i', f])
-        cmd.extend(['-filter_complex', filter_complex, '-map', '[a]', str(combined_audio)])
+        cmd.extend(['-filter_complex', filter_complex_audio, '-map', '[a]', str(combined_audio)])
         subprocess.run(cmd, check=True)
 
+        total_duration = get_audio_duration(combined_audio)
+
         # 3. Generate Video with FFmpeg
         bg_video = BACKGROUNDS_DIR / (task.background_video or "nature.mp4")
         ensure_bg_video(bg_video)
 
         output_video = FINAL_VIDEO_DIR / f"reels_{task.id}.mp4"
         
-        # Escape text for drawtext
+        # Build drawtext filters
-        escaped_surah = escape_ffmpeg_text(task.surah_name)
+        drawtext_filters = []
+        
+        # 1. Surah Name at top
+        escaped_surah = escape_ffmpeg_text(task.surah_name)
+        drawtext_filters.append(
+            f"drawtext=text='{escaped_surah}':fontfile={FONT_SANS}:fontcolor={task.text_color}:fontsize=48:x=(w-text_w)/2:y=100:shadowcolor=black:shadowx=2:shadowy=2"
+        )
+        
+        # 2. Verse text in middle (sequential)
+        current_time = 0
+        for verse in verses_data:
+            escaped_text = escape_ffmpeg_text(verse['text'])
+            start = current_time
+            end = current_time + verse['duration']
+            
+            font_to_use = FONT_ARABIC if os.path.exists(FONT_ARABIC) else FONT_SANS
+            
+            drawtext_filters.append(
+                f"drawtext=text='{escaped_text}':fontfile={font_to_use}:fontcolor={task.text_color}:fontsize=42:x=(w-text_w)/2:y=(h-text_h)/2:shadowcolor=black:shadowx=2:shadowy=2:enable='between(t,{start},{end})'"
+            )
+            current_time += verse['duration']
+
+        vf_chain = [
+            "scale=720:1280:force_original_aspect_ratio=increase",
+            "crop=720:1280",
+            *drawtext_filters
+        ]
+        vf_string = ",".join(vf_chain)
 
-        # Simple FFmpeg command: loop background, add audio, trim to audio length
-        # Using a vertical aspect ratio 720x1280 (common for Reels)
         subprocess.run([
-            'ffmpeg', '-y', '-stream_loop', '-1', '-i', str(bg_video), 
+            'ffmpeg', '-y', 
-            '-i', str(combined_audio), '-shortest', '-map', '0:v:0', '-map', '1:a:0',
+            '-stream_loop', '-1', '-i', str(bg_video), 
+            '-i', str(combined_audio),
+            '-t', str(total_duration),
+            '-map', '0:v:0', '-map', '1:a:0',
             '-pix_fmt', 'yuv420p', 
-            '-vf', f"scale=720:1280:force_original_aspect_ratio=increase,crop=720:1280,drawtext=text='{escaped_surah}':fontcolor={task.text_color}:fontsize=64:x=(w-text_w)/2:y=(h-text_h)/2:shadowcolor=black:shadowx=2:shadowy=2",
+            '-vf', vf_string,
             str(output_video)
         ], check=True)