Update app.py

app.py

@@ -232,9 +232,11 @@ def process_entry(entry, i, add_voiceover, target_language):
         font="./NotoSansSC-Regular.ttf",
         method='caption',
         color='yellow',
-        font_size=int(video.h // 15),
+        stroke_color='black',  # Border color
+        stroke_width=2,  # Border thickness
+        font_size=int(video.h // 20),
         size=(int(video.w * 0.8), None)
-    ).with_start(entry["start"]).with_duration(entry["end"] - entry["start"]).with_position(('bottom')).with_opacity(0.7)
+    ).with_start(entry["start"]).with_duration(entry["end"] - entry["start"]).with_position(('bottom')).with_opacity(0.8)
 
     audio_segment = None
     if add_voiceover:
@@ -248,54 +250,36 @@ def add_transcript_voiceover(video_path, translated_json, output_path, add_voice
     """
     Add transcript and voiceover to a video, segment by segment.
     """
-    # Load the video file
     video = VideoFileClip(video_path)
+    font_path = "./NotoSansSC-Regular.ttf"
 
-    # Create text clips based on timestamps
     text_clips = []
     audio_segments = []
 
-    # Define relative font size based on video height (adjust this value as necessary)
-    subtitle_font_size = int(video.h // 15)  # Ensure it's an integer
-    max_subtitle_width = int(video.w * 0.8)  # 80% of video width
-
-    font_path = "./NotoSansSC-Regular.ttf"
-
-    for i, entry in enumerate(translated_json):
-        logger.debug(f"Processing entry {i}: {entry}")
-
-        # Ensure `entry` is a dictionary with keys "start", "end", and "translated"
-        if isinstance(entry, dict) and "translated" in entry:
-            # Create text clip for subtitles
-            txt_clip = TextClip(
-                text=entry["translated"], 
-                font=font_path, 
-                method='caption', 
-                color='yellow', 
-                font_size=subtitle_font_size,  
-                size=(max_subtitle_width, None)  
-            ).with_start(entry["start"]).with_duration(entry["end"] - entry["start"]).with_position(('bottom')).with_opacity(0.7)
-            text_clips.append(txt_clip)
-
-            # Generate voiceover for this segment, if needed
-            if add_voiceover:
-                segment_audio_path = f"segment_{i}_voiceover.wav"
-                generate_voiceover([entry], target_language, segment_audio_path)
-                audio_segment = AudioFileClip(segment_audio_path).set_duration(entry["end"] - entry["start"])
-                audio_segments.append(audio_segment)
-        else:
-            raise ValueError(f"Invalid entry format: {entry}")
-
-    # Combine the text clips
+    with concurrent.futures.ThreadPoolExecutor() as executor:
+        futures = [executor.submit(process_entry, entry, i, video.w, video.h, font_path, add_voiceover, target_language) 
+                   for i, entry in enumerate(translated_json)]
+
+        for future in concurrent.futures.as_completed(futures):
+            try:
+                txt_clip, audio_segment = future.result()
+                text_clips.append(txt_clip)
+                if add_voiceover and audio_segment:
+                    audio_segments.append(audio_segment)
+            except Exception as e:
+                logger.error(f"Error processing entry: {e}")
+
+    # Sort text clips and audio segments based on their start times
+    text_clips.sort(key=lambda clip: clip.start)
+    
     final_video = CompositeVideoClip([video] + text_clips)
 
-    # Concatenate all audio segments if voiceover was added
-    if add_voiceover:
-        final_audio = sum(audio_segments, AudioFileClip("silent_audio.wav"))  # Mix all audio segments
-        final_audio = final_audio.set_duration(video.duration)  # Ensure the final audio matches the video duration
+    if add_voiceover and audio_segments:
+        audio_segments.sort(key=lambda segment: segment.start)
+        final_audio = concatenate_audioclips(audio_segments)
+        final_audio = final_audio.set_duration(video.duration)
         final_video = final_video.set_audio(final_audio)
 
-    # Write the result to a file
     logger.info(f"Saving the final video to: {output_path}")
     final_video.write_videofile(output_path, codec="libx264", audio_codec="aac")