Upload 2 files

required files for web app execution

app.py

@@ -0,0 +1,95 @@
+import os
+import asyncio
+import re
+import numpy as np
+import whisper
+import edge_tts
+import gradio as gr
+from moviepy import VideoFileClip, TextClip, CompositeVideoClip, AudioFileClip, vfx, concatenate_audioclips, AudioArrayClip, CompositeAudioClip
+
+# TikTok-style Pop-up Animation
+def apply_tiktok_animation(clip, duration=0.2):
+    def zoom(t):
+        if t < duration / 2:
+            return 0.5 + 1.4 * (t / duration)
+        elif t < duration:
+            return 1.2 - 0.4 * ((t - duration/2) / (duration/2))
+        return 1.0
+    return clip.with_effects([vfx.Resize(zoom)])
+
+async def generate_video(text, video_path, voice, voice_speed, font_size, font_color, stroke_width, stroke_color):
+    try:
+        # 1. Generate Speech
+        parts = re.split(r'\[PAUSE\s+(\d+(?:\.\d+)?)\]', text)
+        audio_clips = []
+        temp_files = []
+        for i, part in enumerate(parts):
+            if i % 2 == 0:
+                if part.strip():
+                    chunk_file = f"temp_chunk_{i}.mp3"
+                    communicate = edge_tts.Communicate(part.strip(), voice, rate=voice_speed)
+                    await communicate.save(chunk_file)
+                    audio_clips.append(AudioFileClip(chunk_file))
+                    temp_files.append(chunk_file)
+            else:
+                silence = AudioArrayClip(np.zeros((int(44100 * float(part)), 2)), fps=44100)
+                audio_clips.append(silence)
+
+        combined_audio = concatenate_audioclips(audio_clips)
+        
+        # 2. Process Video
+        video_clip = VideoFileClip(video_path)
+        if video_clip.duration < combined_audio.duration:
+            final_video_clip = video_clip.with_effects([vfx.Loop(duration=combined_audio.duration)])
+        else:
+            final_video_clip = video_clip.subclipped(0, combined_audio.duration)
+
+        if final_video_clip.audio is not None:
+            bg_audio = final_video_clip.audio.with_volume_scaled(0.5)
+            final_video_clip = final_video_clip.with_audio(CompositeAudioClip([bg_audio, combined_audio]))
+        else:
+            final_video_clip = final_video_clip.with_audio(combined_audio)
+
+        # 3. Subtitles
+        temp_audio = "temp_full.mp3"
+        combined_audio.write_audiofile(temp_audio, logger=None)
+        model = whisper.load_model("tiny") # Using 'tiny' for faster processing in free cloud tier
+        result = model.transcribe(temp_audio, word_timestamps=True)
+        
+        subtitle_clips = []
+        for segment in result['segments']:
+            for word_info in segment.get('words', []):
+                txt_clip = TextClip(
+                    text=word_info['word'].strip().upper(),
+                    font_size=font_size, color=font_color,
+                    stroke_color=stroke_color, stroke_width=stroke_width,
+                    method='caption', size=(int(final_video_clip.w * 0.9), None)
+                ).with_start(word_info['start']).with_duration(word_info['end'] - word_info['start']).with_position(('center', int(final_video_clip.h * 0.65)))
+                
+                subtitle_clips.append(apply_tiktok_animation(txt_clip))
+
+        # 4. Export
+        output_path = "output_video.mp4"
+        final_video = CompositeVideoClip([final_video_clip] + subtitle_clips)
+        final_video.write_videofile(output_path, codec="libx264", audio_codec="aac", fps=24, logger=None)
+
+        return output_path, "Done!"
+    except Exception as e:
+        return None, str(e)
+
+# UI
+with gr.Blocks(theme=gr.themes.Soft()) as demo:
+    gr.Markdown("# 🎬 tts-captions-app")
+    with gr.Row():
+        with gr.Column():
+            txt = gr.Textbox(label="Text", value="HELLO WORLD! [PAUSE 0.5] Testing Hugging Face Spaces!")
+            vid = gr.File(label="Background Video", type="filepath")
+            voice = gr.Dropdown(label="Voice", choices=["en-US-ChristopherNeural", "en-US-AriaNeural"], value="en-US-ChristopherNeural")
+            btn = gr.Button("Generate")
+        with gr.Column():
+            out_v = gr.Video()
+            out_t = gr.Textbox(label="Log")
+
+    btn.click(generate_video, [txt, vid, voice, gr.State("+20%"), gr.State(80), gr.State("white"), gr.State(10), gr.State("black")], [out_v, out_t])
+
+demo.launch()

requirements.txt

@@ -0,0 +1,7 @@
+gradio
+moviepy>=2.2.1
+edge-tts
+openai-whisper
+numpy
+torch
+asyncio