Spaces:
Sleeping
Sleeping
import tempfile | |
import logging | |
import os | |
import asyncio | |
from moviepy.editor import * | |
import edge_tts | |
import gradio as gr | |
from pydub import AudioSegment | |
# Configuraci贸n de Logs | |
logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s") | |
# CONSTANTES DE ARCHIVOS | |
INTRO_VIDEO = "introvideo.mp4" | |
OUTRO_VIDEO = "outrovideo.mp4" | |
MUSIC_BG = "musicafondo.mp3" | |
EJEMPLO_VIDEO = "ejemplo.mp4" | |
# Validar existencia de archivos | |
for file in [INTRO_VIDEO, OUTRO_VIDEO, MUSIC_BG, EJEMPLO_VIDEO]: | |
if not os.path.exists(file): | |
logging.error(f"Falta archivo necesario: {file}") | |
raise FileNotFoundError(f"Falta: {file}") | |
# Configuraci贸n de chunks | |
CHUNK_SIZE = 60 # 1 minuto por chunk | |
MAX_CHUNKS = 50 | |
def eliminar_archivo_tiempo(ruta, delay=1800): | |
def eliminar(): | |
try: | |
if os.path.exists(ruta): | |
os.remove(ruta) | |
logging.info(f"Archivo eliminado: {ruta}") | |
except Exception as e: | |
logging.error(f"Error al eliminar {ruta}: {e}") | |
from threading import Timer | |
Timer(delay, eliminar).start() | |
async def generar_tts(texto, voz, duracion_total): | |
try: | |
logging.info("Generando TTS") | |
communicate = edge_tts.Communicate(texto, voz) | |
with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_tts: | |
await communicate.save(tmp_tts.name) | |
tts_audio = AudioFileClip(tmp_tts.name) | |
# Asegurar que el TTS no exceda la duraci贸n del video | |
if tts_audio.duration > duracion_total: | |
tts_audio = tts_audio.subclip(0, duracion_total) | |
return tts_audio, tmp_tts.name | |
except Exception as e: | |
logging.error(f"Fallo en TTS: {str(e)}") | |
raise | |
def crear_musica_fondo(duracion_total): | |
bg_music = AudioSegment.from_mp3(MUSIC_BG) | |
needed_ms = int(duracion_total * 1000) | |
repeticiones = needed_ms // len(bg_music) + 1 | |
bg_music = bg_music * repeticiones | |
bg_music = bg_music[:needed_ms].fade_out(1000) | |
with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_bg: | |
bg_music.export(tmp_bg.name, format="mp3") | |
return AudioFileClip(tmp_bg.name).volumex(0.15), tmp_bg.name | |
async def procesar_video(video_input, texto_tts, voz_seleccionada): | |
temp_files = [] | |
intro, outro, video_original = None, None, None | |
try: | |
logging.info("Iniciando procesamiento") | |
video_original = VideoFileClip(video_input, target_resolution=(720, 1280)) | |
duracion_video = video_original.duration | |
# Generar TTS y m煤sica de fondo para todo el video | |
tts_audio, tts_path = await generar_tts(texto_tts, voz_seleccionada, duracion_video) | |
bg_audio, bg_path = crear_musica_fondo(duracion_video) | |
temp_files.extend([tts_path, bg_path]) | |
# Combinar audios | |
audio_original = video_original.audio.volumex(0.7) if video_original.audio else None | |
audios = [bg_audio.set_duration(duracion_video)] | |
if audio_original: | |
audios.append(audio_original) | |
audios.append(tts_audio.set_start(0).volumex(0.85)) | |
audio_final = CompositeAudioClip(audios).set_duration(duracion_video) | |
# Dividir video en chunks para procesamiento | |
chunks = [] | |
for start in range(0, int(duracion_video), CHUNK_SIZE): | |
end = min(start + CHUNK_SIZE, duracion_video) | |
chunk = video_original.subclip(start, end) | |
chunks.append(chunk) | |
# Procesar cada chunk | |
processed_clips = [] | |
for chunk in chunks: | |
processed_clips.append(chunk) | |
# Combinar chunks (sin efectos) | |
video_final = concatenate_videoclips(processed_clips, method="compose") | |
video_final = video_final.set_audio(audio_final) | |
# Agregar intro y outro | |
intro = VideoFileClip(INTRO_VIDEO, target_resolution=(720, 1280)) | |
outro = VideoFileClip(OUTRO_VIDEO, target_resolution=(720, 1280)) | |
video_final = concatenate_videoclips([intro, video_final, outro], method="compose") | |
# Renderizado final | |
with tempfile.NamedTemporaryFile(suffix=".mp4", delete=False) as tmp: | |
video_final.write_videofile( | |
tmp.name, | |
codec="libx264", | |
audio_codec="aac", | |
fps=24, | |
threads=2, | |
bitrate="3M", | |
ffmpeg_params=[ | |
"-preset", "ultrafast", | |
"-crf", "28", | |
"-movflags", "+faststart", | |
"-vf", "scale=1280:720" | |
], | |
verbose=False | |
) | |
eliminar_archivo_tiempo(tmp.name, 1800) | |
logging.info(f"Video final guardado: {tmp.name}") | |
return tmp.name | |
except Exception as e: | |
logging.error(f"Fallo general: {str(e)}") | |
raise | |
finally: | |
try: | |
if video_original: | |
video_original.close() | |
if intro: | |
intro.close() | |
if outro: | |
outro.close() | |
for file in temp_files: | |
try: | |
os.remove(file) | |
except Exception as e: | |
logging.warning(f"Error limpiando {file}: {e}") | |
except Exception as e: | |
logging.warning(f"Error al cerrar recursos: {str(e)}") | |
# Interfaz Gradio | |
with gr.Blocks() as demo: | |
gr.Markdown("# Editor de Video con IA") | |
with gr.Tab("Principal"): | |
video_input = gr.Video(label="Subir video") | |
texto_tts = gr.Textbox( | |
label="Texto para TTS", | |
lines=3, | |
placeholder="Escribe aqu铆 tu texto..." | |
) | |
voz_seleccionada = gr.Dropdown( | |
label="Voz", | |
choices=["es-ES-AlvaroNeural", "es-MX-BeatrizNeural"], | |
value="es-ES-AlvaroNeural" | |
) | |
procesar_btn = gr.Button("Generar Video") | |
video_output = gr.Video(label="Video Procesado") | |
with gr.Accordion("Ejemplos de Uso", open=False): | |
gr.Examples( | |
examples=[[EJEMPLO_VIDEO, "隆Hola! Esto es una prueba. Suscr铆bete al canal."]], | |
inputs=[video_input, texto_tts], | |
label="Ejemplos" | |
) | |
procesar_btn.click( | |
procesar_video, | |
inputs=[video_input, texto_tts, voz_seleccionada], | |
outputs=video_output | |
) | |
if __name__ == "__main__": | |
demo.queue().launch() |