Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -243,22 +243,28 @@ def create_subtitle_clips(script: str, video_width: int, video_height: int, dura
|
|
243 |
if sentence_duration < 0.5:
|
244 |
continue
|
245 |
|
246 |
-
|
247 |
-
|
248 |
-
|
249 |
-
|
250 |
-
|
251 |
-
|
252 |
-
|
253 |
-
|
254 |
-
|
255 |
-
|
|
|
|
|
|
|
|
|
|
|
256 |
)
|
257 |
-
|
258 |
-
|
259 |
-
|
260 |
-
|
261 |
-
|
|
|
262 |
current_time += sentence_duration
|
263 |
|
264 |
return clips
|
@@ -268,6 +274,8 @@ def create_subtitle_clips(script: str, video_width: int, video_height: int, dura
|
|
268 |
|
269 |
def loop_audio_to_duration(audio_clip: AudioFileClip, target_duration: float) -> AudioFileClip:
|
270 |
"""Hace loop del audio hasta alcanzar la duraci贸n objetivo"""
|
|
|
|
|
271 |
try:
|
272 |
if audio_clip.duration >= target_duration:
|
273 |
return audio_clip.subclip(0, target_duration)
|
@@ -287,11 +295,17 @@ def create_video(script_text: str, generate_script: bool, music_path: str | None
|
|
287 |
|
288 |
def normalize_clip(clip):
|
289 |
"""Normaliza un clip de video a resoluci贸n y FPS est谩ndar"""
|
290 |
-
if clip
|
291 |
-
|
292 |
-
|
293 |
-
clip
|
294 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
295 |
|
296 |
try:
|
297 |
# Paso 1: Generar o usar gui贸n
|
@@ -311,11 +325,20 @@ def create_video(script_text: str, generate_script: bool, music_path: str | None
|
|
311 |
if not generate_tts_audio(script, audio_path):
|
312 |
raise RuntimeError("Error generando el audio TTS")
|
313 |
|
314 |
-
voice_clip =
|
315 |
-
|
316 |
-
|
317 |
-
|
318 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
319 |
|
320 |
# Paso 3: Buscar y descargar videos
|
321 |
update_task_progress(task_id, "Paso 3/7: Buscando videos en Pexels...")
|
@@ -352,12 +375,19 @@ def create_video(script_text: str, generate_script: bool, music_path: str | None
|
|
352 |
clip = None
|
353 |
try:
|
354 |
clip = VideoFileClip(path)
|
|
|
|
|
|
|
355 |
# Tomar m谩ximo 8 segundos de cada clip
|
356 |
duration = min(8, clip.duration)
|
357 |
processed_clip = clip.subclip(0, duration)
|
|
|
358 |
# Normalizar el clip (resoluci贸n y FPS)
|
359 |
processed_clip = normalize_clip(processed_clip)
|
360 |
-
|
|
|
|
|
|
|
361 |
except Exception as e:
|
362 |
logger.error(f"Error procesando video {path}: {e}")
|
363 |
finally:
|
@@ -368,7 +398,15 @@ def create_video(script_text: str, generate_script: bool, music_path: str | None
|
|
368 |
raise RuntimeError("No se pudieron procesar los videos")
|
369 |
|
370 |
# Concatenar videos
|
371 |
-
base_video =
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
372 |
|
373 |
# Extender video si es m谩s corto que el audio con transiciones suaves
|
374 |
if base_video.duration < video_duration:
|
@@ -380,60 +418,95 @@ def create_video(script_text: str, generate_script: bool, music_path: str | None
|
|
380 |
for _ in range(loops_needed - 1):
|
381 |
# Crear un clip con fundido de entrada para la transici贸n
|
382 |
fade_in_clip = base_video.crossfadein(fade_duration)
|
383 |
-
|
384 |
-
|
|
|
385 |
|
386 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
387 |
|
388 |
# Asegurar que el video tenga la duraci贸n exacta del audio
|
389 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
390 |
|
391 |
# Paso 5: Componer audio final
|
392 |
update_task_progress(task_id, "Paso 5/7: Componiendo audio...")
|
|
|
|
|
393 |
if music_path and os.path.exists(music_path):
|
|
|
394 |
try:
|
395 |
music_clip = AudioFileClip(music_path)
|
396 |
-
music_clip
|
397 |
-
|
|
|
|
|
|
|
398 |
except Exception as e:
|
399 |
logger.error(f"Error con m煤sica: {e}")
|
400 |
-
|
401 |
-
|
402 |
-
|
403 |
|
404 |
# Paso 6: Agregar subt铆tulos
|
405 |
update_task_progress(task_id, "Paso 6/7: Agregando subt铆tulos...")
|
406 |
subtitle_clips = create_subtitle_clips(script, base_video.w, base_video.h, video_duration)
|
407 |
if subtitle_clips:
|
408 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
409 |
|
410 |
# Paso 7: Renderizar video final
|
411 |
update_task_progress(task_id, "Paso 7/7: Renderizando video final...")
|
412 |
-
final_video =
|
413 |
-
|
414 |
-
|
415 |
-
|
416 |
-
|
417 |
-
|
418 |
-
|
419 |
-
|
420 |
-
|
421 |
-
|
422 |
-
|
423 |
-
|
424 |
-
|
425 |
-
|
426 |
-
|
427 |
-
|
428 |
-
|
429 |
-
|
430 |
-
|
431 |
-
|
432 |
-
|
433 |
-
|
434 |
-
|
435 |
-
|
436 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
437 |
|
438 |
except Exception as e:
|
439 |
logger.error(f"Error creando video: {e}")
|
|
|
243 |
if sentence_duration < 0.5:
|
244 |
continue
|
245 |
|
246 |
+
try:
|
247 |
+
txt_clip = (
|
248 |
+
TextClip(
|
249 |
+
sentence,
|
250 |
+
fontsize=max(20, int(video_height * 0.05)),
|
251 |
+
color="white",
|
252 |
+
stroke_color="black",
|
253 |
+
stroke_width=2,
|
254 |
+
method="caption",
|
255 |
+
size=(int(video_width * 0.9), None),
|
256 |
+
font="Arial-Bold"
|
257 |
+
)
|
258 |
+
.set_start(current_time)
|
259 |
+
.set_duration(sentence_duration)
|
260 |
+
.set_position(("center", "bottom"))
|
261 |
)
|
262 |
+
if txt_clip is not None:
|
263 |
+
clips.append(txt_clip)
|
264 |
+
except Exception as e:
|
265 |
+
logger.error(f"Error creando subt铆tulo para '{sentence}': {e}")
|
266 |
+
continue
|
267 |
+
|
268 |
current_time += sentence_duration
|
269 |
|
270 |
return clips
|
|
|
274 |
|
275 |
def loop_audio_to_duration(audio_clip: AudioFileClip, target_duration: float) -> AudioFileClip:
|
276 |
"""Hace loop del audio hasta alcanzar la duraci贸n objetivo"""
|
277 |
+
if audio_clip is None:
|
278 |
+
return None
|
279 |
try:
|
280 |
if audio_clip.duration >= target_duration:
|
281 |
return audio_clip.subclip(0, target_duration)
|
|
|
295 |
|
296 |
def normalize_clip(clip):
|
297 |
"""Normaliza un clip de video a resoluci贸n y FPS est谩ndar"""
|
298 |
+
if clip is None:
|
299 |
+
return None
|
300 |
+
try:
|
301 |
+
if clip.size != TARGET_RESOLUTION:
|
302 |
+
clip = clip.resize(TARGET_RESOLUTION)
|
303 |
+
if clip.fps != TARGET_FPS:
|
304 |
+
clip = clip.set_fps(TARGET_FPS)
|
305 |
+
return clip
|
306 |
+
except Exception as e:
|
307 |
+
logger.error(f"Error normalizando clip: {e}")
|
308 |
+
return None
|
309 |
|
310 |
try:
|
311 |
# Paso 1: Generar o usar gui贸n
|
|
|
325 |
if not generate_tts_audio(script, audio_path):
|
326 |
raise RuntimeError("Error generando el audio TTS")
|
327 |
|
328 |
+
voice_clip = None
|
329 |
+
try:
|
330 |
+
voice_clip = AudioFileClip(audio_path)
|
331 |
+
if voice_clip is None:
|
332 |
+
raise RuntimeError("No se pudo cargar el clip de audio")
|
333 |
+
|
334 |
+
video_duration = voice_clip.duration
|
335 |
+
|
336 |
+
if video_duration < 1:
|
337 |
+
raise ValueError("El audio generado es demasiado corto")
|
338 |
+
except Exception as e:
|
339 |
+
if voice_clip is not None:
|
340 |
+
voice_clip.close()
|
341 |
+
raise e
|
342 |
|
343 |
# Paso 3: Buscar y descargar videos
|
344 |
update_task_progress(task_id, "Paso 3/7: Buscando videos en Pexels...")
|
|
|
375 |
clip = None
|
376 |
try:
|
377 |
clip = VideoFileClip(path)
|
378 |
+
if clip is None:
|
379 |
+
continue
|
380 |
+
|
381 |
# Tomar m谩ximo 8 segundos de cada clip
|
382 |
duration = min(8, clip.duration)
|
383 |
processed_clip = clip.subclip(0, duration)
|
384 |
+
|
385 |
# Normalizar el clip (resoluci贸n y FPS)
|
386 |
processed_clip = normalize_clip(processed_clip)
|
387 |
+
|
388 |
+
if processed_clip is not None:
|
389 |
+
video_clips.append(processed_clip)
|
390 |
+
|
391 |
except Exception as e:
|
392 |
logger.error(f"Error procesando video {path}: {e}")
|
393 |
finally:
|
|
|
398 |
raise RuntimeError("No se pudieron procesar los videos")
|
399 |
|
400 |
# Concatenar videos
|
401 |
+
base_video = None
|
402 |
+
try:
|
403 |
+
base_video = concatenate_videoclips(video_clips, method="chain")
|
404 |
+
if base_video is None:
|
405 |
+
raise RuntimeError("No se pudo concatenar los videos")
|
406 |
+
except Exception as e:
|
407 |
+
if base_video is not None:
|
408 |
+
base_video.close()
|
409 |
+
raise e
|
410 |
|
411 |
# Extender video si es m谩s corto que el audio con transiciones suaves
|
412 |
if base_video.duration < video_duration:
|
|
|
418 |
for _ in range(loops_needed - 1):
|
419 |
# Crear un clip con fundido de entrada para la transici贸n
|
420 |
fade_in_clip = base_video.crossfadein(fade_duration)
|
421 |
+
if fade_in_clip is not None:
|
422 |
+
looped_clips.append(fade_in_clip)
|
423 |
+
looped_clips.append(base_video)
|
424 |
|
425 |
+
try:
|
426 |
+
base_video = concatenate_videoclips(looped_clips)
|
427 |
+
if base_video is None:
|
428 |
+
raise RuntimeError("No se pudo extender el video")
|
429 |
+
except Exception as e:
|
430 |
+
if base_video is not None:
|
431 |
+
base_video.close()
|
432 |
+
raise e
|
433 |
|
434 |
# Asegurar que el video tenga la duraci贸n exacta del audio
|
435 |
+
try:
|
436 |
+
base_video = base_video.subclip(0, video_duration)
|
437 |
+
if base_video is None:
|
438 |
+
raise RuntimeError("No se pudo recortar el video")
|
439 |
+
except Exception as e:
|
440 |
+
if base_video is not None:
|
441 |
+
base_video.close()
|
442 |
+
raise e
|
443 |
|
444 |
# Paso 5: Componer audio final
|
445 |
update_task_progress(task_id, "Paso 5/7: Componiendo audio...")
|
446 |
+
final_audio = voice_clip
|
447 |
+
|
448 |
if music_path and os.path.exists(music_path):
|
449 |
+
music_clip = None
|
450 |
try:
|
451 |
music_clip = AudioFileClip(music_path)
|
452 |
+
if music_clip is not None:
|
453 |
+
music_clip = loop_audio_to_duration(music_clip, video_duration)
|
454 |
+
if music_clip is not None:
|
455 |
+
music_clip = music_clip.volumex(0.2)
|
456 |
+
final_audio = CompositeAudioClip([music_clip, voice_clip])
|
457 |
except Exception as e:
|
458 |
logger.error(f"Error con m煤sica: {e}")
|
459 |
+
finally:
|
460 |
+
if music_clip is not None:
|
461 |
+
music_clip.close()
|
462 |
|
463 |
# Paso 6: Agregar subt铆tulos
|
464 |
update_task_progress(task_id, "Paso 6/7: Agregando subt铆tulos...")
|
465 |
subtitle_clips = create_subtitle_clips(script, base_video.w, base_video.h, video_duration)
|
466 |
if subtitle_clips:
|
467 |
+
try:
|
468 |
+
base_video = CompositeVideoClip([base_video] + subtitle_clips)
|
469 |
+
if base_video is None:
|
470 |
+
raise RuntimeError("No se pudo agregar subt铆tulos")
|
471 |
+
except Exception as e:
|
472 |
+
logger.error(f"Error creando video con subt铆tulos: {e}")
|
473 |
+
# Continuar sin subt铆tulos si falla
|
474 |
|
475 |
# Paso 7: Renderizar video final
|
476 |
update_task_progress(task_id, "Paso 7/7: Renderizando video final...")
|
477 |
+
final_video = None
|
478 |
+
try:
|
479 |
+
final_video = base_video.set_audio(final_audio)
|
480 |
+
if final_video is None:
|
481 |
+
raise RuntimeError("No se pudo combinar video y audio")
|
482 |
+
|
483 |
+
output_path = os.path.join(RESULTS_DIR, f"video_{task_id}.mp4")
|
484 |
+
final_video.write_videofile(
|
485 |
+
output_path,
|
486 |
+
fps=TARGET_FPS,
|
487 |
+
codec="libx264",
|
488 |
+
audio_codec="aac", # Mejor calidad de audio
|
489 |
+
bitrate="8000k", # Controlar calidad de video
|
490 |
+
threads=4, # Mejor uso de CPU
|
491 |
+
preset="slow", # Mejor compresi贸n
|
492 |
+
logger=None,
|
493 |
+
verbose=False
|
494 |
+
)
|
495 |
+
|
496 |
+
return output_path
|
497 |
+
except Exception as e:
|
498 |
+
raise e
|
499 |
+
finally:
|
500 |
+
# Limpiar clips
|
501 |
+
if voice_clip is not None:
|
502 |
+
voice_clip.close()
|
503 |
+
if base_video is not None:
|
504 |
+
base_video.close()
|
505 |
+
if final_video is not None:
|
506 |
+
final_video.close()
|
507 |
+
for clip in video_clips:
|
508 |
+
if clip is not None:
|
509 |
+
clip.close()
|
510 |
|
511 |
except Exception as e:
|
512 |
logger.error(f"Error creando video: {e}")
|