Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -3,8 +3,8 @@ import moviepy.editor as mp
|
|
| 3 |
import numpy as np
|
| 4 |
import librosa
|
| 5 |
from PIL import Image, ImageDraw
|
| 6 |
-
import os
|
| 7 |
import tempfile
|
|
|
|
| 8 |
import logging
|
| 9 |
|
| 10 |
# Configuraci贸n de logging
|
|
@@ -17,8 +17,11 @@ logger = logging.getLogger("audio_to_video")
|
|
| 17 |
|
| 18 |
def generate_video(audio_file, image_file):
|
| 19 |
try:
|
|
|
|
|
|
|
| 20 |
# 1. Cargar audio
|
| 21 |
-
|
|
|
|
| 22 |
duration = librosa.get_duration(y=y, sr=sr)
|
| 23 |
logger.info(f"Audio cargado: {duration:.1f} segundos")
|
| 24 |
|
|
@@ -32,10 +35,12 @@ def generate_video(audio_file, image_file):
|
|
| 32 |
audio_envelope_zoom = audio_envelope * 0.2 + 0.9 # Para zoom
|
| 33 |
audio_envelope_wave = audio_envelope * (img_h // 6) # Para waveform
|
| 34 |
|
| 35 |
-
# 4. Generar frames
|
| 36 |
def make_frame(t):
|
| 37 |
-
#
|
| 38 |
time_idx = int(t * sr)
|
|
|
|
|
|
|
| 39 |
zoom_factor = audio_envelope_zoom[time_idx] if time_idx < len(audio_envelope_zoom) else 1.0
|
| 40 |
new_size = (int(img_w * zoom_factor), int(img_h * zoom_factor))
|
| 41 |
zoomed_img = img.resize(new_size, Image.LANCZOS)
|
|
@@ -50,7 +55,7 @@ def generate_video(audio_file, image_file):
|
|
| 50 |
y_offset + img_h
|
| 51 |
))
|
| 52 |
|
| 53 |
-
# --- Waveform ---
|
| 54 |
frame = ImageDraw.Draw(cropped_img)
|
| 55 |
start_y = int(img_h * 0.8) # 80% hacia abajo
|
| 56 |
|
|
@@ -68,7 +73,7 @@ def generate_video(audio_file, image_file):
|
|
| 68 |
points.extend([(x, y_pos), (x, y_neg)])
|
| 69 |
|
| 70 |
if len(points) > 2:
|
| 71 |
-
frame.polygon(points, fill=(255, 0, 0, 150))
|
| 72 |
|
| 73 |
return np.array(cropped_img)
|
| 74 |
|
|
@@ -77,10 +82,11 @@ def generate_video(audio_file, image_file):
|
|
| 77 |
video.fps = 24
|
| 78 |
video = video.set_audio(mp.AudioFileClip(audio_file))
|
| 79 |
|
| 80 |
-
# 6. Guardar video
|
| 81 |
temp_dir = tempfile.mkdtemp()
|
| 82 |
output_path = os.path.join(temp_dir, "output.mp4")
|
| 83 |
|
|
|
|
| 84 |
video.write_videofile(
|
| 85 |
output_path,
|
| 86 |
codec="libx264",
|
|
@@ -89,12 +95,16 @@ def generate_video(audio_file, image_file):
|
|
| 89 |
logger=None
|
| 90 |
)
|
| 91 |
|
| 92 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93 |
return output_path # Retornar la ruta completa
|
| 94 |
|
| 95 |
except Exception as e:
|
| 96 |
-
logger.error(f"Error cr铆tico: {str(e)}")
|
| 97 |
-
return None
|
| 98 |
|
| 99 |
# Interfaz Gradio
|
| 100 |
iface = gr.Interface(
|
|
@@ -109,4 +119,4 @@ iface = gr.Interface(
|
|
| 109 |
)
|
| 110 |
|
| 111 |
if __name__ == "__main__":
|
| 112 |
-
iface.queue().launch()
|
|
|
|
| 3 |
import numpy as np
|
| 4 |
import librosa
|
| 5 |
from PIL import Image, ImageDraw
|
|
|
|
| 6 |
import tempfile
|
| 7 |
+
import os
|
| 8 |
import logging
|
| 9 |
|
| 10 |
# Configuraci贸n de logging
|
|
|
|
| 17 |
|
| 18 |
def generate_video(audio_file, image_file):
|
| 19 |
try:
|
| 20 |
+
logger.info("Iniciando generaci贸n del video...")
|
| 21 |
+
|
| 22 |
# 1. Cargar audio
|
| 23 |
+
logger.info(f"Cargando audio: {audio_file}")
|
| 24 |
+
y, sr = librosa.load(audio_file, sr=None, mono=True) # Carga completa del audio
|
| 25 |
duration = librosa.get_duration(y=y, sr=sr)
|
| 26 |
logger.info(f"Audio cargado: {duration:.1f} segundos")
|
| 27 |
|
|
|
|
| 35 |
audio_envelope_zoom = audio_envelope * 0.2 + 0.9 # Para zoom
|
| 36 |
audio_envelope_wave = audio_envelope * (img_h // 6) # Para waveform
|
| 37 |
|
| 38 |
+
# 4. Generar frames con zoom y waveform
|
| 39 |
def make_frame(t):
|
| 40 |
+
# Calcular posici贸n en el audio
|
| 41 |
time_idx = int(t * sr)
|
| 42 |
+
|
| 43 |
+
# --- Efecto de Zoom ---
|
| 44 |
zoom_factor = audio_envelope_zoom[time_idx] if time_idx < len(audio_envelope_zoom) else 1.0
|
| 45 |
new_size = (int(img_w * zoom_factor), int(img_h * zoom_factor))
|
| 46 |
zoomed_img = img.resize(new_size, Image.LANCZOS)
|
|
|
|
| 55 |
y_offset + img_h
|
| 56 |
))
|
| 57 |
|
| 58 |
+
# --- Dibujar Waveform ---
|
| 59 |
frame = ImageDraw.Draw(cropped_img)
|
| 60 |
start_y = int(img_h * 0.8) # 80% hacia abajo
|
| 61 |
|
|
|
|
| 73 |
points.extend([(x, y_pos), (x, y_neg)])
|
| 74 |
|
| 75 |
if len(points) > 2:
|
| 76 |
+
frame.polygon(points, fill=(255, 0, 0, 150)) # Rojo semitransparente
|
| 77 |
|
| 78 |
return np.array(cropped_img)
|
| 79 |
|
|
|
|
| 82 |
video.fps = 24
|
| 83 |
video = video.set_audio(mp.AudioFileClip(audio_file))
|
| 84 |
|
| 85 |
+
# 6. Guardar video en un directorio temporal persistente
|
| 86 |
temp_dir = tempfile.mkdtemp()
|
| 87 |
output_path = os.path.join(temp_dir, "output.mp4")
|
| 88 |
|
| 89 |
+
logger.info(f"Exportando video a: {output_path}")
|
| 90 |
video.write_videofile(
|
| 91 |
output_path,
|
| 92 |
codec="libx264",
|
|
|
|
| 95 |
logger=None
|
| 96 |
)
|
| 97 |
|
| 98 |
+
# Verificar que el archivo existe
|
| 99 |
+
if not os.path.exists(output_path):
|
| 100 |
+
raise Exception("Error: El archivo de video no se gener贸 correctamente.")
|
| 101 |
+
|
| 102 |
+
logger.info(f"Video guardado correctamente: {output_path}")
|
| 103 |
return output_path # Retornar la ruta completa
|
| 104 |
|
| 105 |
except Exception as e:
|
| 106 |
+
logger.error(f"Error cr铆tico: {str(e)}", exc_info=True)
|
| 107 |
+
return None
|
| 108 |
|
| 109 |
# Interfaz Gradio
|
| 110 |
iface = gr.Interface(
|
|
|
|
| 119 |
)
|
| 120 |
|
| 121 |
if __name__ == "__main__":
|
| 122 |
+
iface.queue(max_size=1).launch(show_error=True)
|