Update app.py
Browse files
app.py
CHANGED
@@ -3,10 +3,9 @@ from tts_module import get_voices, text_to_speech
|
|
3 |
from pixabay_api import search_pixabay
|
4 |
from moviepy.editor import (
|
5 |
AudioFileClip, VideoFileClip, CompositeAudioClip,
|
6 |
-
concatenate_audioclips, concatenate_videoclips,
|
7 |
ColorClip
|
8 |
)
|
9 |
-
from scipy.ndimage import gaussian_filter
|
10 |
import asyncio
|
11 |
import os
|
12 |
import json
|
@@ -23,12 +22,14 @@ if service_account_info:
|
|
23 |
with open('service-account.json', 'w') as f:
|
24 |
json.dump(service_account_info, f)
|
25 |
|
|
|
26 |
output_folder = "outputs"
|
27 |
temp_dir = "temp_files"
|
28 |
os.makedirs(output_folder, exist_ok=True)
|
29 |
os.makedirs(temp_dir, exist_ok=True)
|
30 |
|
31 |
-
|
|
|
32 |
|
33 |
def cleanup_temp_files():
|
34 |
"""Elimina todos los archivos temporales de la carpeta temp_files."""
|
@@ -40,25 +41,16 @@ def cleanup_temp_files():
|
|
40 |
except Exception as e:
|
41 |
print(f"Error deleting {file_path}: {e}")
|
42 |
|
43 |
-
def
|
44 |
-
"""Redimensiona el video al tama帽o 1080p (16:9)
|
45 |
try:
|
46 |
w, h = clip.size
|
47 |
current_aspect_ratio = w / h
|
48 |
target_aspect_ratio = target_width / target_height
|
49 |
|
50 |
if abs(current_aspect_ratio - target_aspect_ratio) < 0.1:
|
51 |
-
# Si la relaci贸n de aspecto ya es cercana a 16:9, solo redimensionamos
|
52 |
return clip.resize((target_width, target_height))
|
53 |
|
54 |
-
# Crear un fondo borroso con las dimensiones objetivo
|
55 |
-
background = ColorClip(size=(target_width, target_height), color=[0, 0, 0]).set_duration(clip.duration)
|
56 |
-
try:
|
57 |
-
# Aplicar filtro gaussiano como alternativa al blur
|
58 |
-
background = background.fx(vfx, lambda gf: gaussian_filter(gf, sigma=5))
|
59 |
-
except Exception as e:
|
60 |
-
print(f"Error al aplicar filtro gaussiano: {e}")
|
61 |
-
|
62 |
# Redimensionar el video original para mantener su proporci贸n
|
63 |
if current_aspect_ratio < target_aspect_ratio: # Video vertical
|
64 |
new_height = target_height
|
@@ -71,10 +63,11 @@ def resize_and_blur_video(clip, target_width=1920, target_height=1080):
|
|
71 |
y_center = (target_height - new_height) / 2
|
72 |
resized_clip = clip.resize(height=new_height).set_position((0, y_center))
|
73 |
|
74 |
-
#
|
|
|
75 |
return CompositeVideoClip([background, resized_clip], size=(target_width, target_height))
|
76 |
except Exception as e:
|
77 |
-
print(f"Error en
|
78 |
return clip
|
79 |
|
80 |
def download_video(link):
|
@@ -109,7 +102,7 @@ def concatenate_pixabay_videos(keywords, num_videos_per_keyword=1):
|
|
109 |
temp_video_path = download_video(links[0])
|
110 |
if temp_video_path:
|
111 |
clip = VideoFileClip(temp_video_path)
|
112 |
-
processed_clip =
|
113 |
video_clips.append(processed_clip)
|
114 |
os.remove(temp_video_path) # Eliminar archivo temporal despu茅s de usarlo
|
115 |
except Exception as e:
|
@@ -202,24 +195,30 @@ def upload_to_google_drive(file_path, folder_id):
|
|
202 |
def process_input(text, txt_file, mp3_file, selected_voice, rate, pitch, keywords):
|
203 |
"""Procesa la entrada del usuario y genera el video final."""
|
204 |
try:
|
|
|
205 |
if text.strip():
|
206 |
final_text = text
|
207 |
elif txt_file is not None:
|
208 |
final_text = txt_file.decode("utf-8")
|
209 |
else:
|
210 |
raise ValueError("No text input provided")
|
|
|
211 |
audio_file = asyncio.run(text_to_speech(final_text, selected_voice, rate, pitch))
|
212 |
if not audio_file:
|
213 |
raise ValueError("Failed to generate audio")
|
|
|
214 |
video_clip = concatenate_pixabay_videos(keywords, num_videos_per_keyword=1)
|
215 |
if not video_clip:
|
216 |
raise ValueError("Failed to generate video")
|
|
|
217 |
music_clip = None
|
218 |
if mp3_file is not None:
|
219 |
music_clip = adjust_background_music(video_clip.duration, mp3_file.name)
|
|
|
220 |
final_video_path = combine_audio_video(audio_file, video_clip, music_clip)
|
221 |
if not final_video_path:
|
222 |
raise ValueError("Failed to combine audio and video")
|
|
|
223 |
download_link = upload_to_google_drive(final_video_path, folder_id=FOLDER_ID)
|
224 |
if download_link:
|
225 |
return f"[Descargar video]({download_link})"
|
@@ -229,8 +228,9 @@ def process_input(text, txt_file, mp3_file, selected_voice, rate, pitch, keyword
|
|
229 |
print(f"Error durante el procesamiento: {e}")
|
230 |
return None
|
231 |
finally:
|
232 |
-
cleanup_temp_files()
|
233 |
|
|
|
234 |
with gr.Blocks() as demo:
|
235 |
gr.Markdown("# Text-to-Video Generator")
|
236 |
with gr.Row():
|
@@ -247,7 +247,8 @@ with gr.Blocks() as demo:
|
|
247 |
rate_slider = gr.Slider(minimum=-50, maximum=50, value=0, label="Speech Rate Adjustment (%)", step=1)
|
248 |
pitch_slider = gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch Adjustment (Hz)", step=1)
|
249 |
with gr.Column():
|
250 |
-
output_link = gr.Markdown("")
|
|
|
251 |
btn = gr.Button("Generate Video")
|
252 |
btn.click(
|
253 |
process_input,
|
@@ -255,5 +256,8 @@ with gr.Blocks() as demo:
|
|
255 |
outputs=output_link
|
256 |
)
|
257 |
|
|
|
258 |
port = int(os.getenv("PORT", 7860))
|
|
|
|
|
259 |
demo.launch(server_name="0.0.0.0", server_port=port, share=True, show_error=True)
|
|
|
3 |
from pixabay_api import search_pixabay
|
4 |
from moviepy.editor import (
|
5 |
AudioFileClip, VideoFileClip, CompositeAudioClip,
|
6 |
+
concatenate_audioclips, concatenate_videoclips, CompositeVideoClip,
|
7 |
ColorClip
|
8 |
)
|
|
|
9 |
import asyncio
|
10 |
import os
|
11 |
import json
|
|
|
22 |
with open('service-account.json', 'w') as f:
|
23 |
json.dump(service_account_info, f)
|
24 |
|
25 |
+
# Define la carpeta de salida y la carpeta temporal
|
26 |
output_folder = "outputs"
|
27 |
temp_dir = "temp_files"
|
28 |
os.makedirs(output_folder, exist_ok=True)
|
29 |
os.makedirs(temp_dir, exist_ok=True)
|
30 |
|
31 |
+
# ID de la carpeta de destino en Google Drive
|
32 |
+
FOLDER_ID = "12S6adpanAXjf71pKKGRRPqpzbJa5XEh3" # Reemplaza con tu ID de carpeta
|
33 |
|
34 |
def cleanup_temp_files():
|
35 |
"""Elimina todos los archivos temporales de la carpeta temp_files."""
|
|
|
41 |
except Exception as e:
|
42 |
print(f"Error deleting {file_path}: {e}")
|
43 |
|
44 |
+
def resize_video(clip, target_width=1920, target_height=1080):
|
45 |
+
"""Redimensiona el video al tama帽o 1080p (16:9) sin aplicar desenfoque."""
|
46 |
try:
|
47 |
w, h = clip.size
|
48 |
current_aspect_ratio = w / h
|
49 |
target_aspect_ratio = target_width / target_height
|
50 |
|
51 |
if abs(current_aspect_ratio - target_aspect_ratio) < 0.1:
|
|
|
52 |
return clip.resize((target_width, target_height))
|
53 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
# Redimensionar el video original para mantener su proporci贸n
|
55 |
if current_aspect_ratio < target_aspect_ratio: # Video vertical
|
56 |
new_height = target_height
|
|
|
63 |
y_center = (target_height - new_height) / 2
|
64 |
resized_clip = clip.resize(height=new_height).set_position((0, y_center))
|
65 |
|
66 |
+
# Crear un fondo negro y combinar con el video redimensionado
|
67 |
+
background = ColorClip(size=(target_width, target_height), color=[0, 0, 0]).set_duration(clip.duration)
|
68 |
return CompositeVideoClip([background, resized_clip], size=(target_width, target_height))
|
69 |
except Exception as e:
|
70 |
+
print(f"Error en resize_video: {e}")
|
71 |
return clip
|
72 |
|
73 |
def download_video(link):
|
|
|
102 |
temp_video_path = download_video(links[0])
|
103 |
if temp_video_path:
|
104 |
clip = VideoFileClip(temp_video_path)
|
105 |
+
processed_clip = resize_video(clip)
|
106 |
video_clips.append(processed_clip)
|
107 |
os.remove(temp_video_path) # Eliminar archivo temporal despu茅s de usarlo
|
108 |
except Exception as e:
|
|
|
195 |
def process_input(text, txt_file, mp3_file, selected_voice, rate, pitch, keywords):
|
196 |
"""Procesa la entrada del usuario y genera el video final."""
|
197 |
try:
|
198 |
+
# Determinamos el texto a usar
|
199 |
if text.strip():
|
200 |
final_text = text
|
201 |
elif txt_file is not None:
|
202 |
final_text = txt_file.decode("utf-8")
|
203 |
else:
|
204 |
raise ValueError("No text input provided")
|
205 |
+
# Generamos el audio
|
206 |
audio_file = asyncio.run(text_to_speech(final_text, selected_voice, rate, pitch))
|
207 |
if not audio_file:
|
208 |
raise ValueError("Failed to generate audio")
|
209 |
+
# Generamos el video
|
210 |
video_clip = concatenate_pixabay_videos(keywords, num_videos_per_keyword=1)
|
211 |
if not video_clip:
|
212 |
raise ValueError("Failed to generate video")
|
213 |
+
# Procesamos la m煤sica de fondo si existe
|
214 |
music_clip = None
|
215 |
if mp3_file is not None:
|
216 |
music_clip = adjust_background_music(video_clip.duration, mp3_file.name)
|
217 |
+
# Combinamos todo
|
218 |
final_video_path = combine_audio_video(audio_file, video_clip, music_clip)
|
219 |
if not final_video_path:
|
220 |
raise ValueError("Failed to combine audio and video")
|
221 |
+
# Subimos a Google Drive y obtenemos el enlace
|
222 |
download_link = upload_to_google_drive(final_video_path, folder_id=FOLDER_ID)
|
223 |
if download_link:
|
224 |
return f"[Descargar video]({download_link})"
|
|
|
228 |
print(f"Error durante el procesamiento: {e}")
|
229 |
return None
|
230 |
finally:
|
231 |
+
cleanup_temp_files() # Limpiar archivos temporales al finalizar
|
232 |
|
233 |
+
# Interfaz Gradio
|
234 |
with gr.Blocks() as demo:
|
235 |
gr.Markdown("# Text-to-Video Generator")
|
236 |
with gr.Row():
|
|
|
247 |
rate_slider = gr.Slider(minimum=-50, maximum=50, value=0, label="Speech Rate Adjustment (%)", step=1)
|
248 |
pitch_slider = gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch Adjustment (Hz)", step=1)
|
249 |
with gr.Column():
|
250 |
+
output_link = gr.Markdown("") # Mostrar el enlace de descarga
|
251 |
+
|
252 |
btn = gr.Button("Generate Video")
|
253 |
btn.click(
|
254 |
process_input,
|
|
|
256 |
outputs=output_link
|
257 |
)
|
258 |
|
259 |
+
# Leer el puerto asignado por Hugging Face
|
260 |
port = int(os.getenv("PORT", 7860))
|
261 |
+
|
262 |
+
# Lanzar la aplicaci贸n
|
263 |
demo.launch(server_name="0.0.0.0", server_port=port, share=True, show_error=True)
|