144 lines
4.4 KiB
Python
144 lines
4.4 KiB
Python
import asyncio
|
|
import os
|
|
import subprocess
|
|
|
|
import imageio_ffmpeg
|
|
|
|
import config
|
|
|
|
FFMPEG = imageio_ffmpeg.get_ffmpeg_exe()
|
|
W, H, FPS = 1280, 720, 20
|
|
FUENTE = None
|
|
for p in (r"C:\Windows\Fonts\arialbd.ttf",
|
|
r"C:\Windows\Fonts\arial.ttf",
|
|
r"C:\Windows\Fonts\segoeui.ttf"):
|
|
if os.path.exists(p):
|
|
FUENTE = p
|
|
break
|
|
|
|
|
|
def _duracion_audio(ruta):
|
|
try:
|
|
r = subprocess.run(
|
|
[FFMPEG, "-i", ruta, "-f", "null", "-"],
|
|
capture_output=True, text=True, timeout=30,
|
|
)
|
|
for linea in (r.stderr or "").splitlines():
|
|
if "Duration:" in linea:
|
|
hh, mm, ss = linea.split("Duration:")[1].split(",")[0].strip().split(":")
|
|
return int(hh) * 3600 + int(mm) * 60 + float(ss)
|
|
except Exception:
|
|
pass
|
|
return None
|
|
|
|
|
|
def sintetizar_con_oraciones(texto, nombre="anuncio.mp3"):
|
|
"""Genera el audio y devuelve (ruta_audio, oraciones_con_tiempos)."""
|
|
from PIL import Image # noqa: F401 (import de dependencia temprano)
|
|
import edge_tts
|
|
|
|
ruta = os.path.join(os.path.dirname(os.path.abspath(__file__)),
|
|
"voz_media", nombre)
|
|
os.makedirs(os.path.dirname(ruta), exist_ok=True)
|
|
audio = []
|
|
oraciones = []
|
|
|
|
async def _gen():
|
|
comunicador = edge_tts.Communicate(
|
|
texto, config.VOZ_EDGE,
|
|
rate=config.VOZ_RATE, pitch=config.VOZ_PITCH,
|
|
)
|
|
async for ch in comunicador.stream():
|
|
if ch["type"] == "audio":
|
|
audio.append(ch["data"])
|
|
elif ch["type"] == "SentenceBoundary":
|
|
inicio = ch["offset"] / 1e7
|
|
fin = inicio + ch["duration"] / 1e7
|
|
oraciones.append((ch["text"], inicio, fin))
|
|
|
|
asyncio.run(_gen())
|
|
if not audio:
|
|
return None, []
|
|
with open(ruta, "wb") as f:
|
|
f.write(b"".join(audio))
|
|
return ruta, oraciones
|
|
|
|
|
|
def _palabras(oraciones):
|
|
"""Distribuye las palabras dentro de cada frase por longitud."""
|
|
palabras = []
|
|
for texto, s, e in oraciones:
|
|
tokens = texto.split()
|
|
total = sum(len(t) for t in tokens) or 1
|
|
t = s
|
|
for tok in tokens:
|
|
dur = (e - s) * (len(tok) / total)
|
|
palabras.append((tok, t, t + dur))
|
|
t += dur
|
|
return palabras
|
|
|
|
|
|
def generar_video(texto, ruta_audio, oraciones, nombre="anuncio.mp4"):
|
|
from PIL import Image, ImageDraw, ImageFont
|
|
|
|
ruta_video = os.path.join(os.path.dirname(os.path.abspath(__file__)),
|
|
"voz_media", nombre)
|
|
palabras = _palabras(oraciones)
|
|
|
|
dur = _duracion_audio(ruta_audio) or ((palabras[-1][2] if palabras else 0) + 2)
|
|
total = dur + 0.5
|
|
n_frames = int(total * FPS)
|
|
|
|
font = ImageFont.truetype(FUENTE, 52) if FUENTE else ImageFont.load_default()
|
|
gap = 18
|
|
|
|
# agrupar en lineas
|
|
lineas = []
|
|
linea = []
|
|
ancho_linea = 0
|
|
for tok, s, e in palabras:
|
|
ancho = font.getlength(tok)
|
|
if ancho_linea + ancho + gap > W - 80 and linea:
|
|
lineas.append(linea)
|
|
linea = [(tok, s, e)]
|
|
ancho_linea = ancho
|
|
else:
|
|
linea.append((tok, s, e))
|
|
ancho_linea += ancho + gap
|
|
if linea:
|
|
lineas.append(linea)
|
|
|
|
cmd = [FFMPEG, "-y", "-f", "rawvideo", "-vcodec", "rawvideo",
|
|
"-s", f"{W}x{H}", "-pix_fmt", "rgb24", "-r", str(FPS), "-i", "-",
|
|
"-i", ruta_audio,
|
|
"-c:v", "libx264", "-preset", "veryfast", "-pix_fmt", "yuv420p",
|
|
"-c:a", "aac", "-b:a", "128k", "-shortest", ruta_video]
|
|
proc = subprocess.Popen(cmd, stdin=subprocess.PIPE,
|
|
stderr=subprocess.DEVNULL)
|
|
|
|
for i in range(n_frames):
|
|
t = i / FPS
|
|
img = Image.new("RGB", (W, H), (8, 8, 8))
|
|
d = ImageDraw.Draw(img)
|
|
y = H // 2 - (len(lineas) // 2) * 80 + 30
|
|
for linea_words in lineas:
|
|
x = 40
|
|
for tok, s, e in linea_words:
|
|
if s <= t < e:
|
|
color = (255, 220, 0)
|
|
elif t >= e:
|
|
color = (255, 255, 255)
|
|
else:
|
|
color = (90, 90, 90)
|
|
d.text((x, y), tok, fill=color, font=font)
|
|
x += font.getlength(tok) + gap
|
|
y += 80
|
|
try:
|
|
proc.stdin.write(img.tobytes())
|
|
except BrokenPipeError:
|
|
break
|
|
|
|
proc.stdin.close()
|
|
proc.wait()
|
|
return ruta_video
|