70 lines
2.6 KiB
Python
70 lines
2.6 KiB
Python
import os
|
||
import urllib.request
|
||
import numpy as np
|
||
import torch
|
||
import soundfile as sf
|
||
|
||
try:
|
||
import folder_paths
|
||
_CACHE_DIR = os.path.join(folder_paths.models_dir, "omnivoice", "presets")
|
||
except ImportError:
|
||
_CACHE_DIR = os.path.join(os.path.expanduser("~"), ".cache", "omnivoice", "presets")
|
||
|
||
# Each entry: (url, transcript)
|
||
# transcript="" means run ref_audio through a Whisper node and connect to ref_text
|
||
PRESETS = {
|
||
"Nature – male, warm (F5-TTS ref)": (
|
||
"https://raw.githubusercontent.com/SWivid/F5-TTS/main/src/f5_tts/infer/examples/basic/basic_ref_en.wav",
|
||
"Some call me nature, others call me mother nature.",
|
||
),
|
||
"Shadowheart – female, expressive (Chatterbox ref)": (
|
||
"https://storage.googleapis.com/chatterbox-demo-samples/prompts/female_shadowheart4.flac",
|
||
"That place in the distance, it's huge and dedicated to Lady Shar. It can only mean one thing. I have a hidden place close to the cloister where night orchids bloom.",
|
||
),
|
||
}
|
||
|
||
|
||
def _load_audio(url):
|
||
"""Download (once) and return (waveform_tensor, sample_rate)."""
|
||
os.makedirs(_CACHE_DIR, exist_ok=True)
|
||
filename = os.path.basename(url.split("?")[0])
|
||
cache_path = os.path.join(_CACHE_DIR, filename)
|
||
if not os.path.exists(cache_path):
|
||
urllib.request.urlretrieve(url, cache_path)
|
||
audio_np, sr = sf.read(cache_path, dtype="float32")
|
||
if audio_np.ndim == 1:
|
||
audio_np = audio_np[np.newaxis, :] # (1, samples)
|
||
else:
|
||
audio_np = audio_np.T # (channels, samples)
|
||
waveform = torch.from_numpy(audio_np).unsqueeze(0) # (1, channels, samples)
|
||
return waveform, sr
|
||
|
||
|
||
class OmniVoiceVoicePreset:
|
||
@classmethod
|
||
def INPUT_TYPES(cls):
|
||
return {
|
||
"required": {
|
||
"preset": (
|
||
list(PRESETS.keys()),
|
||
{
|
||
"tooltip": (
|
||
"Pre-fetched reference voice for OmniVoice Generate.\n"
|
||
"Connect ref_audio → ref_audio and ref_text → ref_text.\n"
|
||
"If ref_text is blank, connect a Whisper node to supply the transcript."
|
||
),
|
||
},
|
||
),
|
||
},
|
||
}
|
||
|
||
RETURN_TYPES = ("AUDIO", "STRING")
|
||
RETURN_NAMES = ("ref_audio", "ref_text")
|
||
FUNCTION = "load_preset"
|
||
CATEGORY = "OmniVoice"
|
||
|
||
def load_preset(self, preset):
|
||
url, transcript = PRESETS[preset]
|
||
waveform, sr = _load_audio(url)
|
||
return ({"waveform": waveform, "sample_rate": sr}, transcript)
|