Federico Volovnik-Jose
284d50b44c
- Use relative imports across nodes/ and lib/ so ComfyUI's loader resolves the local subpackages instead of colliding with ComfyUI's own top-level nodes module. - tts_output_to_audio_dict now returns CPU tensors so downstream nodes don't see CUDA tensors. - model_state cache key uses resolved dtype string, so dtype=auto and the matching explicit choice share one cache slot (no spurious 13 GB reload). - Add cleanup_one(key) helper as called for in spec section 4. - Fix pyproject Homepage URL to point at this repo.
37 lines
1.1 KiB
Python
37 lines
1.1 KiB
Python
"""Resolve a ComfyUI AUDIO input or string path into a voice-reference dict."""
|
|
|
|
import logging
|
|
import os
|
|
from typing import Optional
|
|
|
|
from .audio import audio_dict_to_wav
|
|
|
|
log = logging.getLogger(__name__)
|
|
|
|
|
|
def build_voice_reference(
|
|
audio: Optional[dict],
|
|
audio_path: str,
|
|
temp_dir: str,
|
|
) -> dict:
|
|
"""Return a voice-reference container.
|
|
|
|
Priority:
|
|
1. If `audio` is set, write it to a temp WAV and use that path.
|
|
If `audio_path` is also set, log a warning and ignore it.
|
|
2. Else if `audio_path` is non-empty, validate the path exists and use it.
|
|
3. Else return source="none" / path=None (default voice).
|
|
"""
|
|
if audio is not None:
|
|
if audio_path:
|
|
log.warning("Both audio and audio_path provided; ignoring audio_path=%r", audio_path)
|
|
path = audio_dict_to_wav(audio, dest_dir=temp_dir)
|
|
return {"path": path, "source": "audio"}
|
|
|
|
if audio_path:
|
|
if not os.path.isfile(audio_path):
|
|
raise FileNotFoundError(f"Voice reference audio not found: {audio_path}")
|
|
return {"path": audio_path, "source": "path"}
|
|
|
|
return {"path": None, "source": "none"}
|