turn it into a package

GPT_SoVITS/TTS_infer_pack/TTS.py

@@ -1,60 +1,66 @@
 from copy import deepcopy
 import math
 import os, sys, gc
-import random
-import traceback
 
-from tqdm import tqdm
+
 now_dir = os.getcwd()
 sys.path.append(now_dir)
 import ffmpeg
-import os
+import librosa
-from typing import Generator, List, Tuple, Union
 import numpy as np
+import random
 import torch
 import torch.nn.functional as F
+import traceback
 import yaml
-from transformers import AutoModelForMaskedLM, AutoTokenizer
+
+from huggingface_hub import snapshot_download, hf_hub_download
+from importlib.resources import files
+from time import time as ttime
+from typing import Generator, List, Tuple, Union
+from tqdm import tqdm
+
 
 from AR.models.t2s_lightning_module import Text2SemanticLightningModule
 from feature_extractor.cnhubert import CNHubert
 from module.models import SynthesizerTrn
-import librosa
+from module.mel_processing import spectrogram_torch
-from time import time as ttime
 from tools.i18n.i18n import I18nAuto, scan_language_list
 from tools.my_utils import load_audio
-from module.mel_processing import spectrogram_torch
+from transformers import AutoModelForMaskedLM, AutoTokenizer
 from TTS_infer_pack.text_segmentation_method import splits
 from TTS_infer_pack.TextPreprocessor import TextPreprocessor
+
 language=os.environ.get("language","Auto")
 language=sys.argv[-1] if sys.argv[-1] in scan_language_list() else language
 i18n = I18nAuto(language=language)
+LIBRARY_NAME = "GPT_SoVITS"
 
 # configs/tts_infer.yaml
 """
 custom:
-  bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
+  bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
-  cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
+  cnhuhbert_base_path: pretrained_models/chinese-hubert-base
   device: cpu
   is_half: false
-  t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
+  t2s_weights_path: pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
-  vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
+  vits_weights_path: pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
   version: v2
 default:
-  bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
+  bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
-  cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
+  cnhuhbert_base_path: pretrained_models/chinese-hubert-base
   device: cpu
   is_half: false
-  t2s_weights_path: GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
+  t2s_weights_path: pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
-  vits_weights_path: GPT_SoVITS/pretrained_models/s2G488k.pth
+  vits_weights_path: pretrained_models/s2G488k.pth
   version: v1
 default_v2:
-  bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
+  bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
-  cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
+  cnhuhbert_base_path: pretrained_models/chinese-hubert-base
   device: cpu
   is_half: false
-  t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
+  t2s_weights_path: pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
-  vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
+  vits_weights_path: pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
   version: v2
 """
 
@@ -86,19 +92,19 @@ class TTS_Config:
                 "device": "cpu",
                 "is_half": False,
                 "version": "v1",
-                "t2s_weights_path": "GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt",
+                "t2s_weights_path": "pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt",
-                "vits_weights_path": "GPT_SoVITS/pretrained_models/s2G488k.pth",
+                "vits_weights_path": "pretrained_models/s2G488k.pth",
-                "cnhuhbert_base_path": "GPT_SoVITS/pretrained_models/chinese-hubert-base",
+                "cnhuhbert_base_path": "pretrained_models/chinese-hubert-base",
-                "bert_base_path": "GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large",
+                "bert_base_path": "pretrained_models/chinese-roberta-wwm-ext-large",
             },
         "default_v2":{
                 "device": "cpu",
                 "is_half": False,
                 "version": "v2",
-                "t2s_weights_path": "GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt",
+                "t2s_weights_path": "pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt",
-                "vits_weights_path": "GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth",
+                "vits_weights_path": "pretrained_models/gsv-v2final-pretrained/s2G2333k.pth",
-                "cnhuhbert_base_path": "GPT_SoVITS/pretrained_models/chinese-hubert-base",
+                "cnhuhbert_base_path": "pretrained_models/chinese-hubert-base",
-                "bert_base_path": "GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large",
+                "bert_base_path": "pretrained_models/chinese-roberta-wwm-ext-large",
             },
     }
     configs:dict = None
@@ -120,7 +126,7 @@ class TTS_Config:
     def __init__(self, configs: Union[dict, str]=None):
         
         # 设置默认配置文件路径
-        configs_base_path:str = "GPT_SoVITS/configs/"
+        configs_base_path:str = "configs/"
         os.makedirs(configs_base_path, exist_ok=True)
         self.configs_path:str = os.path.join(configs_base_path, "tts_infer.yaml")
         
@@ -153,21 +159,22 @@ class TTS_Config:
         self.cnhuhbert_base_path = self.configs.get("cnhuhbert_base_path", None)
         self.languages = self.v2_languages if self.version=="v2" else self.v1_languages
         
-        
         if (self.t2s_weights_path in [None, ""]) or (not os.path.exists(self.t2s_weights_path)):
-            self.t2s_weights_path = self.default_configs[default_config_key]['t2s_weights_path']
+            self.t2s_weights_path = str(files(LIBRARY_NAME).joinpath(self.default_configs[default_config_key]['t2s_weights_path']))
             print(f"fall back to default t2s_weights_path: {self.t2s_weights_path}")
         if (self.vits_weights_path in [None, ""]) or (not os.path.exists(self.vits_weights_path)):
-            self.vits_weights_path = self.default_configs[default_config_key]['vits_weights_path']
+            self.vits_weights_path = str(files(LIBRARY_NAME).joinpath(self.default_configs[default_config_key]['vits_weights_path']))
             print(f"fall back to default vits_weights_path: {self.vits_weights_path}")
         if (self.bert_base_path in [None, ""]) or (not os.path.exists(self.bert_base_path)):
-            self.bert_base_path = self.default_configs[default_config_key]['bert_base_path']
+            self.bert_base_path = str(files(LIBRARY_NAME).joinpath(self.default_configs[default_config_key]['bert_base_path']))
             print(f"fall back to default bert_base_path: {self.bert_base_path}")
         if (self.cnhuhbert_base_path in [None, ""]) or (not os.path.exists(self.cnhuhbert_base_path)):
-            self.cnhuhbert_base_path = self.default_configs[default_config_key]['cnhuhbert_base_path']
+            self.cnhuhbert_base_path = str(files(LIBRARY_NAME).joinpath(self.default_configs[default_config_key]['cnhuhbert_base_path']))
             print(f"fall back to default cnhuhbert_base_path: {self.cnhuhbert_base_path}")
-        self.update_configs()
             
+        repo_name="lj1995/GPT-SoVITS"
+        snapshot_download(repo_id=repo_name, local_dir=os.path.dirname(self.bert_base_path))
+        self.update_configs()
         
         self.max_sec = None
         self.hz:int = 50

GPT_SoVITS/api_v2.py

@@ -114,6 +114,8 @@ from fastapi import FastAPI, Request, HTTPException, Response
 from fastapi.responses import StreamingResponse, JSONResponse
 from fastapi import FastAPI, UploadFile, File
 import uvicorn
+
+from importlib.resources import files
 from io import BytesIO
 from tools.i18n.i18n import I18nAuto
 from GPT_SoVITS.TTS_infer_pack.TTS import TTS, TTS_Config
@@ -125,7 +127,7 @@ i18n = I18nAuto()
 cut_method_names = get_cut_method_names()
 
 parser = argparse.ArgumentParser(description="GPT-SoVITS api")
-parser.add_argument("-c", "--tts_config", type=str, default="GPT_SoVITS/configs/tts_infer.yaml", help="tts_infer路径")
+parser.add_argument("-c", "--tts_config", type=str, default=None, help="tts_infer路径")
 parser.add_argument("-a", "--bind_addr", type=str, default="127.0.0.1", help="default: 127.0.0.1")
 parser.add_argument("-p", "--port", type=int, default="9880", help="default: 9880")
 args = parser.parse_args()
@@ -136,7 +138,7 @@ host = args.bind_addr
 argv = sys.argv
 
 if config_path in [None, ""]:
-    config_path = "GPT-SoVITS/configs/tts_infer.yaml"
+    config_path = str(files("GPT_SoVITS").joinpath("configs/tts_infer.yaml"))
 
 tts_config = TTS_Config(config_path)
 print(tts_config)
@@ -394,7 +396,7 @@ async def tts_get_endpoint(
 
 @APP.post("/tts")
 async def tts_post_endpoint(request: TTS_Request):
-    req = request.dict()
+    req = request.model_dump()
     return await tts_handle(req)
 
 
@@ -449,7 +451,8 @@ async def set_sovits_weights(weights_path: str = None):
 
 
 
-if __name__ == "__main__":
+def main():
+    global port, host, argv
     try:
         if host == 'None':   # 在调用时使用 -a None 参数，可以让api监听双栈
             host = None
@@ -458,3 +461,6 @@ if __name__ == "__main__":
         traceback.print_exc()
         os.kill(os.getpid(), signal.SIGTERM)
         exit(0)
+
+if __name__ == "__main__":
+    main()

GPT_SoVITS/configs/tts_infer.yaml

@@ -1,24 +1,24 @@
 custom:
-  bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
+  bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
-  cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
+  cnhuhbert_base_path: pretrained_models/chinese-hubert-base
   device: cuda
   is_half: true
-  t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
+  t2s_weights_path: pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
   version: v2
-  vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
+  vits_weights_path: pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
 default:
-  bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
+  bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
-  cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
+  cnhuhbert_base_path: pretrained_models/chinese-hubert-base
   device: cpu
   is_half: false
-  t2s_weights_path: GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
+  t2s_weights_path: pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
   version: v1
-  vits_weights_path: GPT_SoVITS/pretrained_models/s2G488k.pth
+  vits_weights_path: pretrained_models/s2G488k.pth
 default_v2:
-  bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
+  bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
-  cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
+  cnhuhbert_base_path: pretrained_models/chinese-hubert-base
   device: cpu
   is_half: false
-  t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
+  t2s_weights_path: pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
   version: v2
-  vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
+  vits_weights_path: pretrained_models/gsv-v2final-pretrained/s2G2333k.pth

MANIFEST.in

@@ -0,0 +1 @@
+include GPT_SoVITS/configs/*

README.md

@@ -1,3 +1,6 @@
+# Jarod's NOTE
+Working on turning this into a package.  Right now, the API *does in fact* work to make requests to and this can be installed.  Will get instrucitons out sometime in the future.
+
 <div align="center">
 
 

pyproject.toml

@@ -0,0 +1,63 @@
+[build-system]
+requires = ["setuptools >= 61.0", "setuptools-scm>=8.0"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "GPT_SoVITS"
+dynamic = ["version"]
+description = ""
+readme = "README.md"
+license = {text = "MIT License"}
+classifiers = [
+    "License :: OSI Approved :: MIT License",
+    "Operating System :: OS Independent",
+    "Programming Language :: Python :: 3",
+]
+dependencies = [
+    "numpy==1.26.4",
+    "scipy",
+    "tensorboard",
+    "librosa==0.9.2",
+    "numba",
+    "pytorch-lightning",
+    "gradio>=4.0,<=4.24.0",
+    "ffmpeg-python",
+    "onnxruntime; sys_platform == 'darwin'",
+    "onnxruntime-gpu; sys_platform != 'darwin'",
+    "tqdm",
+    "funasr==1.0.27",
+    "cn2an",
+    "pypinyin",
+    "pyopenjtalk>=0.3.4",
+    "g2p_en",
+    "torchaudio",
+    "modelscope==1.10.0",
+    "sentencepiece",
+    "transformers",
+    "chardet",
+    "PyYAML",
+    "psutil",
+    "jieba_fast",
+    "jieba",
+    "LangSegment>=0.2.0",
+    "Faster_Whisper",
+    "wordsegment",
+    "rotary_embedding_torch",
+    "pyjyutping",
+    "g2pk2",
+    "ko_pron",
+    "opencc; sys_platform != 'linux'",
+    "opencc==1.1.1; sys_platform == 'linux'",
+    "python_mecab_ko; sys_platform != 'win32'",
+    "fastapi<0.112.2",
+]
+
+
+[project.urls]
+Homepage = "https://github.com/RVC-Boss/GPT-SoVITS"
+
+[project.scripts]
+"gpt_sovits_api" = "GPT_SoVITS.api_v2:main"
+
+[tool.setuptools.packages.find]
+include = ["GPT_SoVITS"]

requirements.txt

@@ -1,8 +1,8 @@
-numpy==1.23.4
+numpy==1.26.4
 scipy
 tensorboard
 librosa==0.9.2
-numba==0.56.4
+numba
 pytorch-lightning
 gradio>=4.0,<=4.24.0
 ffmpeg-python