datascale-ai-opentalking/pyproject.toml

[build-system]
requires = ["setuptools>=68", "wheel"]
build-backend = "setuptools.build_meta"

[project]
name = "opentalking"
version = "0.1.0"
description = "Real-time digital human framework with FlashTalk 14B inference"
readme = "README.md"
requires-python = ">=3.10"
license = { text = "Apache-2.0" }
authors = [{ name = "OpenTalking Contributors" }]
dependencies = [
  "fastapi>=0.109",
  "uvicorn[standard]>=0.27",
  "pydantic>=2",
  "pydantic-settings>=2",
  "redis>=5",
  "numpy>=1.24,<2",
  "httpx>=0.26",
  "edge-tts>=6.1",
  "dashscope>=1.25.11",
  "aiortc>=1.6",
  'av>=14,<14.3; python_version < "3.13"',
  "python-multipart>=0.0.9",
  "pillow>=10",
  "websockets>=13",
  "loguru>=0.7",
  "PyYAML>=6",
  "opencv-python>=4.8,<4.12",
  'mediapipe==0.10.21; python_version < "3.13" and platform_machine != "aarch64"',
  'mediapipe==0.10.18; python_version < "3.13" and platform_machine == "aarch64"',
  "kornia>=0.8.2",
  "onnx2torch>=1.5.15",
  "insightface>=0.7.3",
  "transformers>=4.57,<6",
  "openai>=1.0",
  "lightrag-hku>=1.4.9",
  "mem0ai>=0.1.115",
  "huggingface-hub[cli]<1.0",
  "rembg>=2.0.69",
]

[project.optional-dependencies]
engine = [
  "torch>=2.0,<2.11",
  "torchaudio>=2.0,<2.11",
  "torchvision>=0.15,<0.26",
  "diffusers>=0.34",
  "transformers>=4.46",
  "accelerate>=1.0",
  "opencv-python>=4.8,<4.12",
  "xfuser>=0.4",
  "librosa>=0.10",
  "pyloudnorm>=0.1.1",
  "easydict>=1.13",
  "imageio>=2.34",
  "imageio-ffmpeg>=0.5",
  "xformers>=0.0.28",
  "optimum>=1.22",
  "safetensors>=0.4",
  "ftfy>=6.2",
  "regex>=2024.0.0",
  "sentencepiece>=0.2",
]
models = [
  "torch>=2.0,<2.11",
  "torchaudio>=2.0,<2.11",
  "torchvision>=0.15,<0.26",
  "opencv-python>=4.8,<4.12",
  "kornia>=0.7",
  "insightface>=0.7",
  "librosa>=0.10",
  "scipy>=1.10",
  "tqdm>=4.65",
  "diffusers>=0.34",
  "accelerate>=1.0",
  "openai-whisper>=20240930",
]
local-audio = [
  "torch>=2.0,<2.11",
  "torchaudio>=2.0,<2.11",
  "torchvision>=0.15,<0.26",
  "funasr>=1.2.7",
  "modelscope>=1.31.0",
  "huggingface-hub>=0.30",
  "sherpa-onnx>=1.12.0",
  "soundfile>=0.12",
  "librosa>=0.10",
  "scipy>=1.10",
]
local-qwen3-tts-service = [
  "qwen-tts>=0.1.1",
  "transformers==4.57.3",
  "accelerate>=1.12",
  "torch>=2.0,<2.11",
  "torchaudio>=2.0,<2.11",
  "soundfile>=0.12",
  "fastapi>=0.109",
  "uvicorn[standard]>=0.27",
]
quicktalk-cpu = [
  "imageio-ffmpeg>=0.5",
  "onnxruntime>=1.24.3",
]
quicktalk-cuda = [
  "imageio-ffmpeg>=0.5",
  "onnxruntime-gpu>=1.24.0,<1.27",
]
local-cosyvoice-service = [
  "fastapi>=0.109",
  "uvicorn[standard]>=0.27",
  "numpy>=1.24,<2",
  "torch>=2.0,<2.11",
  "torchaudio>=2.0,<2.11",
  "soundfile>=0.12",
  "openai-whisper>=20240930",
  "modelscope>=1.31.0",
  "huggingface-hub>=0.30",
  "onnxruntime-gpu>=1.24.0",
  "librosa>=0.10",
]
ascend = ["torch-npu>=2.1"]
demo = ["gradio>=5.0"]
dev = [
  "pytest>=7.4",
  "pytest-asyncio>=0.23",
  "opencv-python-headless>=4.8,<4.12",
  "ruff>=0.4",
  "mypy>=1.8",
  "types-PyYAML>=6",
  "soundfile>=0.12",
  "pre-commit>=3",
  'mediapipe==0.10.21; python_version < "3.13" and platform_machine != "aarch64"',
  'mediapipe==0.10.18; python_version < "3.13" and platform_machine == "aarch64"',
]

[project.scripts]
opentalking = "apps.cli.main:main"
opentalking-api = "apps.api.main:main"
opentalking-worker = "opentalking.runtime.main:main"
opentalking-unified = "apps.unified.main:main"
opentalking-download = "apps.cli.download_models:main"
opentalking-doctor = "apps.cli.doctor:main"
opentalking-quicktalk-bench = "apps.cli.quicktalk_bench:main"
opentalking-prepare-cache = "apps.cli.prepare_cache:main"
opentalking-persona = "apps.cli.persona:main"

[tool.setuptools]
include-package-data = true

[tool.setuptools.package-data]
opentalking = [
  "assets/voices/system/*/*.json",
  "assets/voices/system/*/*.txt",
  "assets/voices/system/*/*.wav",
  "assets/reference_drivers/*.wav",
  "assets/scene_backgrounds/*.jpg",
]

[tool.setuptools.packages.find]
where = ["."]
include = ["opentalking*", "apps*"]
exclude = [
  "apps.web*",
  "apps.*.tests*",
  "tests*",
  "examples*",
  "configs*",
  "deploy*",
  "docs*",
  "scripts*",
]

[tool.ruff]
line-length = 100
target-version = "py310"

[tool.pytest.ini_options]
asyncio_mode = "auto"
testpaths = ["tests", "apps/api/tests"]
filterwarnings = [
  "ignore:Using `httpx` with `starlette.testclient` is deprecated.*",
  "ignore:Type google\\._upb\\._message\\..* uses PyType_Spec.*:DeprecationWarning",
]

[[tool.mypy.overrides]]
module = [
  "opentalking.providers.*",
  "opentalking.pipeline.*",
  "opentalking.runtime.*",
  "opentalking.voice.*",
  "apps.api.tests.*",
  "tests.*",
]
ignore_errors = true

[tool.uv]
conflicts = [
  [
    { extra = "quicktalk-cpu" },
    { extra = "quicktalk-cuda" },
  ],
  [
    { extra = "quicktalk-cpu" },
    { extra = "local-cosyvoice-service" },
  ],
  [
    { extra = "demo" },
    { extra = "local-cosyvoice-service" },
  ],
]

[[tool.uv.index]]
url = "https://pypi.tuna.tsinghua.edu.cn/simple"
default = true

[[tool.uv.index]]
url = "https://www.piwheels.org/simple"
name = "piwheels"
explicit = true