LMCache/pyproject.toml at dev · kanthgithub/LMCache · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
[build-system]
# Requirements for package build
# Should be mirrored in requirements/build.txt (for builds external to setuptools)
requires = [
    "ninja",
    "packaging>=24.2",
    "setuptools>=77.0.3,<81.0.0",
    "setuptools_scm>=8",
    "torch",
    "wheel",
]
build-backend = "setuptools.build_meta"

[project]
name = "lmcache"
authors = [{name = "LMCache Team", email = "lmcacheteam@gmail.com"}]
license = "Apache-2.0"
license-files = ["LICENSE"]
readme = "README.md"
description = "A LLM serving engine extension to reduce TTFT and increase throughput, especially under long-context scenarios."
classifiers = [
    "Development Status :: 3 - Alpha",
    "Operating System :: POSIX :: Linux",
    "Environment :: GPU",
    "Topic :: Scientific/Engineering :: Artificial Intelligence",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Python :: 3.12",
    "Programming Language :: Python :: Implementation :: CPython",
]
requires-python = ">=3.10,<3.13"
dynamic = ["dependencies", "version"]

[project.scripts]
lmcache_v0_server="lmcache.server.__main__:main"
lmcache_server="lmcache.v1.server.__main__:main"
lmcache_controller="lmcache.v1.api_server.__main__:main"

[project.urls]
homepage = "https://docs.lmcache.ai"
source = "https://github.com/LMCache/LMCache"
issues = "https://github.com/LMCache/LMCache"

[tool.setuptools_scm]
version_file = "lmcache/_version.py"
# do not include +gREV local version, required for Test PyPI upload
local_scheme = "no-local-version"

[tool.setuptools.dynamic]
dependencies = { file = ["requirements/common.txt"] }

[tool.setuptools.packages.find]
where = [""]
include = ["lmcache", "lmcache*"]

[tool.ruff]
# same as Black's default line length
line-length = 88

[tool.ruff.lint]
select = [
    # pycodestyle
    "E",
    # Pyflakes
    "F",
    # pyupgrade
    # "UP",
    # flake8-bugbear
    "B",
    # flake8-simplify
    #"SIM",
    # Ruff does not support isort's import_headings feature, yet.
    # "I",
    # flake8-logging-format
    #"G",
]
ignore = [
    # star imports
    "F405", "F403",
    # lambda expression assignment
    "E731",
    # Loop control variable not used within loop body
    "B007",
    # f-string format
    "UP032",
]

[tool.ruff.lint.isort]
# same as .isort.cfg
from-first = true
# not supported yet
# import-heading-future=Future
# import-heading-stdlib=Standard
# import-heading-thirdparty=Third Party
# import-heading-firstparty=First Party
# import-heading-localfolder=Local

[tool.mypy]
modules = ["lmcache", "tests"]

# TODO: tighten MyPy checks by enabling these checks over time.
disable_error_code = [
    "annotation-unchecked",
    "union-attr",
    "var-annotated",
    "arg-type",
    "call-arg",
    "import-untyped",
    "attr-defined",
    "return-value",
    "assignment",
    "call-overload",
    "misc",
]

ignore_missing_imports = true
explicit_package_bases = true

# TODO: tighten MyPy checks by enabling these checks over time.
check_untyped_defs = false
disallow_incomplete_defs = false
disallow_untyped_defs = false
disallow_untyped_calls = false
warn_return_any = false

follow_imports = "silent"

[tool.cibuildwheel]
build = "cp3*-manylinux_x86_64"
skip = "pp*"

# see https://developer.nvidia.com/cuda-gpus for compute capabilities
# "CUDA-Enabled Datacenter Products"
# 7.0: V100
# 7.5: T4
# 8.0: A100, A30
# 8.6: A40, A10, A16, A2
# 8.9: L4, L40, L40S
# 9.0: H100
environment = {TORCH_CUDA_ARCH_LIST = "7.0;7.5;8.0;8.6;8.9;9.0"}

# Use the PyTorch manylinux image version '2_28' that contains CUDA 12.8
# and torch 2.7 supports (https://pypi.org/project/torch/2.7.0/#files)
manylinux-x86_64-image = "docker.io/pytorch/manylinux2_28-builder:cuda12.8"

[tool.cibuildwheel.linux]
repair-wheel-command = """
auditwheel repair \
  --plat manylinux_2_28_x86_64 \
  --exclude libtorch.so \
  --exclude libtorch_cuda.so \
  --exclude libtorch_python.so \
  --exclude libtorch_cpu.so \
  --exclude libc10.so \
  --exclude libc10_cuda.so \
  --exclude libcudart.so.12 \
  -w {dest_dir} {wheel}
"""