-
Notifications
You must be signed in to change notification settings - Fork 72
Expand file tree
/
Copy pathpyproject.toml
More file actions
158 lines (146 loc) · 5.64 KB
/
Copy pathpyproject.toml
File metadata and controls
158 lines (146 loc) · 5.64 KB
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
[project]
name = "openhands-benchmarks"
version = "0.1.0"
description = "OpenHands Benchmarks"
readme = "README.md"
requires-python = ">=3.12"
dependencies = [
"datasets",
"huggingface-hub>=1.10.1",
"jinja2",
"pandas",
"Pillow",
"toml",
"tqdm",
"openhands-sdk",
"openhands-tools",
"unidiff>=0.7.5,<0.8.0",
"openhands-agent-server",
"openhands-workspace",
"modal>=1.1.4",
"swebench==4.1.0",
"swesmith>=0.0.9",
"commit0",
"pytest-json-report",
# SWT-Bench dependencies (since the git package doesn't install them properly)
"requests",
"docker",
"python-dotenv",
"fire",
"editdistance",
"GitPython",
"pyright[nodejs]>=1.1.405",
"docker-registry-client>=0.5.2",
"deprecation>=2.1.0",
"fastmcp>=2.11.3",
"httpx>=0.27.0",
"litellm==1.84.1",
"pydantic>=2.12.0",
"python-frontmatter>=1.1.0",
"python-json-logger>=3.3.0",
"tenacity>=9.1.2",
"websockets>=12",
"lmnr>=0.7.41",
# TODO Remove the macOS exclusion once https://github.com/multi-swe-bench/multi-swe-bench/pull/93 is merged and released
"multi-swe-bench>=1.1.1; sys_platform != 'darwin'",
"swt-bench @ git+https://github.com/logic-star-ai/swt-bench.git@5fdcd446ff05e248ecfffc19d560a210699f71f8",
# See https://github.com/OpenHands/benchmarks/issues/449
"hf-xet>=1.1.10,<2.0",
# ProgramBench upstream harness — provides the `programbench` CLI used
# by `programbench-eval`, plus task metadata loaders we use during
# inference. Pinned by minor version while the upstream API stabilises.
"programbench>=1.0,<2.0; sys_platform != 'darwin'",
]
[project.scripts]
validate-cfg = "benchmarks.scripts.validate_cfg:main"
harbor-infer = "benchmarks.harbor.run_infer:main"
harbor-eval = "benchmarks.harbor.eval_infer:main"
swebench-infer = "benchmarks.swebench.run_infer:main"
swtbench-infer = "benchmarks.swtbench.run_infer:main"
swebench-eval = "benchmarks.swebench.eval_infer:main"
swebenchpro-infer = "benchmarks.swebenchpro.run_infer:main"
swebenchpro-eval = "benchmarks.swebenchpro.eval_infer:main"
swtbench-eval = "benchmarks.swtbench.eval_infer:main"
swtbench-list-images = "benchmarks.swtbench.image_utils:main"
swtbench-build-eval-images = "benchmarks.swtbench.build_eval_env_images:main"
gaia-infer = "benchmarks.gaia.run_infer:main"
gaia-eval = "benchmarks.gaia.eval_infer:main"
commit0-infer = "benchmarks.commit0.run_infer:main"
commit0-eval = "benchmarks.commit0.eval_infer:main"
multiswebench-infer = "benchmarks.multiswebench.run_infer:main"
multiswebench-eval = "benchmarks.multiswebench.eval_infer:main"
openagentsafety-infer = "benchmarks.openagentsafety.run_infer:main"
openagentsafety-eval = "benchmarks.openagentsafety.eval_infer:main"
swebenchmultimodal-infer = "benchmarks.swebenchmultimodal.run_infer:main"
swebenchmultimodal-eval = "benchmarks.swebenchmultimodal.eval_infer:main"
swebenchmultilingual-infer = "benchmarks.swebenchmultilingual.run_infer:main"
swebenchmultilingual-eval = "benchmarks.swebenchmultilingual.eval_infer:main"
swefficiency-infer = "benchmarks.swefficiency.run_infer:main"
terminalbench-infer = "benchmarks.terminalbench.run_infer:main"
terminalbench-eval = "benchmarks.terminalbench.eval_infer:main"
skillsbench-infer = "benchmarks.skillsbench.run_infer:main"
skillsbench-eval = "benchmarks.skillsbench.eval_infer:main"
hybridgym-funclocalize-infer = "benchmarks.hybridgym_funclocalize.run_infer:main"
hybridgym-funclocalize-eval = "benchmarks.hybridgym_funclocalize.eval_infer:main"
hybridgym-depsearch-infer = "benchmarks.hybridgym_depsearch.run_infer:main"
hybridgym-depsearch-eval = "benchmarks.hybridgym_depsearch.eval_infer:main"
hybridgym-funcgen-infer = "benchmarks.hybridgym_funcgen.run_infer:main"
hybridgym-funcgen-eval = "benchmarks.hybridgym_funcgen.eval_infer:main"
hybridgym-issuelocalize-infer = "benchmarks.hybridgym_issuelocalize.run_infer:main"
hybridgym-issuelocalize-eval = "benchmarks.hybridgym_issuelocalize.eval_infer:main"
swesmith-infer = "benchmarks.swesmith.run_infer:main"
swesmith-eval = "benchmarks.swesmith.eval_infer:main"
programbench-infer = "benchmarks.programbench.run_infer:main"
programbench-eval = "benchmarks.programbench.eval_infer:main"
[build-system]
requires = ["setuptools>=61.0", "wheel"]
build-backend = "setuptools.build_meta"
[tool.setuptools.packages.find]
where = ["."]
include = ["benchmarks"]
[tool.setuptools]
# Install the top-level sitecustomize module so Python auto-loads our Modal logging patch.
py-modules = ["sitecustomize"]
[dependency-groups]
dev = [
"pre-commit>=4.3.0",
"psutil>=7.0.0",
"pyright>=1.1.405",
"ruff>=0.12.10",
"pycodestyle>=2.12.0",
"pytest>=8.0.0",
"pytest-asyncio>=1.0.0",
"pytest-cov>=6.0.0",
"pytest-forked>=1.6.0",
]
[tool.pytest.ini_options]
asyncio_mode = "auto"
# Ruff configuration
[tool.ruff]
target-version = "py312"
line-length = 88
exclude = ["legacy"]
[tool.ruff.format]
quote-style = "double"
indent-style = "space"
[tool.ruff.lint]
select = ["E", "F", "I"]
ignore = ["E501"]
[tool.ruff.lint.isort]
known-first-party = ["benchmarks", "openhands"]
combine-as-imports = true
force-single-line = false
lines-after-imports = 2
[tool.uv.sources]
openhands-sdk = { workspace = true }
openhands-tools = { workspace = true }
openhands-workspace = { workspace = true }
openhands-agent-server = { workspace = true }
swt-bench = { git = "https://github.com/logic-star-ai/swt-bench.git", rev = "5fdcd446ff05e248ecfffc19d560a210699f71f8" }
[tool.uv.workspace]
members = [
"vendor/software-agent-sdk/openhands-sdk",
"vendor/software-agent-sdk/openhands-tools",
"vendor/software-agent-sdk/openhands-workspace",
"vendor/software-agent-sdk/openhands-agent-server",
]