social-data-attribution-w2/pyproject.toml

[project]
name = "data-attribution"
version = "0.1.0"
requires-python = ">=3.12,<3.13"
dependencies = [
    "torch==2.9.0; sys_platform == 'darwin'",
    "torch==2.9.0+cpu; sys_platform != 'darwin' and sys_platform != 'linux'",
    "torch==2.9.0+cu129; sys_platform == 'linux'",
    "numpy<2",
    "requests",
    "boto3>=1.34,<2.0",
    "python-dotenv",
    "bergson",
    "transformers>=4.57.1,<5.0.0",
    "datasets>=3.0.0",
    "pyarrow",
    "zstandard>=0.24.0",
    "accelerate>=1.11.0",
    "safetensors>=0.6.2",
    "sentencepiece>=0.2.1",
    "huggingface-hub>=0.25",
    "torchvision==0.24.0",
    "pyyaml>=6.0.0",
    "empath==0.89",
    "pytest>=9.0.2",
    "pytest-cov",
    "pandas>=2.0.0",
    "ruff",
    "jj>=2.14.0",
    "plotly>=6.5.0",
    "kaleido>=1.2.0",
    "tqdm>=4.66.0",
    "matplotlib>=3.8",
    "pillow>=10",
    "fasttext>=0.9.3",
    "boto3",
    "modal",
    "ai2-olmes",
    "vllm==0.12.0; sys_platform == 'linux'",
    "polars>=1.0.0",
    "peft>=0.15",
    "lm-eval>=0.4",
    "hydra-core==1.4.0.dev1",
    "omegaconf==2.4.0.dev3",
    "pytrec-eval-terrier>=0.5.6",
    "torch-c-dlpack-ext>=0.1.5; sys_platform == 'linux'",
    "ipykernel>=7.2.0",
    "sentence-transformers>=5.2.3",
]

[project.scripts]
data-attribution = "data_attribution.cli.main:main"
socialiqa-export = "data_attribution.data.socialiqa_export:main"
data-attribution-score = "data_attribution.scoring.cli:main"
data-attribution-precache = "data_attribution.cache.precache:main"
data-attribution-dolma-train = "data_attribution.recipes.dolma_training_index:main"
data-attribution-dolma-metadata = "data_attribution.recipes.dolma_metadata_export:main"
data-attribution-socialiqa-queries = "data_attribution.recipes.socialiqa_queries:main"
data-attribution-run-olmes-evaluation = "data_attribution.recipes.olmes.evaluation:main"
data-attribution-export-olmes-predictions = "data_attribution.recipes.olmes.predictions_export:main"
data-attribution-olmes-query-index = "data_attribution.recipes.olmes.query_index:main"
data-attribution-olmes-instruct-manifest = "data_attribution.recipes.olmes_instruct_manifest:main"
data-attribution-olmes-instruct-cot-manifest = "data_attribution.recipes.olmes_instruct_cot_manifest:main"
data-attribution-hardware-notes = "data_attribution.recipes.hardware_notes:main"
data-attribution-enrich = "data_attribution.analysis.enrich:main"
data-attribution-corpus-manifest = "data_attribution.recipes.corpus_manifest:main"
data-attribution-draw-working-sample = "data_attribution.recipes.draw_working_sample:main"
data-attribution-materialize-sample = "data_attribution.recipes.materialize_sample:main"
data-attribution-sidecar-schema-inventory = "data_attribution.recipes.sidecar_schema_inventory.core:main"
dolma-enrich-format = "dolma.cli:main"
dolma-eda = "dolma.eda.core:main"
data-attribution-pool-sample = "dolma.pool_sample.cli:main"
data-attribution-manifest-sample = "dolma.pool_sample.manifest_cli:main"
data-attribution-manifest-analysis = "dolma.manifest_analysis.cli:main"
data-attribution-manifest-analysis-modal = "dolma.manifest_analysis.modal:main"
dolma-corpus-stats = "dolma.corpus_stats.cli:main"
data-attribution-bin-analysis = "dolma.bin_analysis.cli:main"
data-attribution-weborganizer-report = "dolma.distribution_report.cli:main"
data-attribution-paper-figures = "dolma.distribution_report.cli:main"
data-attribution-quality-sidecars = "dolma.quality.cli:main"
data-attribution-quality-validation = "dolma.quality.validation.cli:main"
data-attribution-trackstar-query = "data_attribution.attribution.trackstar.queries:main"
data-attribution-trackstar-prepare = "data_attribution.attribution.trackstar.prepare:main"
data-attribution-trackstar-aggregate = "data_attribution.attribution.trackstar.aggregate:main"
data-attribution-trackstar-extract = "data_attribution.attribution.trackstar.extract:main"
data-attribution-trackstar-dot-score = "data_attribution.attribution.trackstar.dot_score:main"
data-attribution-trackstar-bin-aggregate = "data_attribution.attribution.trackstar.bin_aggregate:main"
data-attribution-trackstar-bin-aggregate-split = "data_attribution.attribution.trackstar.bin_aggregate_split:main"
data-attribution-trackstar-bin-aggregate-perquery = "data_attribution.attribution.trackstar.bin_aggregate_perquery:main"
data-attribution-trackstar-proponent-examples = "data_attribution.attribution.trackstar.proponent_examples:main"
data-attribution-socialtda-trackstar-analysis = "data_attribution.analysis.socialtda_trackstar.cli:main"

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

[tool.setuptools]
package-dir = { "" = "src" }

[tool.setuptools.packages.find]
where = ["src"]

[tool.uv]
package = true
required-environments = [
    "sys_platform == 'darwin' and platform_machine == 'arm64'",
    "sys_platform == 'linux' and platform_machine == 'x86_64'",
    "sys_platform == 'linux' and platform_machine == 'aarch64'",
]
override-dependencies = [
    "transformers>=4.57.1,<5.0.0",
]

[[tool.uv.index]]
name = "pytorch-cu129"
url = "https://download.pytorch.org/whl/cu129"
explicit = true

[[tool.uv.index]]
name = "pytorch-cpu"
url = "https://download.pytorch.org/whl/cpu"
explicit = true

[tool.uv.extra-build-dependencies]
fasttext = ["pybind11"]

[tool.uv.sources]
bergson = { path = "./bergson" }
ai2-olmes = { path = "./olmes" }
torch = [
    { index = "pytorch-cu129", marker = "sys_platform == 'linux'" },
    { index = "pytorch-cpu", marker = "sys_platform != 'linux'" },
]
torchvision = [
    { index = "pytorch-cu129", marker = "sys_platform == 'linux'" },
    { index = "pytorch-cpu", marker = "sys_platform != 'linux'" },
]

[project.optional-dependencies]
dev = []
gpu = [
    "xformers>=0.0.33.post1; sys_platform == 'linux'",
]
modal = [
    "modal>=1.1,<2.0",
]

[tool.pytest.ini_options]
testpaths = ["tests"]
addopts = ""

[dependency-groups]
dev = [
    "jupyterlab>=4.5.4",
    "matplotlib>=3.10.8",
    "seaborn>=0.13.2",
]