pyproject.toml

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

[tool.setuptools.packages.find]
include = ["giskard*"]
exclude = ["docs*", "tests*"]


[tool.setuptools.package-data]
giskard = [
    "**/*.md",
    "**/*.html",
    "**/*.css",
    "**/*.js",
    "**/*.json",
    "**/*titanic.csv",
    "**/*prompts.csv",
    "**/*.txt",
    "**/injection_data/*.csv",
]





[tool.pdm.scripts]
_.env = { GSK_DISABLE_ANALYTICS = "True", GISKARD_DEV_MODE = "True" }
# add "-n auto" to the pytest command to parallelize the execution
test.cmd = "pdm base-test tests/"
# for some reason github runners don't work when calling 'pdm test -m "not slow"'
base-test.cmd = "pytest -c pyproject.toml --cov=giskard --cov-report=xml --cov-append --disable-warnings --no-header -vv --durations=0"
test-fast.cmd = "pdm base-test -n auto -m 'not slow and not concurrency' --use-subprocess tests"
test-slow.cmd = "pdm base-test -m 'slow and not concurrency' --use-subprocess"
test-worker.cmd = "pdm base-test -m 'concurrency' tests/"
lint = "ruff giskard tests"
doc = "sphinx-build docs docs/_build/html"
#  -W --keep-going -n => temporary disable
check-doc = "python ./docs/scrapper.py"
watch-doc = "python -m sphinx_autobuild --watch giskard docs docs/_build/html"
clean = "rm -rf coverage.xml coverage* .coverage*"
notebook = "jupyter notebook --ip 0.0.0.0 --port 8888 --no-browser --notebook-dir ./docs/reference/notebooks --NotebookApp.token=''"
check-deps = "deptry ."
debug-worker = "python3 -Xfrozen_modules=off -m debugpy --listen localhost:5678 --wait-for-client giskard/cli.py worker start"
worker = "python3 giskard/cli.py worker start"
debug-internal-worker = "python3 -Xfrozen_modules=off -m debugpy --listen localhost:5678 --wait-for-client giskard/cli.py worker start -s"
internal-worker = "python3 giskard/cli.py worker start -s"
update-lock = "docker run --platform linux/amd64 --rm -it -w /work -v ${PWD}/pdm.lock:/work/pdm.lock -v ${PWD}/pyproject.toml:/work/pyproject.toml python:3.10-slim bash -c \"apt update && apt install curl -y && curl -sSL https://pdm.fming.dev/install-pdm.py | python3 - && /root/.local/bin/pdm lock -G :all -v\""

[tool.pdm.dev-dependencies]
dev = [
    "typing-extensions",
    "jupyter>=1.0.0",
    "jupyterlab>=3.4.2",
    "pre-commit>=2.19.0",
    "mypy>=0.982",
    "deptry>=0.5.13",
    "ruff>=0.0.271",
    "mlflow>2",
    "black[d]>=22.12.0",
    "pip>=23.1.2",
    "pytest-asyncio>=0.21.1",
    "pydantic>=2",
    "avidtools",
    "mistralai>=0.1.8",
    "boto3>=1.34.88",
]
ml_runtime = [
    "langchain>=0.0.275",
    "nltk>=3.8.1",
    "xgboost>=1.7.5",
    "lightgbm>=3.3.5",
    "imbalanced-learn>=0.12.0",
    "catboost>=1.1.1",
    "requests-mock>=1.10.0",
    "tensorflow-hub>=0.12.0",
    "transformers>=4.34.1",
    "sentencepiece",                                                                      # needed for some transformers stuff with tokenizer
    "torch>=2.0.0",
    "torchdata>=0.6.0",
    "portalocker>=2.0.0",                                                                 # Needed by torchdata for test_newspaper_classification_pytorch_dataset
    "torchtext>=0.15.1",
    "tensorflow-macos==2.14.0; sys_platform == 'darwin' and platform_machine == 'arm64'",
    "tensorflow==2.14.0",
    # tensorflow-text is only available on linux or mac, but not for arm64 nor windows
    "tensorflow-text==2.14.0; python_version < '3.11' and (sys_platform == 'linux' or sys_platform == 'darwin') and platform_machine == 'x86_64'",
    "mlflow>2,<2.12", # version 2.12 throws error with imbalance-learn
    "wandb",
    "tensorflow-io-gcs-filesystem<0.32; platform_machine != 'arm64'", # Tensorflow io does not work for windows from 0.32, but does not work for arm64 before...
]
test = [
    "pytest-cov>=4.0.0",
    "pytest>=7.1.2",
    "pytest-xdist>=3.3.1",
    "polyfactory",
    "pytest-env>=1.1.0",
    "openai>=1",
    "pytest-memray; sys_platform == 'linux' or sys_platform == 'darwin'",
    "pytest-reportlog>=0.4.0",
    "shap<0.45",                                                          # Fixing this to avoid changed on insights
]
doc = [
    "furo>=2023.5.20",
    "myst-parser>=1.0.0",
    "sphinx-autobuild>=2021.3.14",
    "sphinx>=6.1.3",
    "sphinxcontrib-napoleon>=0.7",
    "sphinx-autoapi>=2.1.0",
    "sphinx-rtd-theme>=1.2.0",
    "sphinx-tabs>=3.4.1",
    "sphinx-design>=0.4.1",
    "sphinx-copybutton>=0.5.2",
    "sphinx-click>=4.4.0",
    "nbsphinx>=0.9.2",
    "ipython==8.12.0",
    "scrapy",
    "requests",
]

[project.scripts]
giskard = "giskard.cli:cli"

[project.entry-points."mlflow.model_evaluator"]
giskard = "giskard.integrations.mlflow.giskard_evaluator:GiskardEvaluator"

[project.urls]
"Homepage" = "https://giskard.ai"
"Source Code" = "https://github.com/Giskard-AI/giskard"
"Bug Tracker" = "https://github.com/Giskard-AI/giskard/issues"
"Documentation" = "https://docs.giskard.ai/"
"Discord" = "https://discord.gg/ABvfpbu69R"
"Linkedin" = "https://www.linkedin.com/company/giskard-ai"
"Mastodon" = "https://fosstodon.org/@Giskard"
"Twitter" = "https://twitter.com/giskard_ai"


[project]
name = "giskard"
readme = "README.md"
license = { text = "Apache Software License 2.0" }
version = "2.12.0"
description = "The testing framework dedicated to ML models, from tabular to LLMs"
authors = [{ name = "Giskard AI", email = "hello@giskard.ai" }]
keywords = ["Artificial Intelligence", "Machine Learning", "Quality", "MLOps"]

# Pypi classifiers: https://pypi.org/classifiers/
classifiers = [
    "Development Status :: 4 - Beta",
    "Intended Audience :: Developers",
    "Intended Audience :: Science/Research",
    "Operating System :: OS Independent",
    "Topic :: Scientific/Engineering :: Artificial Intelligence",
    "License :: OSI Approved :: Apache Software License",
    "Programming Language :: Python :: 3.9",
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
]

requires-python = ">=3.9, <3.12"
dependencies = [
    "cloudpickle>=1.1.1",
    "zstandard>=0.10.0",
    "mlflow-skinny>=2",
    "numpy>=1.22.0",
    "scikit-learn>=1.0",
    "scipy>=1.7.3, <1.12.0",
    "mixpanel>=4.4.0",
    "requests>=2.19",
    "pydantic<3,>1",
    "pandas>=1.3.4",
    "xxhash>=3.2.0",
    "langdetect>=1.0.9",
    "chardet",                  # text metadata
    "jinja2>=3",                # scan template
    "requests-toolbelt>=0.9.1",
    "setuptools",               # used for pkg_resources
    "typing_extensions",        # used in registry/decorators, for python <3.10
    "pyyaml",
    "packaging",                # used in settings
    "markdown",                 # needed for display of scan results in notebook
    "colorama",                 # needed for the scan
    "num2words>=0.5.13",
    "griffe>=0.36.9",
    "sentry-sdk>=1.40.5",
]

[project.optional-dependencies]
llm = [
    "openai",
    "evaluate>=0.4.1",
    "bert-score>=0.3.13",
    "tenacity>=4.11.0",
    "faiss-cpu>=1.7.4",
    "bokeh>=3.3.4",
    "umap-learn>=0.5.5",
    "scikit-learn>=1.3.2",
]

hub = [
    # Worker deps
    "websockets>=11.0",                        # Not sure what is the minimal version, let's say 11 to handle python 3.11
    "uvloop>=0.18.0; sys_platform != 'win32'",
    # Server deps
    "tenacity>=4.11.0",
    "psutil>=5.4.6",
    "click>=7.0",
    "docker>=6.0.0",
    "shap>=0.41.0",
    "llvmlite>=0.40.0",    # Needed to be installable for Poetry
    "numba>=0.50.0",       # Needed to be installable for Poetry
    "lockfile>=0.12.2",
    "pycryptodome>=3.6.1",
    "pyngrok>=6.0.0",
    "daemon>=1.2",
]
talk = [
    "openai",
    "tenacity>=4.11.0",
    "shap<0.44.1,>=0.41.0",
    "tabulate==0.9.0",
]

[tool.ruff]
line-length = 120
ignore = ["E501"]

[tool.black]
# https://github.com/psf/black
target-version = ['py39', 'py310', 'py311']
line-length = 120
color = true

exclude = '''
/(
    \.git
    | \.hg
    | \.mypy_cache
    | \.tox
    | \.venv
    | _build
    | buck-out
    | build
    | dist
    | env
    | venv
    | giskard/ml_worker/generated
)/
'''

[tool.isort]
# https://github.com/timothycrosley/isort/
py_version = 38
line_length = 100

known_typing = [
    "typing",
    "types",
    "typing_extensions",
    "mypy",
    "mypy_extensions",
]
sections = [
    "FUTURE",
    "TYPING",
    "STDLIB",
    "THIRDPARTY",
    "FIRSTPARTY",
    "LOCALFOLDER",
]
include_trailing_comma = true
profile = "black"
multi_line_output = 3
indent = 4
color_output = true

[tool.mypy]
# https://mypy.readthedocs.io/en/latest/config_file.html#using-a-pyproject-toml-file
python_version = 3.8
pretty = true
show_traceback = true
color_output = true
show_column_numbers = true
show_error_codes = true
show_error_context = true
plugins = ["pydantic.mypy"]

allow_redefinition = false
check_untyped_defs = false       # TODO: Andrey: fix it gradually
disallow_untyped_defs = false    # TODO: Andrey: fix it gradually
disallow_any_generics = false    # TODO: Andrey: fix it gradually
disallow_incomplete_defs = false # TODO: Andrey: fix it gradually
ignore_missing_imports = true
implicit_reexport = false
no_implicit_optional = false     # TODO: Andrey: fix it gradually
strict_equality = false          # TODO: Andrey: fix it gradually
strict_optional = false          # TODO: Andrey: fix it gradually
warn_no_return = false
warn_redundant_casts = true
warn_return_any = false
warn_unreachable = true
warn_unused_configs = false
warn_unused_ignores = false

[tool.pytest.ini_options]
# https://docs.pytest.org/en/6.2.x/customize.html#pyproject-toml
# Directories that are not visited by pytest collector:
testpaths = ["tests"]
norecursedirs = [
    "hooks",
    "*.egg",
    ".eggs",
    "dist",
    "build",
    "docs",
    ".tox",
    ".git",
    "__pycache__",
]
doctest_optionflags = [
    "ELLIPSIS",
    "NUMBER",
    "NORMALIZE_WHITESPACE",
    "IGNORE_EXCEPTION_DETAIL",
]
markers = [
    "slow: marks tests as slow (deselect with '-m \"not slow\"')",
    "concurrency: run MLWorker pool tests",
    "memory_expensive: memory consuming tests",
]

# Extra options:
addopts = [
    "--strict-markers",
    "--tb=short",
    "--doctest-modules",
    "--doctest-continue-on-failure",
]

env = ["WANDB_MODE=disabled"]

[tool.deptry]
extend_exclude = [".history"]
ignore_notebooks = true