pyproject.toml

[build-system]
requires = ["hatchling"]
build-backend = "hatchling.build"

[project]
name = "selectron"
dynamic = ["version"]
description = 'AI web parsing library & CLI'
readme = "README.md"
requires-python = ">=3.12"
license = "MIT"
keywords = []
authors = [{ name = "Substrate Labs Inc.", email = "ben@substrate.run" }]
classifiers = [
    "Development Status :: 4 - Beta",
    "Programming Language :: Python",
    "Programming Language :: Python :: 3.12",
    "Programming Language :: Python :: Implementation :: CPython",
    "Programming Language :: Python :: Implementation :: PyPy",
]
dependencies = [
    "beautifulsoup4>=4.13.4",
    "bs4>=0.0.2",
    "duckdb>=1.2.2",
    "httpx>=0.28.1",
    "lxml>=5.4.0",
    "markdownify>=1.1.0",
    "pillow>=10.4.0",
    "polars>=1.29.0",
    "psutil>=7.0.0",
    "pyarrow>=20.0.0",
    "pydantic-ai>=0.1.9",
    "pydantic>=2.11.4",
    "python-dateutil>=2.9.0.post0",
    "rich>=14.0.0",
    "textual-dev>=1.7.0",
    "textual>=3.2.0",
    "websockets>=15.0.1",
]

[project.scripts]
selectron = "selectron.entry_points:start"

[project.optional-dependencies]
dev = [
    "pytest>=7.0", 
    "pytest-cov>=4.0",
    "pytest-xdist>=3.0",
    "pytest-mock>=3.10.0",
    "pytest-asyncio>=0.21.0",
    "pytest-httpx>=0.24.0",
    "ruff>=0.3.0",
    "pyright>=1.1.350",
    "term-image>=0.7.2"
]

[project.urls]
Documentation = "https://github.com/SubstrateLabs/selectron"
Issues = "https://github.com/SubstrateLabs/selectron/issues"
Source = "https://github.com/SubstrateLabs/selectron"

[tool.hatch.version]
path = "src/selectron/__about__.py"

[tool.hatch.build]
include = [
  "src/**/*.py",
  "src/selectron/dom/buildDomTree.js",
  "src/selectron/parsers/**/*.json"
]
exclude = [
  "tests/**"
]

[tool.hatch.build.targets.wheel]
packages = ["src/selectron"]

[tool.hatch.envs.types]
extra-dependencies = ["mypy>=1.0.0"]
[tool.hatch.envs.types.scripts]
check = "mypy --install-types --non-interactive {args:src/selectron tests}"

[tool.coverage.run]
source_pkgs = ["selectron", "tests"]
branch = true
parallel = true
omit = ["src/selectron/__about__.py"]

[tool.coverage.paths]
selectron = ["src/selectron", "*/cli/src/selectron"]
tests = ["tests", "*/cli/tests"]

[tool.coverage.report]
exclude_lines = ["no cov", "if __name__ == .__main__.:", "if TYPE_CHECKING:"]

[tool.pytest.ini_options]
pythonpath = ["src"]
asyncio_mode = "strict"
asyncio_default_fixture_loop_scope = "function"
# Logging Configuration
log_cli = true
log_cli_level = "DEBUG"
log_cli_format = "%(asctime)s [%(levelname)s] [%(name)s] %(message)s" # Use standard level names
log_cli_date_format = "%H:%M:%S"
log_file = ".pytest_logs/test_session.log" # Log file path
log_file_level = "DEBUG" # Log everything to the file
log_file_format = "%(asctime)s [%(levelname)-8s] [%(name)s:%(lineno)d] %(message)s" # More detail in file
log_file_date_format = "%Y-%m-%d %H:%M:%S"
log_file_mode = "w" # Overwrite log file each run

[tool.pyright]
include = ["src", "tests"]
exclude = ["**/__pycache__", "build", "dist"]
reportMissingImports = true
reportMissingTypeStubs = false
pythonVersion = "3.11"
typeCheckingMode = "basic"
reportUnknownMemberType = false
reportUnknownArgumentType = false
reportUnknownVariableType = false
reportUnknownLambdaType = false
reportPrivateUsage = false
reportUnnecessaryIsInstance = false # trips up on bs4 tags
reportOptionalMemberAccess = true
reportOptionalCall = true
reportOptionalOperand = true
reportUnnecessaryCast = true
reportAssertAlwaysTrue = true
reportIncompatibleMethodOverride = true
reportUninitializedInstanceVariable = true
reportConstantRedefinition = true
reportCallInDefaultInitializer = true
reportUnusedClass = true
reportDuplicateImport = true
reportInvalidStringEscapeSequence = true
reportMissingTypeArgument = false # too many errors
reportMissingSuperCall = false # not needed if not inheriting
reportImplicitStringConcatenation = false # this is fine
reportUnnecessaryTypeIgnoreComment = false # doesn't work well with other rules
reportUnusedVariable = true
reportPropertyTypeMismatch = true
reportUnusedCoroutine = true
reportUnusedImport = true
reportIncompatibleVariableOverride = true
reportFunctionMemberAccess = true
reportInvalidTypeForm = true
reportUnnecessaryComparison = false # annoying

[tool.ruff]
line-length = 100
target-version = "py311"
# Add these common directories to exclude
extend-exclude = [
    ".git",
    ".venv",
    "__pycache__",
    "build",
    "dist",
]

[tool.ruff.lint]
select = [
    "E",   # pycodestyle errors
    "F",   # pyflakes 
    "I",   # isort
]
extend-select = [
    "B",   # flake8-bugbear (optional)
    "C4",  # flake8-comprehensions (optional)
]
ignore = [
    "E501",    # Line too long
]

[tool.ruff.lint.isort]
known-first-party = ["selectron"]

[tool.ruff.format]
quote-style = "double"
line-ending = "auto"
indent-style = "space"