pyproject.toml

[project]
name = "CNVkit"
description = "Copy number variation toolkit for high-throughput sequencing."
readme = "README.rst"
#license = "Apache-2.0"
license = {"text" = "Apache-2.0"}
authors = [
  {name = "Eric Talevich", email = "me+code@etal.mozmail.com"}
]
maintainers = [
  {name = "Eric Talevich", email = "me+code@etal.mozmail.com"}
]
classifiers = [
  "Development Status :: 4 - Beta",
  "Environment :: Console",
  "Intended Audience :: Science/Research",
  "Intended Audience :: Healthcare Industry",
  "Operating System :: MacOS :: MacOS X",
  "Operating System :: POSIX",
  "Operating System :: POSIX :: Linux",
  "Operating System :: Unix",
  "Programming Language :: Python",
  "Programming Language :: Python :: 3",
  "Programming Language :: Python :: 3.11",
  "Programming Language :: Python :: 3.12",
  "Programming Language :: Python :: 3.13",
  "Programming Language :: Python :: 3.14",
  "Topic :: Scientific/Engineering",
  "Topic :: Scientific/Engineering :: Bio-Informatics",
  "Topic :: Scientific/Engineering :: Medical Science Apps.",
  "Topic :: Scientific/Engineering :: Visualization"
]
requires-python = ">=3.11"
dynamic = ["version", "dependencies", "optional-dependencies"]

[project.urls]
homepage = "https://github.com/etal/cnvkit"
documentation = "https://cnvkit.readthedocs.io"
repository = "https://github.com/etal/cnvkit"
changelog = "https://github.com/etal/cnvkit/releases"

[project.scripts]
"cnvkit.py" = "cnvlib.cnvkit:main"
"cnv_annotate.py" = "cnvlib.cli.cnv_annotate:main"
"cnv_expression_correlate.py" = "cnvlib.cli.cnv_expression_correlate:main"
"cnv_gene_info.py" = "cnvlib.cli.cnv_gene_info:main"
"cnv_updater.py" = "cnvlib.cli.cnv_updater:main"
"genome_instability_index.py" = "cnvlib.cli.genome_instability_index:main"
"guess_baits.py" = "cnvlib.cli.guess_baits:main"
"reference2targets.py" = "cnvlib.cli.reference2targets:main"
"skg_convert.py" = "skgenome.cli.skg_convert:main"
# XXX TODO -- figure out whether this script should and could be included
# in the installation
#"snpfilter.sh" = "scripts/snpfilter.sh"

[build-system]
# Minimum requirements for the build system to execute (PEP 508)
requires = ["setuptools"]
build-backend = "setuptools.build_meta"

[tool.setuptools]
packages = ["cnvlib", "cnvlib.cli", "cnvlib.segmentation", "skgenome", "skgenome.cli", "skgenome.tabio"]

[tool.setuptools.dynamic]
version = {attr = "cnvlib._version.__version__"}
dependencies = {file = "requirements/core.txt"}
optional-dependencies = {test = {file = "requirements/tests.txt"}}

[tool.pytest.ini_options]
testpaths = ["test"]
markers = [
    "slow: tests that take >5 seconds (skip with -m 'not slow')",
    "hypothesis: property-based tests using Hypothesis (skip with -m 'not hypothesis')",
]
filterwarnings = [
    "error",
    # bioframe <=0.8.0 uses df.groupby(['col']).groups with a single-element
    # list, which pandas 3.0 deprecates via Pandas4Warning.
    "ignore:In a future version, the keys of:Warning",
    # Hypothesis internal warnings that would otherwise be promoted to errors
    "ignore::hypothesis.errors.NonInteractiveExampleWarning",
]

[tool.coverage.run]
branch = true
source = ["cnvlib", "skgenome"]

[tool.coverage.paths]
source = ["cnvlib", "skgenome"]

[tool.coverage.report]
precision = 2
show_missing = true
skip_covered = false
exclude_lines = [
    "pragma: no cover",
    "raise NotImplementedError",
    "if __name__ == .__main__.:",
    "if TYPE_CHECKING:",
    "@abstractmethod",
]

[tool.coverage.html]
directory = "htmlcov"

[tool.ruff]
target-version = "py311"
line-length = 88

[tool.ruff.lint]
select = [
    "E",       # pycodestyle errors
    "W",       # pycodestyle warnings
    "F",       # pyflakes (incl. F401 unused-import, F841 unused-variable)
    "UP",      # pyupgrade
    "B",       # flake8-bugbear
    "NPY",     # numpy rules
    "PD",      # pandas-vet
    "RUF",     # ruff's extra rules (incl. RUF100 unused-noqa)
    "TC",      # flake8-type-checking
    "PERF",    # perflint (numpy/loop performance)
    "C4",      # flake8-comprehensions
    "PIE",     # flake8-pie (misc. quality)
    "I",       # isort (import ordering)
    "Q",       # flake8-quotes (defensive: format.quote-style enforces double)
    "PLC0415", # pylint: imports outside top level (lazy imports must justify with noqa)
    # Evaluated and deferred -- noqa cost outweighs the signal here, the
    # 3-reviewer code-review protocol catches the real cases better:
    #   "SIM" -- flake8-simplify; SIM401/SIM115 produce false positives on
    #            domain types (GenomicArray has no .get(); ctx-via-with is
    #            invisible to it) and SIM108 compresses readable per-branch
    #            comments into inline annotations.
    #   "PT"  -- flake8-pytest-style; requires unittest.TestCase -> pytest migration
    #   "N"   -- pep8-naming; collides with scientific names (X, Y, log2, chi2)
    #   "D"   -- pydocstyle; needs a docs-style policy decision
    #   "ERA" -- commented-out code; needs per-occurrence judgment in this codebase
    #   "RET" -- return-statement style; RET504 (assign-then-return) is stylistic
    #   "ARG" -- unused-args; mix of real bugs and intentional API contracts (ABC methods, plugin hooks)
    #   "ANN" -- flake8-annotations; defer until full annotation coverage decided
]
ignore = [
    "E402",  # import not at top -- some lazy imports remain legitimate
    "E501",  # line too long -- too noisy without a reflow pass
    "E731",  # inline lambda -- common in argparse setups
    "UP031", # % format -- kept; modernization is a separate focused PR
    "B007",  # loop variable not used in body -- common idiom in this code
    "B023",  # def closing over loop variable -- intentional in some places
    "PD009", # use .iloc instead of .iat -- iat is faster for scalar access
]

[tool.ruff.lint.per-file-ignores]
# Test suites use `from cnvlib import (access, antitarget, ...)` smoke-test
# blocks (see CLAUDE.md "Imports in Tests"). The imports are intentionally
# "unused" -- they verify each submodule loads cleanly. F401 must not fire
# on these blocks but should still catch real unused imports in cnvlib/.
"test/*" = ["F401"]

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

[tool.mypy]
# numpy >= 2.5 ships PEP 695 ``type`` aliases in its stubs, which mypy only
# parses with a 3.12+ target; the type-check CI job runs on 3.12. (The runtime
# baseline stays 3.11, enforced by ruff target-version and the py311 test job.)
python_version = "3.12"
packages = ["cnvlib", "skgenome"]
ignore_missing_imports = true
check_untyped_defs = true
warn_no_return = true
warn_return_any = true
warn_unused_configs = true
warn_unreachable = true
enable_error_code = ["ignore-without-code"]

[tool.bandit]
# Skip security rules inappropriate for bioinformatics toolkit that needs to call external tools
skips = [
    "B101",  # assert_used - asserts are an intentional idiom here, incl. type narrowing
    "B404",  # subprocess module - needed for calling bioinformatics tools like samtools, R
    "B603",  # subprocess without shell=True - this is actually good security practice
    "B607",  # partial executable path - standard practice for bioinformatics pipelines
]