pyproject.toml

[build-system]
requires = ["setuptools>=77", "wheel"]
build-backend = "setuptools.build_meta"

[project]
name = "httk-workflow"
version = "2.1.1"
description = "Filesystem-native workflow execution for httk₂."
readme = "README.md"
license = "AGPL-3.0-or-later"
license-files = ["LICENSE"]
requires-python = ">=3.12"
authors = [{ name = "Rickard Armiento", email = "gitcommits@armiento.net" }]
dependencies = [
  # Collection attaches the DataRecord.product_of provenance edges added in 2.1.2.
  "httk-core>=2.1.2,<3",
]
keywords = [
  "Materials Discovery",
  "Materials Informatics",
  "Materials Science",
  "Machine Learning",
]
classifiers = [
  "Development Status :: 3 - Alpha",
  "Environment :: Console",
  "Intended Audience :: Science/Research",
  # POSIX rather than OS Independent: the engine reaps process groups with
  # os.killpg, relies on POSIX rename and symlink semantics for its protocol,
  # and ships Bash authoring libraries it executes.
  "Operating System :: POSIX",
  "Programming Language :: Python :: 3",
  "Programming Language :: Python :: 3 :: Only",
  "Programming Language :: Python :: 3.12",
  "Programming Language :: Unix Shell",
  "Topic :: Scientific/Engineering :: Chemistry",
  "Topic :: Scientific/Engineering :: Physics",
  "Topic :: System :: Distributed Computing",
  "Typing :: Typed",
]

[project.urls]
Homepage = "https://httk.org"
Documentation = "https://docs.httk.org/httk-workflow/"
Repository = "https://github.com/httk/httk-workflow"
Issues = "https://github.com/httk/httk-workflow/issues"

[project.optional-dependencies]
dev = [
  # Full-source type checks cover the optional workflow language integrations.
  "httk-workflow[cwl,jobflow,atomate2]",
  "pytest>=8.0",
  "pytest-xdist>=3.6",
  "ruff>=0.16,<0.17",
  "pyright",
  "mypy",
  "pydoclint>=0.9,<1"
]
docs = [
  "sphinx",
  "furo",
  "myst-parser",
  "myst-nb",
  "sphinx-autodoc-typehints",
  "sphinx-copybutton",
  "sphinx-autoapi"
]
release = [
  "build>=1.2",
  "twine>=6.0",
]
default = []
# The CWL language realization needs a CWL *parser* when a document is prepared;
# executing the normalized plan does not, so this extra belongs on the machine
# that creates the job. cwltool is deliberately absent: httk executes the plan
# on its own runner and manager, and never shells out to another engine.
# The Python Workflow Definition realization needs no extra at all — its documents
# are plain JSON, and the `python-workflow-definition` package is used as a
# stricter validator only when an environment happens to have it.
cwl = [
  "cwl-utils>=0.32,<1",
  "cwl-upgrader>=1.2,<2",
]
# The jobflow language needs no extra to prepare or collect jobs; these extras
# belong on the machine that runs the job. `jobflow` supports any jobflow Maker
# and structure inputs via pymatgen; `atomate2` additionally provides its Maker library.
jobflow = [
  "jobflow>=0.3.1,<0.4",
  "pymatgen",
]
atomate2 = [
  "atomate2>=0.1.5,<1",
  "jobflow>=0.3.1,<0.4",
]

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

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

[tool.setuptools.package-data]
"httk.workflow" = [
  "py.typed",
  "shell/*.sh",
  # The native C authoring SDK ships beside the shell libraries as one .h/.c
  # pair, designed to be vendored or compiled directly beside a runner.
  "native/c/*.h",
  "native/c/*.c",
  # The native Fortran SDK is one module wrapping that C pair over iso_c_binding;
  # it ships beside it, compiled with a runner and the C source (see docs/sdks/native_fortran_api).
  "native/fortran/*.f90",
  # The native Rust SDK is a std-only crate (no crates.io deps), a bridge client
  # like the Bash and C ones rather than FFI over them; it ships as its Cargo
  # manifest and single source file, path-depended or vendored (see docs/sdks/native_rust_api).
  "native/rust/Cargo.toml",
  "native/rust/src/*.rs",
  # The native Perl SDK is one pure-core module, loaded directly by interpreted runners.
  "native/perl/*.pm",
  # The native Ada SDK is an Ada 2012 spec/body binding over the C SDK.
  "native/ada/*.ads",
  "native/ada/*.adb",
  # The native C++ SDK is a header-only C++17 RAII wrapper over the C SDK.
  "native/cpp/*.hpp",
  # The native Java SDK is one java.base-only source file, compiled beside a runner.
  "native/java/*.java",
  "adapter_templates/*/remote.json",
  "adapter_templates/*/adapter",
  "launch_templates/*/launcher.json",
  "launch_templates/*/launcher",
]
# The bundled legacy shell runtime is data of the v1 compatibility package it now
# lives in, so a job whose httk-v1 backend points HTTK_DIR at it finds it wherever
# httk.workflow.compat.v1 was installed.
"httk.workflow.compat.v1" = [
  "v1_runtime/Execution/tasks/*.sh",
  "v1_runtime/Execution/tasks/vasp/*.sh",
  "v1_runtime/LICENSE",
  "v1_runtime/NOTICE",
  "v1_runtime/README.md",
  "v1_runtime/compat/*.sh",
]
# The Python runners are modules of the package and ship with it; the Bash runner
# beside them is data, and a job referencing pkg:httk.workflow.vasp.runners/...
# needs it installed exactly like the shell libraries above.
"httk.workflow.vasp.runners" = [
  "*.sh",
  "scripts/relaxation_report",
  "scripts/relaxation_plot",
]
# httk.workflow.languages.cwl and httk.workflow.languages.pwd follow the same rule and
# need no entry of their own: cwl_runner.py and pwd_runner.py are modules of those
# packages, so a job referencing pkg:httk.workflow.languages.cwl/... or
# pkg:httk.workflow.languages.pwd/... finds them wherever the package was installed.
# Nothing beside them is data.

[tool.pytest.ini_options]
testpaths = ["tests"]
# Match the host CPU count: a fixed 16 workers starves runner subprocesses on
# small CI hosts and causes manager deadlines to expire. Cap larger hosts at 16.
# An extended run overrides the marker selection with `-m ""`.
addopts = "-n auto --maxprocesses 16 --dist loadgroup -m 'not extended'"
markers = [
  "slow: drives real processes and waits on them; deselect with -m 'not slow'",
  "extended: full-depth test case; selected only by the extended test profile",
  # Timing-sensitive tests belong in the serial `-m timing -n 0` pass; if a
  # timing-sensitive test is missed, it remains in the parallel pass.
  "timing: asserts or depends on wall-clock timing; run serially after the parallel pass",
  "xdist_group(name): schedule related timing-sensitive tests on one worker",
]

[tool.ruff]
target-version = "py312"
src = ["src", "examples", "tests"]
line-length = 120

[tool.ruff.lint]
# TRY002: raising plain Exception is accepted httk style.
# TRY004: the ValueError/TypeError choice is public API; do not churn it.
# BLE001: broad excepts deliberately isolate plugin/loader/handler failures.
ignore = ["TRY002", "TRY004", "BLE001"]

[tool.ruff.format]
# preserve replaces black's skip-string-normalization: existing quote styles are kept.
quote-style = "preserve"

[tool.pyright]
pythonVersion = "3.12"
include = ["src", "examples", "tests"]
extraPaths = ["src", "../httk-core/src"]
venvPath = ".."
venv = ".venv"
reportMissingImports = true

[tool.mypy]
python_version = "3.12"
files = ["src", "examples", "tests"]
mypy_path = ["src", "tests"]
namespace_packages = true
explicit_package_bases = true
warn_unused_ignores = true

[[tool.mypy.overrides]]
module = ["yaml", "markdown", "docutils", "docutils.*"]
ignore_missing_imports = true

# Optional peer integrations remain type-checked whenever those distributions
# are installed; their absence must not break this module's own release gates.
[[tool.mypy.overrides]]
module = ["httk.atomistic", "httk.atomistic.*", "httk.store", "httk.store.*"]
ignore_missing_imports = true

[tool.pydoclint]
style = "sphinx"
arg-type-hints-in-docstring = false
check-return-types = false
check-yield-types = false
check-class-attributes = false
skip-checking-raises = true