Files
ocp/pyproject.toml
stroblme c5f3248fbc tools/bench_ext.py: measure the bulk APIs against the loops they replace
Keeps the numbers next to the code rather than in a commit message, since the
adoption decisions turn on them. Two results are not what the plan assumed:

- face_surface_props gains *nothing* from leaving Python — the serial C++ loop
  costs exactly what the Python loop costs. That agrees with the app's own
  profile, which found SurfaceProperties to be 94 % of face_candidate_anchors:
  the time is inside the kernel, so the interpreter round trip was never the
  problem. The whole 10.9x is OSD_Parallel, reachable only because the GIL is
  released.
- BinTools is already fast enough. Dropping the BytesIO detour saves nothing
  measurable, so the pools' overhead is not in serialisation — worth knowing
  before rewriting seventeen call sites for it.

Mesh extraction is the clear win at 35x, where the Python loop overhead really
does dominate; edge polylines 2.6x.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DfriM8XUkn7uYf5Dwe2xo6
2026-08-10 20:48:45 +02:00

47 lines
1.9 KiB
TOML

[build-system]
requires = ["scikit-build-core>=0.11", "nanobind>=2.7"]
build-backend = "scikit_build_core.build"
[project]
name = "n3xd-ocp"
# <occt-version>.N, so the kernel a wheel wraps is readable from its version
# alone and occt_version() keeps reporting the truth. Iteration builds carry a
# .devN suffix: the registry never allows republishing a version, and its
# cleanup rule collects only the dev ones. CMake asserts the prefix matches
# the OCCT it found.
version = "7.9.3.1.dev5"
description = "nanobind bindings for the OpenCASCADE geometry kernel (drop-in OCP)"
readme = "README.md"
requires-python = ">=3.12"
license = { text = "LGPL-2.1-only" }
authors = [{ name = "N3XD" }]
classifiers = [
"Private :: Do Not Upload",
"Programming Language :: Python :: 3.12",
"Programming Language :: C++",
]
[project.urls]
Repository = "https://git.stroblme.de/N3XD/ocp"
[tool.scikit-build]
minimum-version = "0.11"
build-dir = "build/{wheel_tag}"
wheel.packages = ["python/OCP", "python/n3xd_ocp"]
# Stubs are generated during the build and gitignored, and file selection is
# git-based by default — so they must be re-included by hand or the wheel ships
# without the typing surface that retires the backend's ty suppression.
sdist.include = ["python/OCP/*.pyi", "python/n3xd_ocp/*.pyi"]
# abi3 from day one, matching the forge/assay convention: one wheel spans 3.12
# and 3.13, so the interpreter bump (roadmap 10D) needs no rebuild here.
wheel.py-api = "cp312"
# No sdist is ever published — it could not build without the OCCT builder
# image, and offering one would invite the production host (4 cores) to try.
# That is enforced by only ever building wheels (scripts/build_wheel.sh) rather
# than by an exclude rule here: scikit-build-core feeds sdist.exclude into the
# wheel's package-file mapping too, so excluding "*" silently ships a wheel
# with the compiled extension and none of the Python package.
[tool.pytest.ini_options]
testpaths = ["tests"]