Repository navigation

fix(notebooks): give the header logo a height, so Colab does not distort it #1266

Workflow file for this run

	# Run tests, coverage, and build docs across all supported platforms
	# Triggers on pushes/PRs to main.

	name: CI

	on:
	push:
	branches: [main]
	pull_request:
	branches: [main]

	env:
	FORCE_JAVASCRIPT_ACTIONS_TO_NODE24: true
	MPLBACKEND: Agg
	QT_QPA_PLATFORM: offscreen

	jobs:
	test-core:
	runs-on: ubuntu-latest
	steps:
	- uses: actions/checkout@v7

	- uses: actions/setup-python@v7
	with:
	python-version: "3.12"
	cache: pip

	- name: Install package (core, no swift)
	run: pip install .[dev]

	# Model tests (YuMi, Valkyrie, ...) load URDFs via the robot_descriptions
	# package, which does NOT bundle the actual asset data -- on first use of
	# a given model it lazily `git clone`s a large upstream asset repo
	# straight into this cache dir. That clone is real network I/O and can
	# be slow or occasionally flaky (timeout) on a cold runner. Caching the
	# directory means only the first run after this lands pays that cost.
	# Keyed by python-version too (not just OS) -- see the "Removed:
	# robot_descriptions CI caching" tech-debt entry (git history,
	# tech-debt.md as of commit f46b5c83^) for why a previous attempt at
	# this (PR #530) didn't actually help: an OS-only key can't distinguish
	# between the os x python-version matrix's parallel jobs. Bump "v2" to
	# force a clean re-fetch if the cache ever ends up corrupted/stale.
	- name: Cache robot_descriptions assets
	uses: actions/cache@v6
	with:
	path: ~/.cache/robot_descriptions
	key: robot-descriptions-v2-${{ runner.os }}-3.12

	- name: Verify compiled extensions built
	# Catches a silent build/import regression for _fknm_c/_frne_c --
	# without this, tests/test_fknm_fallback.py's C-vs-Python
	# cross-validation tests would just skip (they're gated on
	# _C_AVAILABLE precisely so a missing extension can't masquerade as
	# a passing comparison), and nothing else would notice.
	shell: bash
	run: \|
	python -c '
	from roboticstoolbox.ets.fknm import _C_AVAILABLE as fknm_c
	from roboticstoolbox.robot.frne import _C_AVAILABLE as frne_c
	assert fknm_c and frne_c, f"compiled extension missing: fknm={fknm_c} frne={frne_c}"
	print("OK: both _fknm_c and _frne_c built")
	'

	- name: Test (core)
	run: pytest tests/ --ignore=tests/test_blocks.py -m "not slow" --timeout=50 --timeout_method=thread -q

	test:
	runs-on: ${{ matrix.os }}
	strategy:
	fail-fast: false
	matrix:
	os: [ubuntu-latest, windows-latest, macos-latest]
	python-version: ["3.10", "3.11", "3.12", "3.13", "3.14"]
	steps:
	- uses: actions/checkout@v7

	- uses: actions/setup-python@v7
	with:
	python-version: ${{ matrix.python-version }}
	cache: pip

	- name: Shorten TEMP path (Windows — avoids MAX_PATH breach during C++ compile)
	if: runner.os == 'Windows'
	run: \|
	New-Item -ItemType Directory -Force -Path C:\T \| Out-Null
	"TEMP=C:\T" >> $env:GITHUB_ENV
	"TMP=C:\T" >> $env:GITHUB_ENV

	- name: Install package
	run: pip install .[dev,swift]

	# See the "Cache robot_descriptions assets" step in test-core above --
	# same reasoning, this job hits the same lazily-cloned model data.
	- name: Cache robot_descriptions assets
	uses: actions/cache@v6
	with:
	path: ~/.cache/robot_descriptions
	key: robot-descriptions-v2-${{ runner.os }}-${{ matrix.python-version }}

	- name: Verify compiled extensions built
	# See the same step in test-core for why this matters.
	shell: bash
	run: \|
	python -c '
	from roboticstoolbox.ets.fknm import _C_AVAILABLE as fknm_c
	from roboticstoolbox.robot.frne import _C_AVAILABLE as frne_c
	assert fknm_c and frne_c, f"compiled extension missing: fknm={fknm_c} frne={frne_c}"
	print("OK: both _fknm_c and _frne_c built")
	'

	- name: Test
	run: pytest tests/ --ignore=tests/test_blocks.py -m "not slow" --timeout=50 --timeout_method=thread -q

	all-tests-passed:
	name: All tests passed
	needs: [test-core, test]
	runs-on: ubuntu-latest
	if: always()
	steps:
	- name: Check test job results
	run: \|
	if [ "${{ needs.test-core.result }}" != "success" ] \|\| [ "${{ needs.test.result }}" != "success" ]; then
	echo "test-core: ${{ needs.test-core.result }}, test: ${{ needs.test.result }}"
	exit 1
	fi

	# Advisory for now: the package is being annotated module by module, and
	# pyproject.toml lists the modules that still have errors. Not in
	# `all-tests-passed`, so it cannot block a merge. Make it required once that
	# list is empty.
	mypy:
	runs-on: ubuntu-latest
	continue-on-error: true
	steps:
	- uses: actions/checkout@v7

	- uses: actions/setup-python@v7
	with:
	python-version: "3.12"
	cache: pip

	- name: Install package
	run: pip install .[dev,swift] mypy

	- name: Type check
	run: mypy

	coverage:
	needs: [test-core, test]
	runs-on: ubuntu-latest
	steps:
	- uses: actions/checkout@v7

	- uses: actions/setup-python@v7
	with:
	python-version: "3.12"
	cache: pip

	- name: Install package
	run: pip install .[dev,swift]

	# See the "Cache robot_descriptions assets" step in test-core above --
	# same reasoning, this job hits the same lazily-cloned model data.
	- name: Cache robot_descriptions assets
	uses: actions/cache@v6
	with:
	path: ~/.cache/robot_descriptions
	key: robot-descriptions-v2-${{ runner.os }}-3.12

	- name: Run coverage
	run: pytest tests/ --ignore=tests/test_blocks.py --cov=roboticstoolbox --cov-report=xml:coverage.xml -q

	- uses: codecov/codecov-action@303a32d7a59b442fa8d48b6a1cc6825c09c847a5 # v7.1.1
	with:
	files: ./coverage.xml

	# Builds the Sphinx docs + JupyterLite site and, on main, deploys them to
	# GitHub Pages. Steps live in docs.yml (a reusable workflow) so this path
	# and the manual docs-publish.yml dispatch can never drift apart.
	docs:
	needs: [test-core, test]
	uses: ./.github/workflows/docs.yml
	permissions:
	contents: read
	pages: write
	id-token: write

Navigation Menu

Repository navigation

fix(notebooks): give the header logo a height, so Colab does not distort it #1266

Workflow file

fix(notebooks): give the header logo a height, so Colab does not distort it #1266

Uh oh!

Workflow file for this run