Manual/semi-automatic performance regression checking #16

	name: Profiling without benchmarking

	# This runs when a PR is opened, synchronised or reopened
	on:
	pull_request:
	types: [opened, synchronize, reopened]
	workflow_call:

	jobs:
	test:
	uses: ./.github/workflows/pyrealm_ci.yaml
	secrets: inherit

	profiling:
	needs: test
	runs-on: ubuntu-latest
	strategy:
	fail-fast: false

	steps:
	- uses: actions/checkout@v4
	with:
	ref: ${{ github.head_ref }}

	- uses: actions/setup-python@v5
	with:
	python-version: "3.10"

	- name: Install Poetry
	uses: abatilo/[email protected]
	with:
	poetry-version: 1.8.2

	- name: Install dependencies and add env vars
	run: \|
	poetry install
	sudo apt-get install graphviz
	calculatedSha=$(git rev-parse --short ${{ github.sha }})
	echo "COMMIT_SHORT_SHA=$calculatedSha" >> $GITHUB_ENV

	- name: Run the profiling test suite, using time to get peak memory usage
	id: profiling
	run: poetry run /usr/bin/time -v pytest -m "profiling" --profile-svg

Provide feedback