Makefile

SHELL=/bin/bash -euo pipefail

install: install-node install-python

install-python:
	poetry install

install-node:
	npm install

test:
	@echo "No tests configured."

lint:
	npm run lint
	poetry run flake8 **/*.py
	find -name '*.sh' | grep -v node_modules | xargs shellcheck

clean:
	rm -rf build
	rm -rf dist

publish: clean
	mkdir -p build
	npm run publish 2> /dev/null

serve: update-examples
	npm run serve

check-licenses:
	npm run check-licenses
	scripts/check_python_licenses.sh

deploy-proxy: update-examples
	scripts/deploy_proxy.sh

deploy-spec: update-examples
	scripts/deploy_spec.sh

format:
	poetry run black **/*.py

build-proxy:
	scripts/build_proxy.sh

_dist_include="pytest.ini poetry.lock poetry.toml pyproject.toml Makefile build/. e2e"

release: clean publish build-proxy
	mkdir -p dist
	for f in $(_dist_include); do cp -r $$f dist; done

.PHONY: e2e e2e-mock

pytest := PYTEST_ADDOPTS="--color=yes" poetry run pytest --reruns 3 --reruns-delay 2 $$f --suppress-no-test-exit-code

# Since different fixtures are used in diferent files to create and modify test
# apps and products, is quite difficult to handle the state of the resources
# they create in Apigee and the order those resources need to be created or
# modified by the test. On top of that this tests rely on the Apigee API wich
# can be a bit temperamental some times presenting an async behaviour due to its
# internal caching layers. To make testing more stable we run a fresh pytest
# session for each test file (meaning fresh fixtures) increasing isolation
# between testing modules.
e2e:
	rm -f reports/e2e.xml  > /dev/null || true 
	@for f in  $$(find  e2e/tests  -name "test_*.py") ; do \
		echo $$f; \
		$(pytest) || exit 1; \
	done