Enhance testing configuration and coverage measurement

[Karrenbelt]

• Patched PYTHONPATH in test_filesystem fixtures to improve test isolation
• Added pytest-xdist support in pyproject.toml for parallel test execution
• Added hypothesis.settings for scaffolded protocols
• Updated pyproject.toml to include additional tool.coverage settings for isolated filesystem paths
• Implemented get_test_cpu_count in auto_dev/tests based on os.environ.get('CI')
• Disabled test_repo.py temporarily for stability. --> Re-enabled in Streamline repo scaffolding & re-write & re-enable repo scaffolding tests #639

Note: PR #639 re-enables test_repo.py

[Karrenbelt]

on local this ensures not all cores will be used (3/4th I believe), while on the CI we take all that are available

[Karrenbelt]

this increased test coverage locally from ~34% to ~53%

We have some issue on CI generating the coverage report tho, it appears (see CI run)

1s
Run MishaKav/pytest-coverage-comment@main
File read successfully "/Users/runner/work/auto_dev/auto_dev/./coverage-report.txt"
Error: Coverage file "/Users/runner/work/auto_dev/auto_dev/./coverage-report.txt" has bad format or wrong data
./coverage-report.txt
Nothing to report

[Karrenbelt]

When running with pytest-xdist, certain harmless issues incorrectly surface. To avoid modifying the scaffolded codebase in a way that affects production code, we inject a workaround into the generated tests/__init__.py during integration tests to ensure tests pass without impacting the production scaffold

[Karrenbelt]

no longer needed now that monkeypatch.setenv is used in the fixtures

[Karrenbelt]

These are the ghosts of a flawed pattern I once endorsed. I’ve since repented; see #639

[8ball030]

Why skip this entire set of tests?

[Karrenbelt]

This was temporary, since these tests were very poorly written and caused interference with other tests. They have been re-enabled in PR #639, which was already merged into this PR, and this is their current state in this PR:

auto_dev/tests/test_repo.py

Lines 1 to 55 in 876757d

	"""Tests for the click cli."""
	import subprocess
	from pathlib import Path
	def _test_repo_scaffold(repo_type, make_commands, cli_runner, test_clean_filesystem):
	repo_root = Path(test_clean_filesystem) / "dummy"
	command = ["adev", "repo", "scaffold", repo_root.name, "-t", repo_type]
	runner = cli_runner(command)
	result = runner.execute()
	makefile = repo_root / "Makefile"
	# Verify the basic repository structure exists
	assert result, runner.output
	assert repo_root.exists(), f"Repository {repo_root} does not exist."
	assert (repo_root / ".git").exists(), f".git directory not found in {repo_root}."
	assert makefile.exists(), f"Makefile not found in {repo_root}."
	# Run each make command and collect any errors.
	error_messages = {}
	for cmd in make_commands:
	proc_result = subprocess.run(
	["make", cmd],
	shell=False,
	capture_output=True,
	text=True,
	check=False,
	cwd=repo_root,
	)
	if proc_result.returncode != 0:
	error_messages[cmd] = proc_result.stderr
	assert not error_messages, f"Errors encountered in make commands: {error_messages}"
	def test_python_repo(cli_runner, test_clean_filesystem):
	"""Test scaffolding a Python repository."""
	_test_repo_scaffold(
	repo_type="python",
	make_commands=("fmt", "lint", "test"),
	cli_runner=cli_runner,
	test_clean_filesystem=test_clean_filesystem,
	)
	def test_autonomy_repo(cli_runner, test_clean_filesystem):
	"""Test scaffolding an Autonomy repository."""
	_test_repo_scaffold(
	repo_type="autonomy",
	make_commands=("fmt", "lint", "test", "hashes"),
	cli_runner=cli_runner,
	test_clean_filesystem=test_clean_filesystem,
	)