test: add option to run tests with network access

[tiran]

The new pytest option --with-network enables test cases that are decoratored with network marker. By default, tests like test_git_clone_real are disabled, because they spend a non-trivial amount of time to download content from a remote server.

Summary by CodeRabbit

Release Notes

• Tests
  • Added support for selectively running network-dependent tests via a new command-line flag
  • Enhanced git integration testing to properly handle version detection in cloned repositories
  • Improved test infrastructure for conditional test execution based on network availability

[coderabbitai]

📝 Walkthrough

Walkthrough

The changes implement conditional execution of network-dependent tests through a new --with-network command-line flag. A pytest marker system was introduced to label tests requiring network access, with automatic skipping logic in conftest.py. The test workflow was updated to enable this flag, setuptools-scm was added as a test dependency, and the git integration test was modified to use it for version detection.

Changes

Cohort / File(s)	Summary
Workflow Configuration .github/workflows/test.yaml	Updated unit test execution command to include --with-network flag, enabling network-dependent test behavior during CI runs.
Test Dependencies & Pytest Configuration pyproject.toml	Added setuptools-scm to test optional dependencies and introduced [tool.pytest.ini_options] section with network marker definition for labeling network-access tests.
Test Framework Hooks tests/conftest.py	Implemented pytest_addoption hook to register --with-network CLI flag and pytest_collection_modifyitems hook to conditionally skip tests marked with network keyword when flag is absent.
Integration Test Update tests/test_gitutils.py	Applied @pytest.mark.network decorator to test_git_clone_real and introduced setuptools_scm_version() helper function to extract and verify package version from cloned repository.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~12 minutes

Poem

🐰 With --with-network flag so bright,
Network tests skip or run just right,
Markers guide the tests along,
Conftest hooks keep flow so strong! 🌟

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The PR title 'test: add option to run tests with network access' directly and clearly summarizes the main change: introducing a --with-network option for pytest to conditionally run network-dependent tests.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

🧹 Nitpick comments (1)

tests/test_gitutils.py (1)

17-26: Consider handling edge cases in output parsing.

The helper assumes the last line contains the version, but if setuptools_scm produces unexpected output (e.g., warnings only, or empty output), splitlines()[-1] could fail or return an invalid version string.

For a test helper this is acceptable since failures will surface as test errors, but you might consider adding a brief comment explaining the expected output format.

Note: The Ruff S603 warning about subprocess with untrusted input is a false positive here—root_dir comes from the controlled tmp_path fixture and the command is hardcoded.

💡 Optional: Add defensive handling

 def setuptools_scm_version(root_dir: pathlib.Path) -> Version:
+    """Extract version from a git repository using setuptools_scm.
+    
+    Expects the last line of output to contain the version string.
+    """
     out = subprocess.check_output(
         [sys.executable, "-m", "setuptools_scm"],
         text=True,
         stderr=subprocess.STDOUT,
         cwd=str(root_dir),
     )
     # last line contains the version
-    lastline = out.strip().splitlines()[-1]
+    lines = out.strip().splitlines()
+    if not lines:
+        raise ValueError(f"setuptools_scm produced no output in {root_dir}")
+    lastline = lines[-1]
     return Version(lastline)

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/test_gitutils.py` around lines 17 - 26, The helper
setuptools_scm_version assumes the last line of subprocess.check_output contains
a valid version and can crash or produce a confusing error on empty/unexpected
output; update the function setuptools_scm_version to validate the subprocess
output before using splitlines(): trim and split into lines, if no lines are
present raise a clear RuntimeError (including the full raw output) and then
parse the last line, and wrap Version(lastline) in a try/except to raise a
descriptive error if parsing fails; also add a short comment noting that
root_dir is controlled by tests so the Ruff S603 warning is a false positive.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@tests/test_gitutils.py`:
- Around line 17-26: The helper setuptools_scm_version assumes the last line of
subprocess.check_output contains a valid version and can crash or produce a
confusing error on empty/unexpected output; update the function
setuptools_scm_version to validate the subprocess output before using
splitlines(): trim and split into lines, if no lines are present raise a clear
RuntimeError (including the full raw output) and then parse the last line, and
wrap Version(lastline) in a try/except to raise a descriptive error if parsing
fails; also add a short comment noting that root_dir is controlled by tests so
the Ruff S603 warning is a false positive.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: cac7fee4-a6ce-4986-a715-3703852dd04c

📥 Commits

Reviewing files that changed from the base of the PR and between a939d01 and 4e01776.

📒 Files selected for processing (4)

• .github/workflows/test.yaml
• pyproject.toml
• tests/conftest.py
• tests/test_gitutils.py

[LalatenduMohanty]

@tiran Can you please explain in the commit about why need this now i.e. what triggered you to do this change?

[LalatenduMohanty]

This PR would help test #962, but I am not sure if that is your primary motivation.

[tiran]

The test improvements are related to #938. I plan to use git_clone_fast in the new resolver and downloader API. I had written a test case for the function, but the test case is currently skipped.