test: fix import order in test_cli_gui_run

[olivermeyer]

Fix an import error causing test_cli_gui_run to fail with

AttributeError: 'module' object at aignostics.utils has no attribute 'utils'

E.g. this job.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.
see 9 files with indirect coverage changes

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[atlantis-platform-engineering]

Error: This repo is not allowlisted for Atlantis.

[claude]

Claude finished @olivermeyer's task in 3m 35s —— View job

---

Code Review Complete ✅

I've reviewed PR #419 and the import order fix looks good! Here's my comprehensive analysis:

Review Checklist

• Read and understand the code changes
• Verify test markers are present
• Check commit message follows conventional commits
• Validate code quality standards
• Check for architectural compliance
• Verify no breaking changes
• Review test coverage impact
• Provide final review summary

---

Summary

Status: ✅ APPROVED - Ready to Merge

This PR successfully fixes an AttributeError in test_cli_gui_run caused by incorrect monkeypatch usage with string paths. The fix is clean, minimal, and follows best practices.

What Changed

File: tests/aignostics/cli_test.py (+5/-1)

Problem:

# Before: Using string path for monkeypatchmonkeypatch.setattr("aignostics.utils._gui.gui_register_pages", mock_gui_register_pages)

This failed in CI with:

AttributeError: 'module' object at aignostics.utils has no attribute 'utils'

Solution:

# Import module explicitly before monkeypatchingimportaignostics.utils._guias_utils_gui# Use the imported module referencemonkeypatch.setattr(_utils_gui, "gui_register_pages", mock_gui_register_pages)

Root Cause: In pytest-xdist parallel execution, string-based monkeypatch paths can fail if the module hasn't been imported yet in that worker process. Importing explicitly ensures the module is loaded before patching.

✅ Verification Results

1. Test Execution

• ✅ test_cli_gui_run now passes
• ✅ All 8 tests in cli_test.py pass (100% success rate)
• ✅ No flaky behavior observed

2. Code Quality (All Checks Pass)

✅ ruff check - No issues
✅ ruff format - 156 files already formatted
✅ pyright - 0 errors, 0 warnings
✅ mypy - Success: no issues in 80 source files

3. Test Markers ✅ (CRITICAL CHECK)

All tests properly marked with @pytest.mark.integration:

• ✅ No unmarked tests found (0 collected with marker filter)
• ✅ All tests will run in CI pipeline

4. Commit Message ✅

test: fix import order in test_cli_gui_run

• ✅ Follows conventional commits format
• ✅ Type: test (appropriate for test fixes)
• ✅ Clear, concise description

5. Architecture & Standards ✅

• ✅ No breaking changes
• ✅ No new dependencies
• ✅ No changes to production code (test-only fix)
• ✅ Follows pytest best practices (explicit imports > string paths)
• ✅ Added helpful explanatory comment

Code Review Details

Lines 132-136 (New code):

# Import module explicitly to ensure it's loaded before monkeypatching with string paths.# Without this, monkeypatch.setattr("aignostics.utils._gui...") fails in CI because# pytest-xdist workers may not have aignostics.utils loaded as an attribute yet.importaignostics.utils._guias_utils_gui

✅ Excellent: Clear comment explaining the reason for this pattern. This helps future maintainers understand why explicit import is needed.

Line 193 (Modified):

monkeypatch.setattr(_utils_gui, "gui_register_pages", mock_gui_register_pages)

✅ Correct: Using the imported module reference instead of string path prevents the AttributeError in pytest-xdist workers.

Why This Fix Works

1. Module Loading: Explicit import ensures aignostics.utils._gui is loaded in the worker's namespace
2. Reference-Based Patching: Using the module object directly (_utils_gui) is more reliable than string paths in parallel execution
3. No Side Effects: The import is scoped to the conditional block (only if nicegui is available)
4. Consistent Pattern: Matches pytest best practices for monkeypatching in parallel test environments

Testing Strategy Compliance

From .github/CLAUDE.md:

• ✅ Test has correct @pytest.mark.integration marker
• ✅ Uses appropriate parallelization (XDIST_WORKER_FACTOR=0.2 for integration tests)
• ✅ Timeout appropriate (default 10s, sufficient for this test)
• ✅ Mocks external dependencies (nicegui.ui.run)
• ✅ Does not call real external services

Impact Assessment

Risk Level: ⚠️MINIMAL

• Single test file modified
• No production code changes
• No API changes
• No dependency changes
• Fixes existing test failure

Coverage Impact: ✅ NEUTRAL (no change to coverage metrics)

CI/CD Impact: ✅ POSITIVE (resolves test failures in parallel execution)

Additional Observations

Good Practices Demonstrated:

1. ✅ Minimal change principle - only fixed what was broken
2. ✅ Added explanatory comment for future maintainers
3. ✅ Follows pytest-xdist best practices
4. ✅ Test remains comprehensive (verifies all expected behaviors)

No Issues Found:

• No linting violations
• No type checking errors
• No missing test markers
• No architectural violations
• No security concerns

Recommendation

✅ APPROVE AND MERGE

This is a textbook example of a clean, focused bug fix:

• Identifies the root cause (pytest-xdist module loading)
• Applies minimal fix (explicit import)
• Documents the reasoning (helpful comment)
• Maintains all existing test coverage
• Passes all CI checks

No further changes needed.

---

Branch: test/fix-import-order

[olivermeyer]

One failure on MacOS runner for a different reason:

FAILED tests/aignostics/bucket/cli_test.py::test_cli_bucket_flow - Failed: Timeout (>900.0s) from pytest-timeout.

I suggest we still merge the fix because (1) test_cli_gui_run succeeded on the same runner and (2) test_cli_bucket_flow is flaky.