- Replace standard logging with loguru across codebase (50+ files)
- Fix logger.exception usage in exception handlers (100+ fixes)
- Update environment variable access to use SettingsManager
- Improve pre-commit hooks to handle legitimate use cases
- Fix formatting and linting issues
Reduced custom code check violations from 268 to 38 (86% reduction)
Reduced env var check violations from 73 to 57 (22% reduction)
* Install Ruff and fix all the Ruff errors.
* Fix pre-commit failures.
* Potential fix for code scanning alert no. 104: Information exposure through an exception
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
* Fix pre-commit failures.
---------
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
- Fix health check endpoint mismatch (use /api/v1/health instead of /health)
- Increase server startup timeout from 10s to 30s with proper health polling
- Add comprehensive server startup logging to diagnose failures
- Fix UI tests by using xvfb-run for headless Chrome in CI
- Improve error reporting when server fails to start
These changes address all the test failures seen in PR #393 by ensuring
the server starts properly and tests check the correct endpoints.
fix: additional CI pipeline fixes for server startup
- Create data directory before starting server in CI
- Add explicit screen configuration for xvfb-run
- Set DISPLAY environment variable for UI tests
The server was failing to start because setup_data_dir() couldn't
create the data directory with the relative path calculation when
installed via PDM.
fix: resolve remaining CI test issues
- Create screenshots directory for UI tests to prevent ENOENT errors
- Fix invalid CSS selector ':contains()' in settings page test
- Add better error handling for screenshot captures
- Improve API test server startup debugging with:
- Environment variable logging
- Immediate process death detection
- Port listening status checks
- Full server log output on failures
These changes should help diagnose why the server isn't starting
in the API tests and fix the remaining UI test failures.
fix: add diagnostics for API server startup issue
- Check if PDM scripts are available with pdm run --list
- Test direct Python module import to isolate the issue
- Try running server with python -m to see actual error
- Add server import test before attempting startup
This will help identify if the issue is with:
- PDM script registration
- Module import paths
- Missing dependencies
- Database initialization
fix: replace hardcoded research ID test with useful error handling test
- Remove test that tries to load non-existent research ID 67
- Replace with test that verifies proper 404 error handling
- Also test history page structure as part of research functionality
- Update test description to reflect new purpose
This makes the test suite more robust by testing actual functionality
rather than relying on specific data that won't exist in CI.
feat: add CI fallback LLM support for tests
- Add USE_FALLBACK_LLM environment variable check in llm_config.py
- Update GitHub Actions workflows to use fallback LLM in CI
- Create test_ci_config.py for CI-specific configurations
- Tests now automatically use FakeListChatModel when USE_FALLBACK_LLM=true
This allows tests to run in CI without requiring actual LLM services,
while still using real LLMs when available locally.
fix: rename test_browsecomp.py to avoid pytest collection issues
- Rename examples/test_browsecomp.py to examples/example_browsecomp.py
- File was causing pytest to fail with SystemExit during test collection
- This is an example script, not a test file, so it shouldn't have test_ prefix
fix: add headless UI test support for CI environments
- Update run_all_tests.py to detect headless environments and use xvfb-run
- Set PUPPETEER_HEADLESS=true when xvfb-run is not available
- Add CI=true environment variable in GitHub Actions workflow
- This fixes the 'Missing X server or :1' error in CI
fix: CI test improvements
- Rename test files to avoid pytest collection errors:
- download_stuff_for_local_test.py -> download_stuff_for_local.py
- test_url_utils_debug.py -> debug_url_utils.py
- Add server startup for API unit tests in GitHub Actions
- API tests now properly start server before running tests
fix: improve CI test reliability
- Add better error handling for API test server startup
- Fix IPv6 localhost detection in normalize_url function
- Add server process monitoring and logging on failure
- Ensure server is actually running before starting tests
