improving-python-coverage

Run the test suite with coverage, export a JSON report, and prioritize files with large size and low percent_covered. Then write pytest tests targeting missing lines in conditional branches, error handling, and edge cases, and re-measure until the target gain is reached.

Generate a coverage JSON report with coverage json, which lists per-file missing_lines arrays. Focus on core modules with below-average coverage and skip files above 97 percent, protobuf, vendor, and static directories.

Test conditional logic, error handling, edge cases like None, empty, zero, and max values, and public API functions. Avoid testing simple accessors, protobufs, and implementation details that add no real bug-catching value.

Import integration-only packages inside the test function rather than at module top level, and mark the test with pytest.mark.require_integration. This lets the test skip gracefully when those packages are not installed in the unit test environment.

Version-specific branches for Python or library versions only run on matching CI jobs, and integration tests run in separate CI jobs with extra packages installed. Verify whether uncovered lines are exercised in CI before adding tests or pragma no cover annotations.

Use it sparingly for platform-specific branches, defensive code that should never execute, or abstract method stubs, always with a reason comment. Do not use it to hide testable logic just to inflate coverage numbers.