| # EA2: Coverage Gap — Add Tests to Reach 90% Branch Coverage |
|
|
| ## Goal |
|
|
| A Python data validation library has insufficient test coverage. Add tests to reach at least |
| 90% branch coverage. The existing tests cover only the happy paths. |
|
|
| ## Coverage Measurement |
|
|
| ``` |
| pytest --cov=validator --cov-branch --cov-report=term-missing |
| ``` |
|
|
| Target: **≥ 90% branch coverage** across all modules in `validator/`. |
|
|
| ## Library Structure |
|
|
| - `validator/core.py` — core validation functions: a domain-specific string validator |
| (email/username/phone depending on seed), `validate_range`, `validate_string` |
| - `validator/rules.py` — rule classes: `RequiredRule`, `TypeRule`, `PatternRule`, `RangeRule`, |
| `ValidationError` |
| - `validator/pipeline.py` — `Pipeline` class for chaining validation rules |
|
|
| ## What the Existing Tests Cover |
|
|
| `tests/test_basic.py` has 7 tests covering only success paths — valid inputs that pass |
| validation. Coverage is approximately 40%. All the error/edge-case branches are uncovered. |
|
|
| ## Guidance on Where Gaps Typically Occur |
|
|
| Validation code commonly has uncovered branches in: |
|
|
| - **Rejection paths**: what happens when the input is invalid, malformed, empty, or None |
| - **Boundary conditions**: behavior when values exactly equal the minimum or maximum thresholds |
| - **Configuration errors**: what happens when the validator itself is misconfigured |
| (e.g., min > max in a range) |
| - **Pipeline behavior**: what happens when there are no rules, when the first rule fails |
| (early exit), when a duplicate rule is added, or when `clear()` is called on an empty pipeline |
| - **Type coercion**: behavior when values are of unexpected types |
|
|
| Use `pytest --cov=validator --cov-branch --cov-report=term-missing` to see exactly which |
| lines and branches are uncovered. |
|
|
| ## Deliverables |
|
|
| - `tests/test_validator.py` with comprehensive tests targeting the uncovered branches |
| - Branch coverage ≥ 90% verified by pytest |
| - All existing tests still pass |
|
|
