⚙️ Test Setup & Guidelines

Automated testing ensures the long-term reliability and maintainability of the metadata generation platform.
This chapter describes the structure, conventions, and execution of all test suites — from database-integrated lineage validation to pure logic modules such as naming, hashing, and validators.

The goal is to keep the test environment pragmatic yet powerful:
- Realistic database tests verify dataset and lineage creation using Django’s ORM.
- Pure logic tests ensure correctness of naming, hashing, and SQL generation — without any database dependency.
- Template tests define the expected behavior of the upcoming SQL Preview pipeline.
- SQL preview tests also validate error surfacing and policy violations

Together, these tests form the foundation for confident releases and safe refactoring across the platform.

🔧 1. Overview

This project uses pytest with pytest-django for both database-integrated and pure-logic testing.
All tests are located in core/tests/ and organized by logical area.

Layer	Scope	Example file
Database / Metadata	Validates TargetDataset, TargetDatasetInput, and lineage integrity.	test_target_lineage.py
Generation (Logic)	Tests for hashing, naming, and validators.	test_generation_hashing_and_naming.py, test_generation_validators.py
SQL Preview (Templates)	These template tests show expected behavior for preview rendering.	test_sql_preview_*.py
Smoke & Import Tests	Sanity check for Django setup and module imports.	test_smoke.py

---

🔧 2. Running Tests

From the project root:

# Run all tests
python runtests.py

🔧 3. Database Tests

Tests that use Django models automatically create a temporary test database (e.g. test_elevata) and destroy it afterward.
Your production data is never touched.

Fixtures that use the database are defined in core/tests/conftest.py:

• source_system_sap
• source_dataset_sap_customer
• target_schemas
• raw_stage_rawcore_datasets

These fixtures provide a realistic Raw → Stage → Rawcore dataset chain for lineage validation.

---

🔧 4. Logic-Only Tests

Logic-based modules (e.g. hashing, naming, validators, logical_plan)
are tested purely in memory — no Django setup or migrations required.

They are fast, deterministic, and ideal for CI pipelines.

Examples:

• test_generation_hashing_and_naming.py
• test_generation_validators.py
• test_logical_union.py

---

🔧 5. Skipped SQL Preview Tests

Files under test_sql_preview_* are template tests for the future SQL Preview pipeline. They describe expected output patterns.

To activate them later:

1. Implement your SQL preview renderer (e.g. build_preview_sql() in preview.py).
2. Update the import paths in the test files.
3. Remove the @pytest.mark.skip decorators.

Example command once wired:

python runtests.py

🔧 6. Local and CI Execution

🧩 Local execution

Run all tests using:

python runtests.py

Run a specific file (verbose mode):

python runtests.py core/tests/test_generation_validators.py -v

Run only tests matching a keyword:

python runtests.py -k "hashing"

🔧 7. CI Integration

If you use GitHub Actions or GitLab CI, your step can simply call the same command:

- name: Run tests
  run: |
    pip install -r requirements-dev.txt
    python runtests.py --maxfail=1

Optional flags:
- --reuse-db speeds up local test runs
- --cov enables coverage reports (requires pytest-cov)

---

🔧 8. Conventions

• Indentation: 2 spaces (consistent with project code style)
• Comments: English, concise, and descriptive
• Naming pattern:
• Test files: test_*.py
• Test functions: test_*
• Database fixtures: defined in core/tests/conftest.py
• Logic-only tests: may use lightweight dummy classes (no mocks)

---

🔧 9. Recommended Next Steps

• Add pytest-cov for coverage tracking: bash pip install pytest-cov python runtests.py --cov=metadata

• Integrate pre-commit hooks to ensure consistent formatting.

• Gradually unskip SQL preview tests as the rendering pipeline evolves.

---

© 2025-2026 elevata Labs — Internal Technical Documentation