test: establish baseline coverage

[sigilmakes]

What

Adds portable baseline coverage before the Core pipeline refactor.

Why

The staged Core pipeline refactor needs a reliable behavioural safety net first. The current tests still depended on private local_data/ pseudopotential files and did not cover several public/error paths. This PR freezes current behaviour with portable synthetic fixtures and tightens the physics assumptions around k-spacing and heavy-element detection.

Closes #10.

How to test

uv run pytest -q
uv run ruff check src tests
uv run ruff format --check src tests
uv run pre-commit run --all-files

Current local result:

37 passed
pre-commit passed: ruff check, ruff format, pytest

Changes

• Replace local_data/ UPF parser tests with synthetic portable UPF fixtures.
• Add CLI main() coverage without shelling out or loading a real model.
• Add structure file loading coverage using a temporary CIF file.
• Add structure analysis tests for transition metals, period-5 heavy elements, and lanthanides.
• Add ML inference error-path tests.
• Add k-mesh baseline tests that assert VASP-style KSPACING / solid-state reciprocal-lattice convention.
• Clarify k-mesh docstrings around the 2π reciprocal-lattice convention.
• Change heavy-element heuristic from Z >= 57 to period 5 and heavier for SOC consideration.
• Update README and architecture docs to describe portable test expectations, VASP-style k-spacing convention, and near-term pipeline direction.

Notes

This is a stacked draft PR based on docs/agents-md while #7 is still open. After #7 merges, this branch should be rebased onto main and the PR base updated to main before review/merge.

---

Written by an agent on behalf of Willow.