Expand test coverage across modules (unit + integration)

[bernalde]

Goal

Grow real test coverage beyond the two existing integration scenarios, with an emphasis on fast unit tests that don't need assimulo. Part of #3; depends on #4.

Current state

• Only integration tests exist (full reactor / flowsheet solves). Reverted PR Fix crystallizer energy balance units and jacket paths #106 estimated baseline coverage ~3%.
• Many pure-Python pieces are unit-testable in isolation without a solver: kinetics rate evaluation, thermodynamic property calls, stream/phase construction, name analysis, interpolation, FVM flux/limiter helpers, moment conversions.

Proposed coverage targets (unit, no assimulo)

• Kinetics.py — RxnKinetics.get_rxn_rates / temp_term / elem_f_model against hand-computed values; CrystKinetics.get_kinetics supersaturation branches.
• Phases.py / Streams.py — construct Liquid/Solid/Vapor streams, check derived properties and moment/distribution conversions.
• ThermoModule.py — property evaluations on a known compound DB.
• Commons.py — FVM helpers (high_resolution_fvm, limiter math), NameAnalysis mapping.
• Connections.py — topological_bfs ordering and recycle detection; unit conversion.

Integration coverage (assimulo, mark @pytest.mark.assimulo)

• Batch / CSTR / Semibatch reactor smoke tests against reference trajectories.
• Batch and MSMPR crystallizer (moments + 1D-FVM) smoke tests.
• At least one separation unit (filtration/drying) end-to-end.

Acceptance criteria

• Coverage measured in CI and trending up from baseline; a target (e.g. 30% near-term) agreed in the PR.
• Unit tests run in the assimulo-free core job (Add a minimal, assimulo-free GitHub Actions CI workflow #5).
• Tests use fixtures/reference data, not network access.

---

Upstream PR

Wave 2 — land upstream as several small, test-only PRs (one module or theme each), not a single large one. See #3.

---

Reverted PR provenance

This issue is part of the selective re-introduction of useful infrastructure from the reverted upstream work:

• Implement CI/CD pipeline and enhance testing framework for PharmaPy #1 and Refactor codebase and enhance CI for quality checks #2: original fork-side CI/testing work.
• Implement CI/CD pipeline and enhance testing framework for PharmaPy CryPTSys/PharmaPy#106: upstream PR that bundled CI, packaging, tests, docs, and broad formatting.
• Revert "Implement CI/CD pipeline and enhance testing framework for PharmaPy" CryPTSys/PharmaPy#107: upstream revert of that bundled PR.

Implementation should stay small and reviewable, avoid repo-wide formatting, and preserve traceability through one focused PR per issue.

[Mazhar331]

Coordination note from PR #119 / issue #81:

The drying/separation end-to-end coverage item here should wait until the open Drying correctness issues are solved and merged. A full Drying.solve_unit transient would be valuable, but asserting realistic physics before the drying model is corrected would make the test brittle or misleading.

Current blockers include #21 / PR #108, #24, #28, #37, #42, #81 / PR #119, and #101. Once those land, this testing umbrella (#7) is the right place to track a class/example-backed Drying integration regression, or to link a dedicated follow-up issue for it.