Validation: src/proteus/outgas/burial_ledger.py¶
This page tracks the @pytest.mark.reference_pinned tests that anchor
proteus.outgas.burial_ledger, which records where in melt fraction trapping
buried each volatile, so that a remelt returns the most recently buried mass.
| Test id | Reference | Scope | Last compared |
|---|---|---|---|
tests/outgas/test_burial_ledger.py::test_the_ledger_matches_a_fine_grid_over_a_random_history |
Cross-implementation check against an independent ledger on a grid of 2e5 melt-fraction bins, which deposits each freeze uniformly over its bins and empties every bin below the melt fraction after each remelt | A seeded history of 400 steps that mostly freeze and sometimes remelt. The mass held below four melt fractions agrees to rel=1e-3, the discretisation error of one grid bin, and the total held to 1e-9. |
2026-10-01 |
Invariants checked alongside¶
- A remelt returns the most recently buried mass first: after freezing from
Phi = 1to 0.7 and on to 0.5, remelting to 0.6 returns half of the second layer, not a fifth of everything. - Returned plus still held equals buried, after every remelt.
- Freezing and remelting the same interval cancel. In
tests/outgas/test_trapping.pya closed melt cycled a hundred times betweenPhi = 0.41and 0.40 keeps its trapped water to1e-12, where releasing the mean of the solid raises it by 4.4% a cycle.