pycalphad / pycalphad / 2083

87%

push

travis-ci

FIX: Two sublattice ionic liquid fixes and tests (#273)

This PR fixes two bugs in the implementation of ionic liquid two sublattice model

1. Neutral species in the anion sublattice can be specified in the TDB using one sublattice parameters. These are special cased now in `Model` to be valid parameters and to correct their sublattice index in the Redlich-Kister sum to produce the correct `v.SiteFraction` objects.
2. A sorting bug was fixed in the anion lattice, where the model from Sundman [1] describes the anion sublattice being sorted by first charged species, then vacancies, then anion species. Within each species group, the species are sorted alphabetically as usual. This is changed within Model itself, rather than the TDB which sorts the parameters at parse time. Because it's specific to this implementation of the Model, I think it makes sense to put it in the `Model` class corresponding to any phase with the `ionic_liquid_2SL` model hint applied.

This PR includes 
* tests that compare the energy to the energies computed by the same database in Thermo-Calc
* some documentation improvements and cleanup for the ionic liquid

There's also an additional commit that adds support for auxiliary text in the TDB `PHASE` keyword, making it easier to read MatCalc databases that frequently use this feature.

[1] Sundman, "Modification of the two-sublattice model for liquids", Calphad 15(2) (1991) 109-119 https://doi.org/d3jppb

45 of 45 new or added lines in 3 files covered. (100.0%)

3810 of 4409 relevant lines covered (86.41%)

5.18 hits per line