Thermo

Added
- Unifac 2.0: Hayer, Nicolas, Thorsten Wendel, Stephan Mandt, Hans Hasse, and Fabian Jirasek. “Advancing Thermodynamic Group-Contribution Methods by Machine Learning: UNIFAC 2.0.” Chemical Engineering Journal 504 (January 15, 2025): 158667. https://doi.org/10.1016/j.cej.2024.158667.

- First pass implementation of Bondi group contribution method for estimating R and Q for UNIQUAC (and regressing new UNIFAC groups). Some definitions are unclear and/or hard to compute programatically.
- More UNIFAC groups have SMARTS groups patterns assigned to them
- Add functions for identifying which functional groups are in a chemical
- Molecule graph structure functions `count_rings_by_atom_counts`, `identify_functional_group_atoms`, `identify_conjugated_bonds`
- New Diky Joback implementation which improves predictions of ideal gas heat capacity, from: Elliott, J. Richard, Vladimir Diky, Thomas A. Knotts IV, and W. Vincent Wilding. The Properties of Gases and Liquids, Sixth Edition. 6th edition. New York: McGraw Hill, 2023.
- Activity coefficient models have `missing_interaction_parameters` function added to return a list of compound interactions that are missing

Changed
- NISTUFSG CH=NOH previously had the group ID 309, but this duplicated another 309 group CH2(O)2, so it was reassigned 1309 in thermo https://github.com/CalebBell/thermo/issues/158
- Reorganized the teperature dependent chemical property data additions to TDependentProperty objects so that data is added through generic methods, instead of hardcoding the method and attributes on the classes that the attributes are stored on. This positions thermo for adding new data sets easier in the future.
- Renamed arguments `cross_B_model` to `B_mixing_rule` of `VirialGas` to avoid confusion with similar parameter in `VirialCSP`. and the same for `cross_C_model` to `C_mixing_rule`
- Fix missing negative sign on `H_dep` of Virial model - this had impact on enthalpy derivatives, `G_dep`, and fugacity related terms as well. This was a serious bug and any `VirialCSP` calculations should be repeated because of it.
- Clean up UNIFAC assignment tests
- Fix inconsistency of using IAPWS name for multiple TDependentProperty methods

Changed
- Fluids version dependency now >= 1.0.27
- Chemicals version dependency now >= 1.3.0
- General code cleanup and further documentation
- Add Flory Huggins and Hansen activity coefficient models
- Further progress on removing legacy scipy interp1d method
- Clean up in code of vapor pressure extrapolation

Changed
- Compatibility with NumPy 2.0 and SciPy 1.14. Note this causes somewhat different results when extrapolating tabular data (Temperature and pressure dependent - temperature dependent only behaves the same)
- Fluids version dependency now >= 1.0.26
- Chemicals version dependency now >= 1.2.0
- General code cleanup and further documentation

Changed
- Previously added accurate fits to pure-component temperature-dependent properties now have analytical integrals implemented so as to speed up enthalpy and entropy calculations.
- Creation of Flasher objects has been sped up
- Add some fits for pure metal solid and liquid heat capacities to the SGTE UNARY database. The fits are quite accurate but do not implement the same equations.
- Add base class for ThermalConductivitySolid
- Add element fits for thermal conductivities of solids from the source: Ho, C. Y., R. W. Powell, and P. E. Liley. "Thermal Conductivity of the Elements." Journal of Physical and Chemical Reference Data 1, no. 2 (April 1, 1972): 279-421. https://doi.org/10.1063/1.3253100.
- Add additional data for sublimation pressure
- Fix an issue with threading and sqlite lookups

Changed
- Code cleanup with ruff (experiment)
- Add accurate fits to pure-component temperature-dependent properties derived using REFPROP. This is the preferred method where available. As part of this effort, a way of adding new data to thermo using json files is being experimented with.
- Add liquid density and viscosity correlation for 8 elements (experiment)