Standardize model chemistry

[cgrambow]

Motivation or Problem

Model chemistry definitions in Arkane are not unique because changes in software and options can lead to different results. The goal is to standardize the representation in Arkane and in the RMG database by defining detailed and unique levels of theory to identify the settings used in quantum chemistry calculations.

Description of Changes

Implements two new classes, LevelOfTheory and CompositeLevelOfTheory, which are used by Arkane internally instead of model chemistry strings. Arkane input files can still be specified using the modelChemistry syntax, but using the new level of theory objects is now also possible. Arkane will first try to find an exact match for the specified level of theory in the RMG database (including options like software, solvation, and additional arguments) but will fall back to just matching method and basis if an exact match is not possible (and will print a warning if this is the case).

The corresponding database PR ReactionMechanismGenerator/RMG-database#409 updates the database to also use the new level of theory objects. This required some changes to RMGObject to enable using hashable objects as dictionary/YAML keys.

Testing and Reviewer Suggestions

Unit tests were adapted and should pass. Make some Arkane input files using the new syntax and check that it works as expected.

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (master@74cee80). Click here to learn what that means.
The diff coverage is n/a.

[alongd]

Thanks for pulling this off so quickly! I understand it's WIP. I Added two minor comments

[lgtm-com]

This pull request introduces 2 alerts when merging 7fe83c3 into 99d009a - view on LGTM.com

new alerts:

• 2 for Wrong name for an argument in a class instantiation

[alongd]

A possible additional attribute to add to the level of theory is solvation.
This is what I have from TCKDB:

solvation (Dict[str, str]): The solvation method used. Keys (values) are:
    - 'method' (str): The solvation method used, e.g., 'SMD'.
    - 'solvent' (str): The considered solvent, e.g., 'water'.
    - 'description' (str): An optional description of the solvation method used if not standard.

[cgrambow]

Thanks for the initial comments @alongd. This is now ready for review.

[amarkpayne]

The code itself looks really good. I have only a few comments code wise. I might have a few more comments about the big picture--we can discuss offline at our meeting today after I have thought about it some more.

We should also figure out how we want to handle the cases where the software is required (you have the check for making sure that the user specifies it, but how should this affect warning messages/behavior when the software does not match?)