Improve support for MRCM

[ekorchmar]

State

Currently, Jackalope core evaluates only a subset of MRCM rules, so there is little indication that concept may break established modelling; this may lead to Jackalope silently failing to find the correct subhierarchy. Precisely, only Attribute Domains(AD) are tested, but not attribute groupping, nor Attribute Ranges(AR).

Proposed solution steps

• Add new rule constructor class that would produce AR rules validation.data_atoms
• Implement AR validation methods for MRCMValidator. Initial subset of rules for testing may be defined by hand, because they use ECL definition language, which Jackalope does not yet speak.
• Similarly, implement groupping validation for AD. This does not require new parsing technologies.
• Add support for ECL grammar. I suggest switching to tree-sitter as a more long-term sustainable solution, primarily for it's ability for error recovery. Existing expression parser should also move off ANTLR for that reason, but that is a task for another day.
• Create a rule factory that creates AR rules based on ECL definitions.
• Extend MRCMValidator to use these AR rules
• Make sure MRCMValidator raises informative exceptions when AD/AR rules are broken. Process them in server.py

Domain Reference has lower priority as it is mostly redundant to AD+AR constraints, and is more useful to authoring tools than to Jackalope core. Module Scope is also way beyond what is required here.

[ekorchmar]

Started working on it here:
https://github.com/ekorchmar/jackalope/tree/mrcm