CatOpt-Query: MVP scaffolding for a category-theoretic distributed query planning framework.

Overview
- A minimal, production-oriented Python MVP that models local shard plans (Objects), inter-shard signals (Morphisms), and vendor adapters (Functors).
- Includes a canonical representation, two starter adapters (PostgreSQL and MongoDB), and a simple coordinator to fuse local plans.
- Designed as a stepping stone toward the 8–12 week MVP plan described in the project proposal.

What you’ll find here
- Core data models for LocalProblem, SharedVariables, DualVariables, PlanDelta, DataContract, and AuditLog.
- A CanonicalPlan representation and a naive joint-planning coordinator.
- Adapters for PostgreSQL and MongoDB that map local plans to the canonical representation.
- Tests validating protocol serialization, adapter mappings, and basic joint planning behavior.
- A lightweight protocol registry (Graph-of-Contracts concept) skeleton and a small DSL skeleton (data classes-only).

Getting started
- Install: python3 -m build && pip install dist/catopt_query-0.1.0-py3-none-any.whl
- Run tests: pytest -q

Notes
- This is an MVP scaffold. It focuses on correctness, testability, and incremental extensibility for the larger CatOpt-Query project.
- No external DB calls are required for the MVP tests; adapters simulate plan mapping.

License: MIT