# CatOpt Studio MVP

A lightweight, open-source framework sketching category-theory-inspired compositional optimization across edge meshes.

Status: MVP scaffold. Core primitives implemented; packaging scaffolding added. Tests to validate the DSL surface exist.

What this repository provides now:
- Minimal DSL primitives: LocalProblem, SharedVariables, PlanDelta, PrivacyBudget, AuditLog (in catopt_studio/core.py)
- Public API exposing the DSL surface (catopt_studio/__init__.py)
- Packaging metadata (pyproject.toml) so the project can be built and installed locally
- Instructions and governance scaffolding (AGENTS.md) for future expansion

How to run tests locally (once dependencies are installed):
- Install dependencies and build the package
  - python3 -m pip install --upgrade build
  - python3 -m build
- Run tests
  - pytest

Roadmap (high level):
- Expand the DSL with more primitives and type safety
- Add a tiny solver placeholder and an adapter interface
- Introduce a basic conformance test suite and registry stubs
- Add a minimal README for onboarding users and developers