SPEC.md

Specification: Generalized Notation Notation (GNN) Ecosystem

Design Requirements

This repository comprises the unified Generalized Notation Notation (GNN) ecosystem. At its foundational architectural root, it is built to orchestrate the translation of conceptual Active Inference generative models (written in highly structured Markdown) into dynamically actionable and executable scientific artifacts over a 25-step execution pipeline.

The repository root governs 100% Zero-Mock compliance across all integrations, strict testing schemas, continuous documentation integration, and CI/CD security definitions. The environment handles multi-parameter rendering targets across complex computational engines including JAX/PyMDP, NumPyro, RxInfer, and PyTorch.

Components

There are no python source packages instantiated directly within this root location. Instead, the root contains the infrastructural map that triggers the pipeline execution and validation layer:

1. src/: Master source tree for the 25-step gnn execution orchestrators.
2. doc/: Deep-linked, versioned, extensive framework mapping and cognitive systems documentation.
3. tests/ (via src/tests): Zero-Mock execution boundaries.
4. .github/: Declarative workflow integration interfaces for continuous validation.
5. scripts/: Specialized tooling.
6. input/ & output/: The decoupled transactional zones mapping the inputs of configurations and .md models toward serialized, visualized, ML-computed deliverables.
7. Infrastructural Configs (pyproject.toml, pytest.ini, uv.lock): Defines environments and strict operational capacities using uv.

Technical Rules

• The root tier must maintain 100% explicit module compliance manifesting exactly three primary documentation files per node (AGENTS.md, README.md, SPEC.md).
• Avoid exposing any .py script behaviors at the root; invoke logic strictly through explicit interfaces inside src/.