README · c2ab42b03dfb

interdependent-lib

A unified PyPI meta-package that brings together the four-letter and five-letter acronym libraries of The Interdependency into a single, installable collection.

Overview

interdependent-lib is the convergence point for the mathematical and AI orchestration libraries developed under The Interdependency. Installing this package makes each sub-library available in a single step.

Included Libraries

Installation

# Install the meta-package (base package only; extras install sub-libraries)
pip install interdependent-lib

# Install only specific libraries
pip install interdependent-lib[pcea]
pip install interdependent-lib[ucns]   # ucns>=0.9.1
pip install interdependent-lib[aimmh]
pip install interdependent-lib[ptcna]  # ptcna>=0.1.1; repaired four-layer runtime

# Install everything
pip install interdependent-lib[all]

Release 0.1.5 pins the candidate-state pair exactly: UCNS b7b6f35cce69c273860923489a1c8b5372d14eb0 and PTCNA c5fa9a599498f19c8345f2790a0636542dfbc6a1. Run interdependent_lib.validate_installed_pair() after installation. Moving, registry-only, and unattributed installs fail closed; compatibility establishes construction, not usefulness.

Quick Start

import interdependent_lib

# Check which sub-libraries are available in this environment
print(interdependent_lib.available())

The Prime-Tensor Stack

The stack is one package — PTCNA (Prime Tensor Circled Neural Architecture) — with four layers. PCEA is the orthogonal guardian; ZFAE is the conceptual inference cap (runtime in a0). Composition counts are variable at every level — the only invariant is that every circle, seed, and core is itself a tensor. The canonical role-and-boundary map lives in docs/prime-tensor-stack.md:

neural tensors ─(circle layer)─► circles ─(seed layer)─► seeds ─(core layer)─► cores
  back-propagation lives ONLY in the neural layer
  circle / seed / core are auditing & timing tensors (non-differentiable);
    fiqs gate core internal propagation per Fick's law (J = −D ∇φ)
PCEA — guardian: "last state as key for this state" at every layer (orthogonal; not a layer)

The 0.1.1 runtime uses one shared circle type, keeps reverse-mode gradients owned by the neural layer, reports UCNS integration as typed and suspended, and consumes EDCM measurements only through an explicitly injected external provider.

Library Details

See libs/ for per-library documentation and source links:

• libs/ptcna/ — PTCNA (consolidated stack: neural/circle/seed/core)
• libs/pcea/ — PCEA (orthogonal guardian)
• libs/ucns/ — UCNS
• libs/zfae/ — ZFAE (conceptual; runtime in a0)
• libs/aimmh/ — AIMMH (five-letter)
• libs/metapat/ — METAPAT (FLAR)

Archived predecessors (edcmbone, pcna, pcta, pcsa, a0replite) are listed with their successors in docs/archive-registry.md — read-only ancestors; do not build against them.

Repository Structure

interdependent-lib/
├── README.md               ← you are here
├── COPILOT.md              ← GitHub Copilot + ecosystem best practices
├── CHANGELOG.md            ← version history
├── CONTRIBUTING.md         ← how to contribute
├── pyproject.toml          ← PyPI meta-package config
├── interdependent_lib/     ← top-level Python package
│   └── __init__.py
└── libs/                   ← per-library documentation and stubs
    ├── README.md
    ├── aimmh/
    ├── metapat/
    ├── pcea/
    ├── ptcna/
    ├── ucns/
    └── zfae/

Contributing

See CONTRIBUTING.md.

Changelog

See CHANGELOG.md.

License

MPL-2.0 (Mozilla Public License 2.0) — see LICENSE. Relicensed from MIT to MPL-2.0 — weak copyleft: embed anywhere, but changes to these files must be published. (Earlier history: AGPL-3.0-or-later + commercial, then MIT.)

Author

Erin Patrick Spencer — interdependentway.org

Open exact README source