edp-mep

Environment Design Pattern (EDP) + Model Environment Protocol (MEP)

OneOrigine / ImperialSchool Research — I.S. License

"A perfect balance between rationality and probability is necessary to fully leverage the logic of artificial intelligence." — Seikatsu-One

Install

pip install edp-mep
# With provider support:
pip install edp-mep[ollama]     # local Ollama
pip install edp-mep[openai]     # OpenAI API
pip install edp-mep[anthropic]  # Anthropic API
pip install edp-mep[all]        # all providers

Mathematical Foundation

Central Equation: E_{t+1} = 𝔘( E_t, 𝔯( x_t, c_t, Σ_t, 𝔄(x_t,c_t,Σ_t,Γ_t, Ψ(D_t,c_t,Σ_t)) ) )

Harmony Function: H = α·cos(A,C) + β·cos(A,S) + γ·cos(R̂,R) − δ·D

Quick Start

from edp import Environment, Context, Action, Circumstance, SenseVector, EnvironmentKind
from mep import MepGateway

# Build environment
env = Environment("MyEnv", EnvironmentKind.REACTIVE)
ctx = env.create_context("Main",
    basis=SenseVector.normative("operations", 0.9))

# Run agents via MEP
gateway = MepGateway(env)
session = gateway.connect("my-agent")
envelope = gateway.build_envelope(session, actor, ctx)
reaction = await gateway.dispatch(session, actor, "action.type", payload, ctx)

Usage (agent demo)

python mep_ollama_agent_v4.py --demo
python mep_ollama_agent_v4.py --model gemma3:12b --inject-memory
python mep_ollama_agent_v4.py --provider openai --model gpt-4o-mini
python mep_ollama_agent_v4.py --provider anthropic --model claude-3-haiku-20240307