TrajectoryOS Models

TrajectoryOS Models

This directory contains the Python-based modeling and inference stack for TrajectoryOS.

Architecture

The models layer is the "brain" that the TypeScript Trajectory-Core service calls into for sophisticated inference:

• skill_graph/: Bayesian inference on skill competencies with graph-based message passing
• alignment/: Embedding-based project coherence and alignment scoring
• gravity_mass/: Constraint-based gravity estimation and structural mass computation
• life_state/: Full latent dynamical system for life-state evolution and forecasting
• echelon_fusion/: Embodied signal → life physics mapping (integrates movement data)
• generators/: Scenario generation and plan evaluation

Setup

cd models
python -m venv venv
source venv/bin/activate  # or `venv\Scripts\activate` on Windows
pip install -r requirements.txt

Key Concepts

The Physics Variables

• Thrust (T): Productive capacity from skills × time investment
• Alignment (A): How coherently projects and skills point in the same direction
• Gravity (G): Downward pull from constraints, obligations, stress
• Mass (M): Structural inertia from complexity and coupling
• Escape Index (η): η = T_eff / (G·M), where T_eff = T × A

Model Flow

Evidence → BayesianSkillModel → Skill beliefs (with uncertainty)
                                      ↓
Projects + Skills → AlignmentScorer → Coherence score (A)
                                      ↓
Constraints → GravityEstimator → Pull (G)
Projects + Dependencies → MassEstimator → Inertia (M)
                                      ↓
Current state + Actions → LifeStateDynamics → Future trajectories
                                      ↓
Scenarios → Evaluator → Ranked plans

Integration with Trajectory-Core

These models are called by:
- Background workers (periodic updates)
- Real-time API endpoints (on-demand inference)
- PlannerAgent (for scenario generation)

All communication happens via well-defined request/response schemas in `shared/types.py`.