Skip to content

rxinfer

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

rxinfer

README.md

RxInfer.jl Integration for GNN

📋 Document Metadata
Type: Framework Integration Guide | Audience: Researchers, Developers | Complexity: Intermediate-Advanced
Cross-References: AGENTS.md | GNN RxInfer Guide | Framework Integration | Main Documentation | doc/SPEC.md (versioning)

Overview

This directory contains documentation, scripts, and resources for integrating GNN (Generalized Notation Notation) models with RxInfer.jl, a Julia-based reactive Bayesian inference framework. RxInfer.jl provides efficient message-passing inference for probabilistic models, making it ideal for Active Inference simulations.

Status: ✅ Production Ready
Version: 1.0

Quick Navigation

This Directory

Main Documentation

Related Directories

Pipeline Integration

Contents

Files: 12+ | Subdirectories: 1

Core Files

  • gnn_rxinfer.md: Complete RxInfer.jl integration guide

    • RxInfer.jl framework overview
    • GNN to RxInfer.jl translation
    • Code generation patterns
    • Example models and usage

  • Multiagent_GNN_RxInfer.jl: Validation script

    • Validates GNN to RxInfer.jl translation
    • Two-stage validation process
    • Configuration file generation testing

  • engineering_rxinfer_gnn.md: Engineering guide

    • Technical implementation details
    • Best practices and patterns

  • AGENTS.md: Technical documentation and agent scaffolding

    • Complete documentation structure
    • Integration with pipeline
    • Cross-references and navigation

  • README.md: Directory overview (this file)

Subdirectories

  • multiagent_trajectory_planning/: Multi-agent trajectory planning examples
    • Complete RxInfer.jl implementations
    • Configuration examples
    • Results and analysis

RxInfer.jl Integration

Framework Overview

RxInfer.jl is a reactive Bayesian inference framework for Julia that provides:

  • Reactive Probabilistic Programming: Dynamic model construction and inference
  • Efficient Message Passing: Optimized inference algorithms
  • Factor Graph Models: Natural representation of Active Inference models
  • Streaming Inference: Real-time belief updating
  • Multi-agent Support: Coordinated multi-agent systems

GNN to RxInfer.jl Translation

The GNN pipeline translates GNN models to RxInfer.jl through:

  1. Model Parsing: GNN syntax parsed into structured representation
  2. Factor Graph Construction: Active Inference components mapped to factor graph
  3. Code Generation: Julia code generation with RxInfer.jl API
  4. Configuration Generation: TOML configuration files for model parameters
  5. Validation: Automated validation of generated code

Validation Process

The Multiagent_GNN_RxInfer.jl script validates the translation pipeline:

Stage 1: Baseline Simulation

  • Locates standard "Multi-agent Trajectory Planning" example
  • Runs with original hand-written config.toml
  • Establishes baseline for successful execution

Stage 2: GNN-Configured Simulation

  • Creates new validation directory
  • Copies Julia script files from original example
  • Replaces config.toml with GNN-generated configuration
  • Executes simulation with GNN-derived configuration
  • Compares results with baseline

Validation Success Criteria

Successful validation demonstrates:

  • Syntactic Correctness: GNN parser produces valid TOML configuration
  • Parameter Translation: GNN parameters correctly translated to RxInfer.jl values
  • End-to-End Functionality: Complete pipeline from GNN model to RxInfer.jl simulation
  • Result Equivalence: GNN-configured results match baseline expectations

Integration with Pipeline

This documentation is integrated with the 25-step GNN processing pipeline:

  1. Core Processing (Steps 0-9): GNN parsing, validation, export

    • GNN models parsed and validated
    • RxInfer.jl code generation (Step 11: Render)

  2. Simulation (Steps 10-16): Model execution and analysis

    • RxInfer.jl execution (Step 12: Execute)
    • Results processing and analysis

  3. Integration (Steps 17-24): System coordination and output

    • RxInfer.jl results integrated into comprehensive outputs
    • Multi-agent coordination and analysis

See src/AGENTS.md for complete pipeline documentation.

Usage Examples

Running Validation

# Ensure Julia environment with required packages
julia doc/rxinfer/Multiagent_GNN_RxInfer.jl

Basic RxInfer.jl Model

GNN models are translated to RxInfer.jl factor graphs:

using RxInfer

# GNN model translated to RxInfer.jl
@model function gnn_model(observations, actions)
    # Hidden state beliefs
    s_f0 ~ Categorical(prior)
    
    # Observations
    o_m0 ~ Categorical(A * s_f0)
    
    # State transitions
    s_f0_next ~ Categorical(B[s_f0, actions])
    
    return s_f0, o_m0, s_f0_next
end

Related Resources

Main GNN Documentation

Framework Integration

Pipeline Architecture

Standards and Guidelines

All documentation in this module adheres to professional standards:

  • Clarity: Concrete, technical writing with Julia and RxInfer.jl foundations
  • Functionality: Describes actual RxInfer.jl integration capabilities
  • Completeness: Comprehensive coverage of RxInfer.jl integration
  • Consistency: Uniform structure and style with GNN documentation ecosystem

See Also

Status: ✅ Production Ready
Compliance: Professional documentation standards
Maintenance: Regular updates with new RxInfer.jl features and integration capabilities