PropFlow

Contents:

  • Installation
    • From PyPI (Recommended)
    • From Source
    • Development Installation
    • Requirements
    • Verifying Installation
  • Quick Start Guide
    • Prepare the Environment
    • Step 1 — Build a Factor Graph (FGBuilder First)
    • Step 2 — Run Belief Propagation
    • Step 3 — Try Engine Variants and Policies
    • Step 4 — Scale Out with the Simulator
    • Optional — Build a Custom Graph Manually
    • Inspecting Runs with Analyzer Tooling
    • Where to Go Next
  • PropFlow User Guide
    • Top-Down Architecture
    • Agents Layer
      • Variable Agents
      • Factor Agents
      • Message Lifecycle
    • Factor Graph Layer
      • Using FGBuilder
      • Config-Driven Graphs
      • Manual Graph Assembly
    • Engine Layer
      • Selecting a Computator
      • Engine Variants and Policies
      • Running a Single Engine
    • Simulation Layer
    • Analysis Layer
    • Chain of Creation
      • Custom Graph Checklist
    • Next Steps
  • PropFlow Handbook
    • Conceptual Overview
      • 1. What PropFlow Solves
      • 2. Core Runtime Concepts
      • 3. Creation Pipeline (Top-Down)
      • 4. Directory Layout
      • 5. Data Flow Through an Engine Step
      • 6. Extensibility Points
      • 7. Operational Roles
    • System Setup & Dependencies
      • 1. Supported Platforms
      • 2. Core Dependencies
      • 3. Environment Bootstrapping
        • Option A – Recommended: uv
        • Option B – Standard virtualenv + pip
        • Option C – Conda
      • 4. Verifying Installation
      • 5. External Services & Data
      • 6. File System Expectations
    • Running Simulations & Tools
      • 1. Quick Smoke Test (main.py)
      • 2. Using the Simulator API
      • 3. Command-Line Interface (bp-sim)
      • 4. Snapshot Capture & Analysis
        • 4.1 EngineSnapshotRecorder
        • 4.2 SnapshotVisualizer
      • 5. Examples Directory
      • 6. Testing & Validation
        • Unit Tests
        • Coverage
        • Static Analysis
      • 7. Logging & Artefacts
    • Deployment Playbooks
      • 1. Local Workstation or Lab Server
        • Steps
        • Tips
      • 2. Containerised Deployment (Docker)
        • Example Dockerfile
        • Build & Run
        • Notes
      • 3. HPC / Batch Scheduling
      • 4. Web Service / API Integration
      • 5. Packaging & Distribution
      • 6. Configuration Management
      • 7. Logging & Monitoring
    • Development Workflow
      • 1. Repository Hygiene
      • 2. Tooling
      • 3. Coding Guidelines
      • 4. Testing Strategy
      • 5. Adding New Features
      • 6. Release Process
      • 7. Documentation Maintenance
      • 8. Collaboration
    • Reference Tables
      • 1. Modules & Key Entrypoints
      • 2. Configuration Defaults (global_config_mapping.py)
      • 3. CLI & Scripts
      • 4. Environment Variables
      • 5. File & Directory Conventions
      • 6. Glossary
    • Troubleshooting & FAQ
      • 1. Installation Issues
        • ModuleNotFoundError: No module named 'propflow'
        • python: command not found
        • Dependency compilation failures
      • 2. Runtime Errors
        • RuntimeError: Invalid default configuration: Missing required engine config key
        • ModuleNotFoundError: No module named 'networkx'
        • Multiprocessing permissions (macOS sandbox errors)
      • 3. Performance & Resource Usage
      • 4. Visualisation Problems
      • 5. Snapshot Recorder Questions
      • 6. Deployment Questions
      • 7. Getting Support
    • Analyzer and Reporting Workflows
      • 1. Capturing Snapshots
      • 2. Parsing Snapshot Payloads
      • 3. Running the Snapshot Analyzer
      • 4. Generating Reports and Artefacts
      • 5. Command Line Interface
      • 6. Worked Examples
      • 7. Best Practices
  • API Reference
    • Core Modules
      • Engines
        • BPEngine
        • Engine
        • SplitEngine
        • CostReductionOnceEngine
        • DampingEngine
        • DampingSCFGEngine
        • DampingCROnceEngine
        • MessagePruningEngine
        • DiscountEngine
      • Factor Graphs
        • FactorGraph
      • Agents
        • FGAgent
        • VariableAgent
        • FactorAgent
      • Components
        • Message
        • MailHandler
      • Policies
        • damp()
        • TD()
        • cost_reduction_all_factors_once()
        • discount_attentive()
        • split_all_factors()
        • MessagePruningPolicy
        • normalize_inbox()
        • normalize_soft_max()
        • normalize_cost_table_sum()
        • init_normalization()
      • Utilities
        • dummy_func()
        • FGBuilder
        • find_project_root()
        • CTFactory
      • Simulator
        • Simulator
  • Examples
    • Basic Graph Coloring
    • Comparing Engine Variants
    • Custom Cost Function
    • Large-Scale Problem
    • Analyzing Convergence
  • Contributing
    • Getting Started
    • Development Workflow
      • Running Tests
      • Code Quality
      • Documentation
    • Contribution Guidelines
      • Code Style
      • Commit Messages
      • Pull Requests
      • Testing
    • What to Contribute
      • Need Ideas?
    • Questions?
    • License
PropFlow
  • Overview: module code

All modules for which code is available

  • propflow.bp.engine_base
  • propflow.bp.engines
  • propflow.bp.factor_graph
  • propflow.core.agents
  • propflow.core.components
  • propflow.policies.cost_reduction
  • propflow.policies.damping
  • propflow.policies.message_pruning
  • propflow.policies.normalize_cost
  • propflow.policies.splitting
  • propflow.simulator
  • propflow.utils.fg_utils
  • propflow.utils.general_utils
  • propflow.utils.path_utils

© Copyright 2025, Or Muller.

Built with Sphinx using a theme provided by Read the Docs.