config.yaml

# GNN Processing Pipeline Configuration
# This file contains configuration settings for the GNN processing pipeline

# Pipeline execution settings
pipeline:
  # Enable/disable pipeline execution
  enabled: true

  # Pipeline steps to execute (0-24)
  # Leave empty to run all steps
  steps: []

  # Skip specific steps by number (e.g. [13] to skip LLM). Default: run all steps.
  # Step 13 needs a local LLM (e.g. Ollama) and the model from src/llm/defaults.py or OLLAMA_MODEL.
  skip_steps: []

  # Step 2 test mode when not overridden on CLI (see src/utils/pipeline_config_merge.py)
  fast_only: true
  # comprehensive: false   # set true to run the full pytest suite in step 2

  # Force execution of specific steps
  force_steps: []

  # Parallel execution settings
  parallel:
    enabled: false
    max_workers: 4

  # Timeout settings (in seconds)
  timeout:
    step: 3600 # 1 hour per step
    total: 7200 # 2 hours total

# Step 1 defaults (read by main.py → 1_setup). CLI flags override these when present.
setup:
  # uv sync --extra dev (pytest-cov, xdist, coverage, linters, …)
  dev: true
  recreate_venv: false
  # install_all_extras: false   # if true: uv sync --all-extras (every optional group; heavy)

# Legacy / documentation only: main.py does not read uv.sync.* (use ``setup:`` above).
# For manual installs: uv sync --extra dev | uv sync --all-extras
uv:
  recreate_uv_env: false
  uv_env_dir: ".venv"

# Input/output settings
io:
  # Input directory for GNN files
  input_dir: "input/gnn_files"

  # Output directory for pipeline artifacts
  output_dir: "output"

  # Temporary directory
  temp_dir: "temp"

  # Log directory
  log_dir: "logs"

# ═══════════════════════════════════════════════════════════════════════════════
# Testing Matrix Configuration
# Maps subdirectories in input/gnn_files/ to specific pipeline steps.
# Edit the lists below to control which steps run on which folders.
# ═══════════════════════════════════════════════════════════════════════════════
#
# ┌──────────────────────────────────────────────────────────────────────────────
# │ STEP REFERENCE (select by number)
# ├──────┬──────────────────────────┬──────────────────────────────────────────
# │ Step │ Script (src/)            │ Purpose
# ├──────┼──────────────────────────┼──────────────────────────────────────────
# │   0  │ 0_template.py            │ Pipeline template & initialization
# │   1  │ 1_setup.py               │ Environment setup & dependency install
# │   2  │ 2_tests.py               │ Test suite execution
# │   3  │ 3_gnn.py                 │ GNN file discovery & parsing
# │   4  │ 4_model_registry.py      │ Model registry & versioning
# │   5  │ 5_type_checker.py        │ GNN type validation
# │   6  │ 6_validation.py          │ Consistency & quality checking
# │   7  │ 7_export.py              │ Multi-format export (JSON, XML, GraphML)
# │   8  │ 8_visualization.py       │ Graph & matrix visualization
# │   9  │ 9_advanced_viz.py        │ Interactive / advanced visualization
# │  10  │ 10_ontology.py           │ Active Inference ontology processing
# │  11  │ 11_render.py             │ Code generation (PyMDP, RxInfer, JAX…)
# │  12  │ 12_execute.py            │ Execute rendered simulation scripts
# │  13  │ 13_llm.py                │ LLM-enhanced analysis
# │  14  │ 14_ml_integration.py     │ Machine learning integration
# │  15  │ 15_audio.py              │ Audio sonification generation
# │  16  │ 16_analysis.py           │ Statistical analysis & reporting
# │  17  │ 17_integration.py        │ System integration checks
# │  18  │ 18_security.py           │ Security validation
# │  19  │ 19_research.py           │ Research tools & literature
# │  20  │ 20_website.py            │ Static HTML website generation
# │  21  │ 21_mcp.py                │ Model Context Protocol processing
# │  22  │ 22_gui.py                │ Interactive GNN constructor GUI
# │  23  │ 23_report.py             │ Comprehensive report generation
# │  24  │ 24_intelligent_analysis  │ AI-powered pipeline analysis
# └──────┴──────────────────────────┴──────────────────────────────────────────
#
# NOTE: Global steps run once before folder-specific steps (not folder-routed).
#       Steps not in any folder's list run once on the root input directory.
# ═══════════════════════════════════════════════════════════════════════════════

testing_matrix:
  enabled: true

  # ─────────────────────────────────────────────────────────────────────────────
  # GLOBAL STEPS — run once (not per-folder), togglable independently
  # Set individual steps to true/false to include/exclude them from the pipeline
  # ─────────────────────────────────────────────────────────────────────────────
  global_steps:
    0_template: true # Pipeline template & initialization
    1_setup: true # Environment setup & dependency install
    2_tests: true # Test suite execution

  # Default steps for folders not explicitly listed below
  default_steps: [3, 5, 6, 10]

  # Folder-specific step configurations
  # Each key = subdirectory name in input/gnn_files/
  # Each value = list of step numbers to run on that folder
  folders:
    # POMDP agents (full pipeline — all steps)
    discrete: [
        3,
        4,
        5,
        6,
        7,
        8,
        9,
        10,
        11,
        12,
        # 13,  # LLM — disabled (slow Ollama calls); re-enable when needed
        14,
        15,
        16,
        17,
        18,
        19,
        20,
        21,
        22,
        23,
        24,
      ]

    # Basic perception models
    basics: [3, 5, 6, 10] # Parse, type-check, validate, ontology

    # Continuous-space models
    continuous: [3, 5, 6] # Parse and type-check

    # Hierarchical models
    hierarchical: [3, 5, 6] # Parse and type-check

    # Multi-agent models
    multiagent: [3, 5, 6] # Parse and type-check

    # Precision-weighted models
    precision: [3, 5, 6] # Parse and type-check

    # Structured / factor graph models
    structured: [3, 5, 6] # Parse and type-check

# LLM settings (Step 13)
llm:
  # Default model for Ollama (overridden by OLLAMA_MODEL env var)
  model: "smollm2:135m-instruct-q4_K_S"

  # Total budget for LLM processing in seconds (pipeline kills at 900s)
  timeout_seconds: 600

  # Per-prompt timeout in seconds
  prompt_timeout: 45

# Logging settings
logging:
  # Log level (DEBUG, INFO, WARNING, ERROR, CRITICAL)
  level: "INFO"

  # Log format
  format: "%(asctime)s - %(name)s - %(levelname)s - %(message)s"

  # Log file settings
  file:
    enabled: true
    max_size: "10MB"
    backup_count: 5

# Validation settings
validation:
  # Enable validation checks
  enabled: true

  # Strict validation mode
  strict: false

  # Validation timeout (seconds)
  timeout: 300

# Performance settings
performance:
  # Memory limit (MB)
  memory_limit: 4096

  # CPU limit (cores)
  cpu_limit: 4

  # Cache settings
  cache:
    enabled: true
    max_size: "1GB"
    ttl: 3600 # 1 hour

# Security settings
security:
  # Enable security checks
  enabled: true

  # Allow network access
  network_access: true

  # File permissions
  file_permissions: "644"