MindOT - Mind Operational Technology

    MindOT is a decision engine for the agent system, providing structured reasoning for complex trade-offs. It breaks down complex tasks into manageable sub-tasks using multiple decision-making methods.
    
    ## Features
    
    - **Task Decomposition**: Break complex tasks into manageable sub-tasks
    - **Hierarchical Planning**: Create Project → Operation → Job → Task structures
    - **Multiple Decision Methods**: Bayesian inference, MCDA, decision trees, rule engines, ML evaluation
    - **CLI Interface**: Easy command-line interaction with all modules
    - **Configurable**: Flexible configuration via pyproject.toml and mindot.yaml
    - **ML Evaluation**: Data-driven option weighting using sklearn metrics (AUC, F1, LogLoss)
    - **Feature Analysis**: PCA and t-SNE for dimensionality reduction
    
    ## Installation
    
    ```bash
    pip install -e .
    ```
    
    ### Requirements
    - Python >= 3.10
    - pyyaml>=6.0
    - numpy
    - scikit-learn (for ML evaluation features)
    
    ## Usage
    
    ### Command Line
    
    ```bash
    # Basic commands
    mindot advanced          # Run advanced analysis
    mindot bayesian --samples 1000   # Bayesian inference
    mindot mcda --criteria 5        # Multi-criteria decision analysis
    mindot simulation --steps 100   # Run simulation
    mindot trees             # Tree operations
    ```
    
    ### Python API
    
    ```python
    from mindot import MindOT
    
    # Initialize MindOT
    mt = MindOT()
    
    # Run analysis
    result = mt.analyze(data)
    ```
    
    #### Bayesian Analysis
    
    ```python
    from mindot.bayesian import BayesianAnalyzer
    
    analyzer = BayesianAnalyzer()
    result = analyzer.infer(evidence)
    ```
    
    #### MCDA
    
    ```python
    from mindot.mcda import MCDA
    
    mcda = MCDA(criteria=['cost', 'quality', 'time'])
    scores = mcda.rank(alternatives)
    ```
    
    #### Decision Trees
    
    ```python
    from mindot.trees import DecisionTree
    
    tree = DecisionTree()
    tree.train(X, y)
    prediction = tree.predict(new_data)
    ```
    
    ## Core Modules
    
    | Module | Description |
    |--------|-------------|
    | `mindot.mindot` | Core decision engine manager |
    | `mindot.models` | Internal ML models and evaluation |
    | `mindot.rules` | Rule and trigger engine |
    | `mindot.engines` | Decision engine implementations |
    | `mindot.bayesian` | Bayesian inference |
    | `mindot.mcda` | Multi-criteria decision analysis |
    | `mindot.trees` | Decision tree implementations |
    | `mindot.cli` | Command-line interface |
    
    ## Development
    
    ```bash
    # Install dev dependencies
    pip install -e ".[dev]"
    
    # Run tests
    pytest tests/ -v
    ```
    
    ## Configuration
    
    MindOT can be configured via:
    - `pyproject.toml` - Build and tool configuration
    - `mindot.yaml` - Runtime behavior settings
    
    ## Contributing
    
    See [CONTRIBUTING.md](CONTRIBUTING.md) for guidelines.
    
    ## License
    
    MIT License - See [LICENSE](LICENSE) file for details.