Docs
Tutorial
YAML Configuration

YAML Configuration

YAML is a human-readable data serialization format used by PETsARD for experiment configuration. This document explains how to structure your YAML configurations effectively.

Basic Structure

YAML configurations in PETsARD follow a three-level hierarchy:

ModuleName:           # First level: Module
    ExperimentName:   # Second level: Experiment
        param1: value # Third level: Parameters
        param2: value

Module Level

The top level defines the processing modules in execution order:

  • Loader: Data loading
  • Preprocessor: Data preprocessing
  • Synthesizer: Data synthesis
  • Postprocessor: Data postprocessing
  • Constrainer: Data constraining
  • Evaluator: Result evaluation
  • Reporter: Report generation

Experiment Level

Each module can have multiple experiment configurations:

Synthesizer:
    exp1_ctgan:        # First experiment
        method: ctgan
        epochs: 100
    exp2_tvae:         # Second experiment
        method: tvae
        epochs: 200

Parameter Level

Parameters follow each module’s specific requirements:

Loader:
    demo_load:
        filepath: 'data/sample.csv'
        na_values:
            age: '?'
            income: 'unknown'
        column_types:
            category:
                - gender
                - occupation

Execution Flow

When multiple experiments are defined, PETsARD executes them in a depth-first order:

Loader -> Preprocessor -> Synthesizer -> Postprocessor -> Constrainer -> Evaluator -> Reporter

For example:

Loader:
    load_a:
        filepath: 'data1.csv'
    load_b:
        filepath: 'data2.csv'
Synthesizer:
    syn_ctgan:
        method: ctgan
    syn_tvae:
        method: tvae

This creates four experiment combinations:

  1. load_a + syn_ctgan
  2. load_a + syn_tvae
  3. load_b + syn_ctgan
  4. load_b + syn_tvae

Reporting Options

Reporter supports two methods:

Data Saving 

Reporter:
    save_data:
        method: 'save_data'
        source: 'Postprocessor'  # Module to save data from

Report Generation 

Reporter:
    save_report:
        method: 'save_report'
        granularity: 'global'    # Report detail level

Best Practices

  1. Use meaningful experiment names
  2. Keep parameters organized by module
  3. Document experiment configurations
  4. Validate YAML syntax before running
TutorialDefault Synthesis