Skip to main content

Synthetic Data - API Specification

Version: v1.0.0 | Status: Active | Last Updated: March 2026

1. Overview​

The synthetic_data module provides template-based and LLM-powered synthetic data generation for training, testing, and augmentation purposes.

2. Core Components​

2.1 Classes​

ClassDescription
SyntheticDataGeneratorMain generator orchestrating data creation from schema definitions
TemplateGeneratorTemplate-based generator using string interpolation and rules
DataSchemaSchema definition for synthetic datasets (fields, types, constraints)

3. Usage Example​

from codomyrmex.synthetic_data import SyntheticDataGenerator, DataSchema

schema = DataSchema(fields={
    "name": {"type": "string", "pattern": "faker.name"},
    "age": {"type": "int", "min": 18, "max": 80},
    "score": {"type": "float", "min": 0.0, "max": 1.0},
})

gen = SyntheticDataGenerator(schema)
dataset = gen.generate(n=1000)
print(f"Generated {len(dataset)} records")

4. Navigation​