Skip to main content

Architecture Overview

Understanding the internal architecture of DataGeneration.

High-Level Architecture​

DSL JSON → Parser → DSL Tree → Visitor → Generated Data
  1. DSL Parsing: JSON is parsed into a tree of DslNode objects
  2. Validation: Tree is validated for correctness
  3. Visitor Pattern: DataGenerationVisitor traverses the tree
  4. Generation: Generators create data based on node types
  5. Output: Data is serialized to JSON, SQL, or custom formats

Core Components​

DSL Tree​

The DSL is represented as a tree of nodes:

  • RootNode - Top-level container
  • CollectionNode - Represents a collection
  • ItemNode - Template for collection items
  • GeneratedFieldNode - Field with generator
  • LiteralFieldNode - Static value field
  • ArrayFieldNode - Array field
  • ReferenceNode - Reference to another collection

Visitor Pattern​

The DataGenerationVisitor traverses the DSL tree and generates data:

public interface DslNodeVisitor {
    void visit(RootNode node);
    void visit(CollectionNode node);
    void visit(GeneratedFieldNode node);
    // ... other node types
}

Generator Registry​

All generators are registered in GeneratorRegistry:

GeneratorRegistry registry = new GeneratorRegistry();
registry.register("uuid", new UuidGenerator());
registry.register("name", new NameGenerator());
// ... etc

Generation Contexts​

Two context implementations:

  • EagerGenerationContext - Generates all data upfront
  • LazyGenerationContext - Generates data on-demand

Design Patterns​

Visitor Pattern​

Used for traversing and processing the DSL tree without modifying node classes.

Builder Pattern​

DslDataGenerator.Builder provides fluent API for configuration.

Strategy Pattern​

FilteringBehavior enum allows different filtering strategies.

Proxy Pattern​

Lazy generation uses proxy objects for on-demand field materialization.

Registry Pattern​

GeneratorRegistry manages all available generators.

Next Steps​