Dependency Graph

The dependency graph shows file-level and module-level dependencies in the legacy codebase. This goes deeper than the context map — it shows which files import which, where shared controllers create coupling, and where circular dependencies exist.

Interactive Visualization

What This Shows

The dependency graph renders cross-module relationships from the ERPNext reference dataset:

• Nodes represent modules or files
• Edges show import/dependency relationships
• Edge weight reflects the strength of coupling (number of shared references)
• Clusters indicate natural bounded context boundaries
• Red edges highlight circular dependencies that complicate extraction

Why Dependencies Matter for Modernization

Legacy systems accumulate implicit dependencies over years. In ERPNext:

• accounts_controller.py (4,412 lines) is imported by every transaction type
• The hooks system creates invisible execution paths across modules
• Regional overrides create country-specific dependency branches

Understanding these dependencies is a prerequisite for safe extraction. The dependency graph makes the invisible visible.

Controller Inheritance Example

ERPNext’s controller inheritance chain demonstrates deep coupling:

A single method change in AccountsController can affect every transaction type in the system.

Identifying Extraction Boundaries

The dependency graph helps identify:

Pattern	Meaning	Action
Dense cluster	High internal cohesion — good bounded context	Extract as a unit
Few cross-cluster edges	Low external coupling	Clean extraction boundary
Hub node	Many incoming/outgoing edges — shared controller	Decompose before extraction
Circular dependency	Modules depend on each other	Break cycle with ACL before extraction

Next Steps

• Complexity Heatmap — See where complexity concentrates
• domains.json Specification — Define bounded contexts