Skip to content

Community Detection

What is it?

Your code naturally forms "neighborhoods" or communities. Even if you didn't plan it, some classes talk to each other more than others.

AST Metrics uses the Louvain Algorithm (a graph clustering algorithm) to detect these communities automatically based on the dependency graph.

Why it matters?

  • Reality Check: Does the code structure match your folder structure?
  • Leak Detection: Are "Domain" classes talking to "Infrastructure" classes when they shouldn't?
  • Modularity: Helps identify natural boundaries for microservices or modules.

Communities Explanation

Interpreting the Results

Purity

AST Metrics calculates the "purity" of a community by checking if its members belong to the same namespace.

  • High Purity: The community corresponds to a namespace. Good modularity.
  • Low Purity: The community is a mix of unrelated namespaces. This often indicates Spaghetti Architecture or hidden dependencies.

Modularity Score

A global score (0-1) indicating how well the network divides into modular communities. High modularity means dense connections within modules and sparse connections between them.