This example demonstrates how to use Graph-sitter to automatically analyze and visualize module dependencies in Python codebases. The script creates a directed graph showing relationships between different modules, making it easier to understand code architecture and dependencies.

The script analyzes module dependencies in several key steps:

1. Graph Initialization

   G = nx.DiGraph()
   list_apps = ["src/sentry/api", "src/sentry/auth", "src/sentry/flags"]
   for app in list_apps:
       G.add_node(app, metadata={"color": "red"})

   • Creates a directed graph using NetworkX
   • Initializes nodes for each major application module
   • Sets up metadata for visualization

2. Import Analysis

   for file in codebase.files:
       if app in file.filepath:
           for import_statement in file.import_statements:
               # Analyze imports and build edges

   • Scans through all files in specified modules
   • Analyzes import statements
   • Creates edges based on module dependencies

3. Graph Cleanup

   nodes_to_remove = [node for node, degree in G.degree() if degree == 1]
   G.remove_nodes_from(nodes_to_remove)

   • Removes isolated nodes
   • Cleans up the graph for better visualization
   • Focuses on meaningful dependencies

1. Automated Dependency Detection

   • Automatically finds module relationships
   • Identifies import patterns
   • No manual tracking needed

2. Visual Representation

   • Clear visualization of dependencies
   • Easy to identify clusters
   • Highlights potential architectural issues

3. Simplified Analysis

   • Quick overview of codebase structure
   • Helps identify tightly coupled modules
   • Assists in refactoring decisions

# The script will detect dependencies like:
from src.sentry.api import endpoint  # Creates edge from current module to api
from src.sentry.auth import tokens  # Creates edge from current module to auth

DiGraph with n nodes and m edges where:
- Nodes represent major modules
- Edges show import relationships
- Node colors indicate module types

1. Better Architecture Understanding

   • Clear view of module relationships
   • Identifies dependency patterns
   • Helps spot architectural issues

2. Refactoring Support

   • Identifies tightly coupled modules
   • Helps plan refactoring
   • Shows impact of changes

3. Documentation Aid

   • Visual documentation of architecture
   • Easy to share and discuss
   • Helps onboard new developers

# install graph-sitter and dependencies
pip install graph-sitter networkx

# Run the visualization
python run.py

The script will:

1. Initialize the codebase
2. Analyze module dependencies
3. Create a dependency graph
4. Output the visualization through codegen.sh

You can customize the analysis by:

• Modifying the list_apps to include different modules
• Adjusting node metadata and colors
• Adding additional filtering criteria

• NetworkX Documentation
• Python Import System
• Graph-sitter Documentation
• Graph visualization

Feel free to submit issues and enhancement requests! Contributions to improve the visualization or add new features are welcome.