code-conservation

Source code spectral analysis — measure structural coherence of codebases via the conservation spectral framework.

Treats source code as a graph: functions/modules are nodes, call/reference edges connect them. Measures how "conserved" structural properties (cyclomatic complexity, depth, naming) are across the codebase. Clean code has high conservation (consistent patterns); spaghetti code has low conservation (chaotic structure).

What This Gives You

• Code graph construction — AST-based call/reference graphs
• Structural attribute tracking — complexity, nesting depth, line count per function
• Conservation scoring — how consistent is the codebase's structure?
• Style comparison — Python idiomatic vs Rust-style Python (included samples)
• Anomaly detection — find functions that break the codebase's structural pattern

Quick Start

python code_conservation.py

Includes sample files: samples/clean-python.py and samples/clean-rust-style.py for comparison.

How It Fits

Part of the SuperInstance ecosystem:

• linguistic-spectral — Text/language spectral analysis
• code-conservation — Source code spectral analysis (this repo)

License

MIT