code_analyzer

Unused Code Detection

Refactoroscope includes a powerful unused code detection feature that helps identify potentially dead code in your Python projects.

How It Works

The unused code detection uses static analysis based on Python’s Abstract Syntax Tree (AST) to:

1. Parse Python source files into ASTs
2. Track definitions of functions, classes, variables, and imports
3. Track usages of these definitions
4. Report definitions that have no corresponding usages

Usage

To analyze your project for unused code, run:

uv run refactoroscope unused /path/to/your/project

This will scan all Python files in the specified directory and report any unused code elements.

Output Format

The output shows:

• File path where unused code is found
• Type of unused element (function, class, variable, import)
• Name of the unused element
• Line number where it’s defined
• Confidence level (higher confidence means more likely to be truly unused)

Confidence Levels

Different types of code elements have different confidence levels:

• Imports: 90% confidence
• Functions and classes: 70% confidence
• Variables: 60% confidence

The confidence level helps you prioritize which findings to investigate first.

Limitations

As with all static analysis tools, there are some limitations:

1. Dynamic references: Code that is referenced dynamically (e.g., through getattr) may be incorrectly flagged as unused
2. Entry points: Code that serves as an entry point but isn’t called within the codebase may be flagged
3. Framework code: Code that integrates with frameworks may have false positives

Best Practices

1. Review findings with high confidence first
2. Consider the context of your application when evaluating findings
3. Add truly necessary code that is flagged as unused to a whitelist or documentation
4. Use version control to track changes when removing unused code

This site is open source. Improve this page.