code-pattern-extractor

Code Pattern Extractor

Overview

Systematically analyze code in directories or modules to identify recurring patterns, code duplication, and implementation patterns that can be abstracted into reusable components, utilities, or design patterns.

Workflow

1. Define Analysis Scope

Identify the code to analyze:

• Directory/module: Analyze all files in a specific directory
• File set: Analyze a specific set of related files
• Component: Analyze files related to a specific feature or component

Use Glob to find relevant files:

**/*.js, **/*.py, **/*.go, etc.

2. Scan for Code Duplication

Identify repeated code blocks that appear multiple times:

Look for:

• Similar function implementations with minor variations
• Repeated code blocks (>5 lines) across files
• Copy-pasted code with slight modifications
• Similar class structures or method patterns

Analysis criteria:

• Similarity threshold: >70% code similarity
• Minimum size: 5+ lines of code
• Frequency: Appears 3+ times

3. Identify Implementation Patterns

Find recurring implementation approaches:

Common patterns:

• API call patterns: Similar fetch/request handling
• Error handling: Repeated try-catch or error checking
• Data validation: Similar input validation logic
• Data transformation: Repeated mapping/filtering operations
• State management: Similar state update patterns
• Configuration: Repeated configuration setup

Example patterns to detect:

// Pattern: API call with error handling
async function fetchX() {
  try {
    const response = await fetch(url);
    if (!response.ok) throw new Error();
    return await response.json();
  } catch (error) {
    console.error(error);
    return null;
  }
}

4. Categorize Patterns

Group identified patterns by type and impact:

Categories:

• High-value: Appears frequently (5+ times), significant code size
• Medium-value: Appears moderately (3-4 times), moderate complexity
• Low-value: Appears rarely (2 times), simple code