agentsqripts/lib/wet-code/ast/semantic-similarity/calculators/calculateContentSimilarity.js

Comprehensive static code analysis toolkit for identifying technical debt, security vulnerabilities, performance issues, and code quality problems

/**
 * @file Content-based similarity calculator using advanced string distance algorithms
 * @description Single responsibility: Assess code block similarity through normalized textual content analysis
 * 
 * This calculator implements sophisticated content-based similarity using Levenshtein distance
 * to detect textual similarity between code blocks. It serves as the foundation layer for
 * similarity assessment, catching cases where structural analysis might miss surface-level
 * code duplication or near-duplicate implementations.
 * 
 * Design rationale:
 * - Content normalization reduces noise from formatting and naming differences
 * - Levenshtein distance provides accurate measurement of textual similarity between implementations
 * - Normalized scoring (0.0 to 1.0) enables consistent similarity comparison across content lengths
 * - Fallback to raw content ensures analysis continues when normalization unavailable
 * - Edge case handling for empty content prevents calculation errors
 * 
 * Content similarity methodology:
 * - Normalized content preferred for cleaner similarity assessment without formatting noise
 * - Levenshtein distance measures minimum edit operations needed to transform one block to another
 * - Distance normalization by maximum content length provides proportional similarity scoring
 * - Perfect similarity (1.0) for identical content after normalization
 * - Zero similarity (0.0) for completely different implementations requiring full rewrite
 */

const levenshtein = require('../../../../utils/levenshtein');

/**
 * Calculate content similarity using normalized content
 */
function calculateContentSimilarity(block1, block2) {
  const content1 = block1.normalizedContent || block1.content;
  const content2 = block2.normalizedContent || block2.content;
  
  // Use Levenshtein distance for content similarity
  const distance = levenshtein(content1, content2);
  const maxLength = Math.max(content1.length, content2.length);
  
  if (maxLength === 0) return 1.0;
  
  return 1 - (distance / maxLength);
}

module.exports = calculateContentSimilarity;