All files / src cluster-detector.ts

100% Statements 92/92
77.27% Branches 34/44
100% Functions 12/12
100% Lines 71/71

Press n or j to go to the next uncovered block, b, p or k for the previous block.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122                    5x   5x 13x 13x 6x   13x 16x   16x 19x 24x 25x   24x 23x 12x 16x   16x 16x 16x 16x 23x 10x 16x 16x 10x 16x 7x 7x 12x   10x 8x     16x   10x     11x 16x 16x 23x 24x 17x 10x   16x 10x 5x 10x 17x   10x 10x   10x 12x 12x 12x   10x 7x     5x       16x 16x 5x 16x 14x 5x 15x 12x 12x 17x 13x 13x 12x     17x       5x 5x   5x                           11x    
import type { DependencyGraph, ModuleCluster } from './types';
import { calculateFragmentation, calculateEnhancedCohesion } from './metrics';
import {
  classifyFile,
  adjustFragmentationForClassification,
} from './classifier';
 
/**
 * Group files by domain to detect module clusters
 * @param graph - The dependency graph to analyze
 * @param options - Optional configuration options
 * @param options.useLogScale - Whether to use logarithmic scaling for calculations
 * @returns Array of module clusters
 */
export function detectModuleClusters(
  graph: DependencyGraph,
  options?: { useLogScale?: boolean }
): ModuleCluster[] {
  const domainMap = new Map<string, string[]>();
 
  for (const [file, node] of graph.nodes.entries()) {
    const primaryDomain = node.exports[0]?.inferredDomain || 'unknown';
    if (!domainMap.has(primaryDomain)) {
      domainMap.set(primaryDomain, []);
    }
    domainMap.get(primaryDomain)!.push(file);
  }

  const clusters: ModuleCluster[] = [];
 
  const generateSuggestedStructure = (
    files: string[],
    Etokens: number,
    fragmentation: number
  ) => {
    const targetFiles = Math.max(1, Math.ceil(tokens / 10000));
    const plan: string[] = [];
 
    if (fragmentation > 0.5) {
      plan.push(
        `Consolidate ${files.length} files scattered across multiple directories into ${targetFiles} core module(s)`
      );
    }
 
    if (tokens > 20000) {
      plan.push(
        `Domain logic is very large (${Math.round(tokens / 1000)}k tokens). Ensure clear sub-domain boundaries.`
      );
    }
 
    return { targetFiles, consolidationPlan: plan };
  };
 
  for (const [domain, files] of domainMap.entries()) {
    if (files.length < 2 || domain === 'unknown') continue;
 
    const totalTokens = files.reduce((sum, file) => {
      Econst node = graph.nodes.get(file);
      return sum + (node?.tokenCost || 0);
    }, 0);
 
    // Calculate shared import ratio for coupling discount
    let sharedImportRatio = 0;
    Eif (files.length >= 2) {
      const allImportSets = files.map(
        (f) => new Set(graph.nodes.get(f)?.imports || [])
      );
      let intersection = new Set(allImportSets[0]);
      const union = new Set(allImportSets[0]);
 
      for (let i = 1; i < allImportSets.length; i++) {
        const nextSet = allImportSets[i];
        intersection = new Set([...intersection].filter((x) => nextSet.has(x)));
        for (const x of nextSet) union.add(x);
      }
 
      sharedImportRatio = union.size > 0 ? intersection.size / union.size : 0;
    }
 
    const rawFragmentation = calculateFragmentation(files, domain, {
      ...options,
      sharedImportRatio,
    });
 
    // Average cohesion and adjusted fragmentation for the cluster
    let totalCohesion = 0;
    let totalAdjustedFragmentation = 0;
 
    files.forEach((f) => {
      const node = graph.nodes.get(f);
      Eif (node) {
        const cohesion = calculateEnhancedCohesion(node.exports);
        totalCohesion += cohesion;
 
        const classification = classifyFile(node, cohesion);
        totalAdjustedFragmentation += adjustFragmentationForClassification(
          rawFragmentation,
          classification
        );
      }
    });
 
    const avgCohesion = totalCohesion / files.length;
    const fragmentationScore = totalAdjustedFragmentation / files.length;
 
    clusters.push({
      domain,
      files,
      totalTokens,
      fragmentationScore,
      avgCohesion,
      suggestedStructure: generateSuggestedStructure(
        files,
        totalTokens,
        fragmentationScore
      ),
    });
  }
 
  return clusters;
}