code-gauge 10.0.0 → 10.2.0

package/dist/crossFileDuplication.cjs

@@ -1,2 +1,2 @@
-"use strict";const e=require("./duplication.cjs"),t=require("./crossFileNearMiss.cjs"),n=require("./duplicateSelection.cjs");function r(t,r){let{minTokens:o,maxGapTokens:d,minSimilarityPercent:f}=e.resolveDuplicationOptions(r),p=t.flatMap(({file:e,candidates:t},n)=>t.map(t=>({...t,regionBucket:n,file:e})));for(let e of a(t,o))p.push(e);let m=n.selectMaximalGroups(p,s,(e,t)=>e.regionBucket-t.regionBucket||e.startIndex-t.startIndex),h=c(t,d),g=l([...m.values()],h,d);for(let e of i(t,g,h,o,f))g.push(e);return u(g,t,h)}function i(e,n,r,i,a){let o=e.map(()=>[]);for(let{fileIndex:e,startTokenIndex:t,endTokenIndex:i}of n.flat()){let n=r[e]??0;o[e]?.push({startTokenIndex:t-n,endTokenIndex:i-n})}return t.collectCrossFileNearMissGroups(e,o,i,a).map(t=>t.map(t=>{let n=r[t.fileIndex]??0;return{...t,file:e[t.fileIndex]?.file??``,segments:t.segments.map(e=>({startTokenIndex:e.startTokenIndex+n,endTokenIndex:e.endTokenIndex+n})),startTokenIndex:t.startTokenIndex+n,endTokenIndex:t.endTokenIndex+n}}))}function a(t,n){let r=[],i=[];for(let[n,{tokens:a,containerStatements:o}]of t.entries())a&&o&&(r.push(n),i.push({tokens:a,literalCountPrefix:e.buildLiteralCountPrefix(a),containers:o}));return i.length<2?[]:e.collectSequenceWindowCandidates(i,n,!0).flatMap(({candidate:e,contextIndex:n})=>{let i=r[n],a=i===void 0?void 0:t[i];return i===void 0||a===void 0?[]:[{...e,regionBucket:i,file:a.file}]})}function o(e){return new Set(e.map(e=>e.file)).size>=2}function s(e){return e.length>=2&&new Set(e.map(e=>e.regionBucket)).size>=2}function c(e,t){let n=[],r=0;for(let{tokens:i,candidates:a}of e){n.push(r);let e=i?.length??0;if(!i)for(let t of a)e=Math.max(e,t.endTokenIndex);r+=e+t+1}return n}function l(t,n,r){let i=t.map(e=>e.map(e=>{let t=e.startTokenIndex+(n[e.regionBucket]??0),r=e.endTokenIndex+(n[e.regionBucket]??0);return{file:e.file,fileIndex:e.regionBucket,spanCountedElsewhere:e.nestedInLargerGroup,nestedInLargerGroup:e.nestedInLargerGroup,segments:[{startTokenIndex:t,endTokenIndex:r}],tokenCount:e.tokenCount,startTokenIndex:t,endTokenIndex:r,startIndex:e.startIndex,endIndex:e.endIndex,startLine:e.startLine,endLine:e.endLine}}).toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex));return e.mergeAdjacentGroups(i,r,o)}function u(t,n,r){let i=[],a=/* @__PURE__ */ new Map,o=new Map(n.map((e,t)=>[e.file,{tokens:e.tokens,codeLineNumbers:e.codeLineNumbers,offset:r[t]??0}])),s=/* @__PURE__ */ new Map,c=0;for(let n of t){c+=e.countRedundantFragments(n);for(let e of n)d(e,o,s);let t=n.map(({file:e,startLine:t,endLine:n})=>({file:e,startLine:t,endLine:n})).toSorted((e,t)=>e.file.localeCompare(t.file)||e.startLine-t.startLine),r=[...new Set(t.map(({file:e})=>e))];for(let e of r)a.set(e,(a.get(e)??0)+1);i.push({files:r,occurrences:t,tokenCount:Math.min(...n.map(({tokenCount:e})=>e))})}return i.sort((e,t)=>t.tokenCount-e.tokenCount||(e.occurrences[0]?.file??``).localeCompare(t.occurrences[0]?.file??``)||(e.occurrences[0]?.startLine??0)-(t.occurrences[0]?.startLine??0)),{duplicateBlockCount:c,duplicateBlockGroupCountByFile:Object.fromEntries(a),duplicateLineNumbersByFile:Object.fromEntries([...s].map(([e,t])=>[e,[...t].toSorted((e,t)=>e-t)])),groups:i}}function d(t,n,r){let i=n.get(t.file);if(!i?.tokens)return;let a=r.get(t.file);a||(a=/* @__PURE__ */ new Set,r.set(t.file,a));for(let n of t.segments)e.collectSegmentLines({startTokenIndex:n.startTokenIndex-i.offset,endTokenIndex:n.endTokenIndex-i.offset},i.tokens,i.codeLineNumbers,a)}exports.measureCrossFileDuplication=r;
+"use strict";const e=require("./duplication.cjs"),t=require("./crossFileNearMiss.cjs"),n=require("./duplicateSelection.cjs");function r(t,r){let{minTokens:a,maxGapTokens:s,minSimilarityPercent:f}=e.resolveDuplicationOptions(r),p=t.flatMap(({file:e,candidates:t},n)=>t.map(t=>({...t,regionBucket:n,file:e})));for(let e of o(t,a))p.push(e);let m=n.selectMaximalGroups(p,c,(e,t)=>e.regionBucket-t.regionBucket||e.startIndex-t.startIndex),h=l(t,s),g=u([...m.values()],h,s);for(let e of i(t,g,h,a,f))g.push(e);return d(g,t,h)}function i(e,n,r,i,o){let s=e.map(()=>[]);for(let{fileIndex:e,startTokenIndex:t,endTokenIndex:i}of n.flat()){let n=r[e]??0;s[e]?.push({startTokenIndex:t-n,endTokenIndex:i-n})}return t.collectCrossFileNearMissGroups(e,s,i,o).map(t=>t.map(t=>{let n=r[t.fileIndex]??0;return{...t,file:e[t.fileIndex]?.file??``,segments:t.segments.map(e=>a(e,n)),matchedRuns:t.matchedRuns?.map(e=>a(e,n)),startTokenIndex:t.startTokenIndex+n,endTokenIndex:t.endTokenIndex+n}}))}function a({startTokenIndex:e,endTokenIndex:t},n){return{startTokenIndex:e+n,endTokenIndex:t+n}}function o(t,n){let r=[],i=[];for(let[n,{tokens:a,containerStatements:o}]of t.entries())a&&o&&(r.push(n),i.push({tokens:a,literalCountPrefix:e.buildLiteralCountPrefix(a),containers:o}));return i.length<2?[]:e.collectSequenceWindowCandidates(i,n,!0).flatMap(({candidate:e,contextIndex:n})=>{let i=r[n],a=i===void 0?void 0:t[i];return i===void 0||a===void 0?[]:[{...e,regionBucket:i,file:a.file}]})}function s(e){return new Set(e.map(e=>e.file)).size>=2}function c(e){return e.length>=2&&new Set(e.map(e=>e.regionBucket)).size>=2}function l(e,t){let n=[],r=0;for(let{tokens:i,candidates:a}of e){n.push(r);let e=i?.length??0;if(!i)for(let t of a)e=Math.max(e,t.endTokenIndex);r+=e+t+1}return n}function u(t,n,r){let i=t.map(e=>e.map(e=>{let t=e.startTokenIndex+(n[e.regionBucket]??0),r=e.endTokenIndex+(n[e.regionBucket]??0);return{file:e.file,fileIndex:e.regionBucket,spanCountedElsewhere:e.nestedInLargerGroup,nestedInLargerGroup:e.nestedInLargerGroup,segments:[{startTokenIndex:t,endTokenIndex:r}],tokenCount:e.tokenCount,startTokenIndex:t,endTokenIndex:r,startIndex:e.startIndex,endIndex:e.endIndex,startLine:e.startLine,endLine:e.endLine}}).toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex));return e.mergeAdjacentGroups(i,r,s)}function d(t,n,r){let i=[],a=/* @__PURE__ */ new Map,o=new Map(n.map((e,t)=>[e.file,{tokens:e.tokens,codeLineNumbers:e.codeLineNumbers,offset:r[t]??0}])),s=/* @__PURE__ */ new Map,c=0;for(let n of t){c+=e.countRedundantFragments(n);let t=n.map(e=>{let{file:t,startLine:n,endLine:r,tokenCount:i}=e,a=f(e,o,s);return{file:t,startLine:n,endLine:r,...a&&{lineNumbers:a},tokenCount:i}}).toSorted((e,t)=>e.file.localeCompare(t.file)||e.startLine-t.startLine),r=[...new Set(t.map(({file:e})=>e))];for(let e of r)a.set(e,(a.get(e)??0)+1);i.push({files:r,occurrences:t,tokenCount:Math.min(...n.map(({tokenCount:e})=>e))})}return i.sort((e,t)=>t.tokenCount-e.tokenCount||(e.occurrences[0]?.file??``).localeCompare(t.occurrences[0]?.file??``)||(e.occurrences[0]?.startLine??0)-(t.occurrences[0]?.startLine??0)),{duplicateBlockCount:c,duplicateBlockGroupCountByFile:Object.fromEntries(a),duplicateLineNumbersByFile:Object.fromEntries([...s].map(([e,t])=>[e,[...t].toSorted((e,t)=>e-t)])),groups:i}}function f(t,n,r){let i=n.get(t.file);if(!i?.tokens)return;let o=r.get(t.file);o||(o=/* @__PURE__ */ new Set,r.set(t.file,o));let s=e.collectMatchedLines(t.segments.map(e=>a(e,-i.offset)),t.matchedRuns?.map(e=>a(e,-i.offset)),i.tokens,i.codeLineNumbers);for(let e of s)o.add(e);return[...s].toSorted((e,t)=>e-t)}exports.measureCrossFileDuplication=r;
 //# sourceMappingURL=crossFileDuplication.cjs.map

package/dist/crossFileDuplication.cjs.map

@@ -1 +1 @@
-{"version":3,"file":"crossFileDuplication.cjs","names":["resolveDuplicationOptions","selectMaximalGroups","collectCrossFileNearMissGroups","buildLiteralCountPrefix","collectSequenceWindowCandidates","mergeAdjacentGroups","countRedundantFragments"],"sources":["../src/crossFileDuplication.ts"],"sourcesContent":["import { collectCrossFileNearMissGroups } from './crossFileNearMiss.js';\nimport { selectMaximalGroups, type SelectableRegion } from './duplicateSelection.js';\nimport {\n  buildLiteralCountPrefix,\n  collectSegmentLines,\n  collectSequenceWindowCandidates,\n  countRedundantFragments,\n  mergeAdjacentGroups,\n  resolveDuplicationOptions,\n  type CountedOccurrence,\n  type CrossFileDuplicateCandidate,\n  type CrossFileDuplicationFileData,\n  type SequenceWindowContext,\n} from './duplication.js';\nimport type { DuplicationOptions } from './types.js';\n\nexport interface CrossFileDuplicationSourceFile extends Partial<CrossFileDuplicationFileData> {\n  file: string;\n  candidates: CrossFileDuplicateCandidate[];\n}\n\nexport interface CrossFileDuplicateOccurrence {\n  endLine: number;\n  file: string;\n  startLine: number;\n}\n\nexport interface CrossFileDuplicateBlockGroup {\n  files: string[];\n  occurrences: CrossFileDuplicateOccurrence[];\n  /**\n   * Token count of the smallest occurrence. For exact and gapped groups it is the matched token\n   * count every occurrence shares (gaps are not counted); for near-miss (Type-3) groups it is the\n   * token count of the smallest whole block or set of matched cores, edited tokens included.\n   */\n  tokenCount: number;\n}\n\nexport interface CrossFileDuplicationMetrics {\n  /** Number of redundant copies across all groups, counted per matched fragment like within-file. */\n  duplicateBlockCount: number;\n  /** Groups the file participates in, keyed by the file name passed in. */\n  duplicateBlockGroupCountByFile: Record<string, number>;\n  /**\n   * Per file, the 1-based code lines covered by the tokens of its cross-file occurrences, sorted\n   * ascending: the matched tokens of exact and gapped occurrences, and every token of a near-miss\n   * block or its matched cores, edited ones included (like within-file near-miss coverage). The unmatched gap of a merged\n   * clone and comment/blank lines inside an occurrence's bounding range are excluded (blank rows\n   * inside multi-row tokens only when the file supplied codeLineNumbers). A file that supplied\n   * only candidates (no `tokens`) has no entry — without its token stream the covered lines are\n   * unknowable, and an approximate bounding range would break this field's exactness.\n   */\n  duplicateLineNumbersByFile: Record<string, number[]>;\n  groups: CrossFileDuplicateBlockGroup[];\n}\n\ninterface SelectableCandidate extends CrossFileDuplicateCandidate, SelectableRegion {\n  regionBucket: number;\n  file: string;\n}\n\n/** A cross-file occurrence: a within-file occurrence in the project-wide token index space. */\ninterface CrossFileOccurrence extends CountedOccurrence {\n  file: string;\n  fileIndex: number;\n}\n\n/**\n * Detects code regions duplicated across files. Per-file candidates (whole block subtrees and full\n * container runs, fingerprinted with the same normalization as within-file duplication) are joined\n * by a project-level window index over per-statement fingerprint sequences (CPD-style), so a\n * copy-pasted partial statement run embedded in different surrounding code is matched even though\n * no single file can know it repeats elsewhere. Candidates are grouped by fingerprint, and only\n * maximal, non-overlapping regions whose group spans at least two files are counted. Groups that\n * shrink to a single file during selection are shed — a within-file repeat is already reported by\n * that file's own duplication metrics. A copy nested inside a larger group's region (two files share\n * a whole function, a third file only a block of it) is reported with its group, so the third\n * file's copy still shows what it duplicates. Groups separated by a small token gap within each file then\n * merge into gapped (Type-3) clone groups under `maxGapTokens`, exactly like within-file merging.\n * Finally, blocks of files that supplied `nearMissBlocks` are compared across files for near-miss\n * (Type-3) clones under `minSimilarityPercent` (see crossFileNearMiss.ts).\n */\nexport function measureCrossFileDuplication(\n  files: CrossFileDuplicationSourceFile[],\n  options?: DuplicationOptions\n): CrossFileDuplicationMetrics {\n  const { minTokens, maxGapTokens, minSimilarityPercent } = resolveDuplicationOptions(options);\n  const candidates: SelectableCandidate[] = files.flatMap(({ file, candidates }, fileIndex) =>\n    candidates.map((candidate) => ({ ...candidate, regionBucket: fileIndex, file }))\n  );\n  // Pushed one by one: spreading the project-scale window-candidate array as call arguments\n  // overflows V8's argument limit (~124k) and crashes on Node, though Bun/JSC tolerates it.\n  for (const candidate of collectWindowCandidates(files, minTokens)) {\n    candidates.push(candidate);\n  }\n  const counted = selectMaximalGroups(\n    candidates,\n    spansMultipleFiles,\n    // File index and position break coverage ties deterministically.\n    (left, right) => left.regionBucket - right.regionBucket || left.startIndex - right.startIndex\n  );\n  const tokenOffsets = computeTokenOffsets(files, maxGapTokens);\n  const groups = mergeGapAdjacentGroups([...counted.values()], tokenOffsets, maxGapTokens);\n  for (const group of collectNearMissGroups(files, groups, tokenOffsets, minTokens, minSimilarityPercent)) {\n    groups.push(group);\n  }\n  return summarize(groups, files, tokenOffsets);\n}\n\n/** Near-miss groups among the blocks the exact groups leave unreported, in the project token space. */\nfunction collectNearMissGroups(\n  files: CrossFileDuplicationSourceFile[],\n  exactGroups: CrossFileOccurrence[][],\n  tokenOffsets: number[],\n  minTokens: number,\n  minSimilarityPercent: number\n): CrossFileOccurrence[][] {\n  const reportedSpansByFile: { startTokenIndex: number; endTokenIndex: number }[][] = files.map(() => []);\n  for (const { fileIndex, startTokenIndex, endTokenIndex } of exactGroups.flat()) {\n    const offset = tokenOffsets[fileIndex] ?? 0;\n    reportedSpansByFile[fileIndex]?.push({\n      startTokenIndex: startTokenIndex - offset,\n      endTokenIndex: endTokenIndex - offset,\n    });\n  }\n  return collectCrossFileNearMissGroups(files, reportedSpansByFile, minTokens, minSimilarityPercent).map((group) =>\n    group.map((occurrence) => {\n      const offset = tokenOffsets[occurrence.fileIndex] ?? 0;\n      return {\n        ...occurrence,\n        file: files[occurrence.fileIndex]?.file ?? '',\n        segments: occurrence.segments.map((segment) => ({\n          startTokenIndex: segment.startTokenIndex + offset,\n          endTokenIndex: segment.endTokenIndex + offset,\n        })),\n        startTokenIndex: occurrence.startTokenIndex + offset,\n        endTokenIndex: occurrence.endTokenIndex + offset,\n      };\n    })\n  );\n}\n\n/** Repeated sub-windows of sibling statements matched across the whole project's files. */\nfunction collectWindowCandidates(files: CrossFileDuplicationSourceFile[], minTokens: number): SelectableCandidate[] {\n  const fileIndexByContext: number[] = [];\n  const contexts: SequenceWindowContext[] = [];\n  for (const [fileIndex, { tokens, containerStatements }] of files.entries()) {\n    if (tokens && containerStatements) {\n      fileIndexByContext.push(fileIndex);\n      contexts.push({ tokens, literalCountPrefix: buildLiteralCountPrefix(tokens), containers: containerStatements });\n    }\n  }\n  if (contexts.length < 2) {\n    return [];\n  }\n  return collectSequenceWindowCandidates(contexts, minTokens, true).flatMap(({ candidate, contextIndex }) => {\n    const fileIndex = fileIndexByContext[contextIndex];\n    const file = fileIndex === undefined ? undefined : files[fileIndex];\n    return fileIndex === undefined || file === undefined\n      ? []\n      : [{ ...candidate, regionBucket: fileIndex, file: file.file }];\n  });\n}\n\n/** A merged group is reported only while it still covers more than one file. */\nfunction spansMultipleFilesAfterMerge(group: CrossFileOccurrence[]): boolean {\n  return new Set(group.map((occurrence) => occurrence.file)).size >= 2;\n}\n\nfunction spansMultipleFiles(group: SelectableCandidate[]): boolean {\n  return group.length >= 2 && new Set(group.map((candidate) => candidate.regionBucket)).size >= 2;\n}\n\n/**\n * Per-file token offsets that map every file into one project-wide token index space: each file's\n * tokens are offset by more than `maxGapTokens` past the previous file's, so occurrences in\n * different files are never gap-adjacent and merged pairs always stay within one file.\n */\nfunction computeTokenOffsets(files: CrossFileDuplicationSourceFile[], maxGapTokens: number): number[] {\n  const tokenOffsets: number[] = [];\n  let offset = 0;\n  for (const { tokens, candidates } of files) {\n    tokenOffsets.push(offset);\n    // Accumulated in a loop: spreading a project-scale candidate array as call arguments would\n    // overflow V8's argument limit (~124k) and crash on Node.\n    let tokenCount = tokens?.length ?? 0;\n    if (!tokens) {\n      for (const candidate of candidates) {\n        tokenCount = Math.max(tokenCount, candidate.endTokenIndex);\n      }\n    }\n    offset += tokenCount + maxGapTokens + 1;\n  }\n  return tokenOffsets;\n}\n\n/** Reuses the within-file gapped (Type-3) merging in the project-wide token index space. */\nfunction mergeGapAdjacentGroups(\n  groups: SelectableCandidate[][],\n  tokenOffsets: number[],\n  maxGapTokens: number\n): CrossFileOccurrence[][] {\n  const occurrenceGroups = groups.map((group) =>\n    group\n      .map((candidate): CrossFileOccurrence => {\n        const start = candidate.startTokenIndex + (tokenOffsets[candidate.regionBucket] ?? 0);\n        const end = candidate.endTokenIndex + (tokenOffsets[candidate.regionBucket] ?? 0);\n        return {\n          file: candidate.file,\n          fileIndex: candidate.regionBucket,\n          spanCountedElsewhere: candidate.nestedInLargerGroup,\n          nestedInLargerGroup: candidate.nestedInLargerGroup,\n          segments: [{ startTokenIndex: start, endTokenIndex: end }],\n          tokenCount: candidate.tokenCount,\n          startTokenIndex: start,\n          endTokenIndex: end,\n          startIndex: candidate.startIndex,\n          endIndex: candidate.endIndex,\n          startLine: candidate.startLine,\n          endLine: candidate.endLine,\n        };\n      })\n      .toSorted((left, right) => left.startTokenIndex - right.startTokenIndex)\n  );\n  return mergeAdjacentGroups(occurrenceGroups, maxGapTokens, spansMultipleFilesAfterMerge);\n}\n\nfunction summarize(\n  groups: CrossFileOccurrence[][],\n  files: CrossFileDuplicationSourceFile[],\n  tokenOffsets: number[]\n): CrossFileDuplicationMetrics {\n  const reported: CrossFileDuplicateBlockGroup[] = [];\n  // Accumulated in Maps: file names are arbitrary strings, and a plain object would read\n  // inherited properties for names like \"constructor\".\n  const groupCountByFile = new Map<string, number>();\n  const fileDataByName = new Map(\n    files.map((file, index) => [\n      file.file,\n      { tokens: file.tokens, codeLineNumbers: file.codeLineNumbers, offset: tokenOffsets[index] ?? 0 },\n    ])\n  );\n  const lineNumbersByFile = new Map<string, Set<number>>();\n  let duplicateBlockCount = 0;\n  for (const group of groups) {\n    // Mirrors within-file counting: each redundant occurrence contributes one count per matched\n    // fragment, gapped merging consolidates the grouping without halving the count, and spans a\n    // partial merge shares between a retained group and the merged group count once.\n    duplicateBlockCount += countRedundantFragments(group);\n    for (const occurrence of group) {\n      collectOccurrenceLines(occurrence, fileDataByName, lineNumbersByFile);\n    }\n    const occurrences = group\n      .map(({ file, startLine, endLine }) => ({ file, startLine, endLine }))\n      .toSorted((left, right) => left.file.localeCompare(right.file) || left.startLine - right.startLine);\n    const files = [...new Set(occurrences.map(({ file }) => file))];\n    for (const file of files) {\n      groupCountByFile.set(file, (groupCountByFile.get(file) ?? 0) + 1);\n    }\n    reported.push({ files, occurrences, tokenCount: Math.min(...group.map(({ tokenCount }) => tokenCount)) });\n  }\n  reported.sort(\n    (left, right) =>\n      right.tokenCount - left.tokenCount ||\n      (left.occurrences[0]?.file ?? '').localeCompare(right.occurrences[0]?.file ?? '') ||\n      (left.occurrences[0]?.startLine ?? 0) - (right.occurrences[0]?.startLine ?? 0)\n  );\n  return {\n    duplicateBlockCount,\n    duplicateBlockGroupCountByFile: Object.fromEntries(groupCountByFile),\n    duplicateLineNumbersByFile: Object.fromEntries(\n      [...lineNumbersByFile].map(([file, lines]) => [file, [...lines].toSorted((left, right) => left - right)])\n    ),\n    groups: reported,\n  };\n}\n\n/**\n * Adds the code lines an occurrence's segment tokens cover (matched tokens of an exact or gapped\n * occurrence, the whole block or matched cores of a near-miss one) to its file's line set, mapping the\n * project-wide token segments back into the file's own token stream. A file that supplied only\n * candidates (no token stream) is skipped rather than approximated from the bounding line range,\n * which would include gap and comment/blank lines and break the field's exactness contract.\n */\nfunction collectOccurrenceLines(\n  occurrence: CrossFileOccurrence,\n  fileDataByName: Map<\n    string,\n    { tokens?: CrossFileDuplicationSourceFile['tokens']; codeLineNumbers?: Set<number>; offset: number }\n  >,\n  lineNumbersByFile: Map<string, Set<number>>\n): void {\n  const fileData = fileDataByName.get(occurrence.file);\n  if (!fileData?.tokens) {\n    return;\n  }\n  let lines = lineNumbersByFile.get(occurrence.file);\n  if (!lines) {\n    lines = new Set();\n    lineNumbersByFile.set(occurrence.file, lines);\n  }\n  for (const segment of occurrence.segments) {\n    collectSegmentLines(\n      {\n        startTokenIndex: segment.startTokenIndex - fileData.offset,\n        endTokenIndex: segment.endTokenIndex - fileData.offset,\n      },\n      fileData.tokens,\n      fileData.codeLineNumbers,\n      lines\n    );\n  }\n}\n"],"mappings":"6HAkFA,SAAgB,EACd,EACA,EAC6B,CAC7B,GAAM,CAAE,YAAW,eAAc,wBAAyBA,EAAAA,0BAA0B,CAAO,EACrF,EAAoC,EAAM,SAAS,CAAE,OAAM,cAAc,IAC7E,EAAW,IAAK,IAAe,CAAE,GAAG,EAAW,aAAc,EAAW,MAAK,EAAE,CACjF,EAGA,IAAK,IAAM,KAAa,EAAwB,EAAO,CAAS,EAC9D,EAAW,KAAK,CAAS,EAE3B,IAAM,EAAUC,EAAAA,oBACd,EACA,GAEC,EAAM,IAAU,EAAK,aAAe,EAAM,cAAgB,EAAK,WAAa,EAAM,UACrF,EACM,EAAe,EAAoB,EAAO,CAAY,EACtD,EAAS,EAAuB,CAAC,GAAG,EAAQ,OAAO,CAAC,EAAG,EAAc,CAAY,EACvF,IAAK,IAAM,KAAS,EAAsB,EAAO,EAAQ,EAAc,EAAW,CAAoB,EACpG,EAAO,KAAK,CAAK,EAEnB,OAAO,EAAU,EAAQ,EAAO,CAAY,CAC9C,CAGA,SAAS,EACP,EACA,EACA,EACA,EACA,EACyB,CACzB,IAAM,EAA8E,EAAM,QAAU,CAAC,CAAC,EACtG,IAAK,GAAM,CAAE,YAAW,kBAAiB,mBAAmB,EAAY,KAAK,EAAG,CAC9E,IAAM,EAAS,EAAa,IAAc,EAC1C,EAAoB,EAAU,EAAE,KAAK,CACnC,gBAAiB,EAAkB,EACnC,cAAe,EAAgB,CACjC,CAAC,CACH,CACA,OAAOC,EAAAA,+BAA+B,EAAO,EAAqB,EAAW,CAAoB,CAAC,CAAC,IAAK,GACtG,EAAM,IAAK,GAAe,CACxB,IAAM,EAAS,EAAa,EAAW,YAAc,EACrD,MAAO,CACL,GAAG,EACH,KAAM,EAAM,EAAW,UAAU,EAAE,MAAQ,GAC3C,SAAU,EAAW,SAAS,IAAK,IAAa,CAC9C,gBAAiB,EAAQ,gBAAkB,EAC3C,cAAe,EAAQ,cAAgB,CACzC,EAAE,EACF,gBAAiB,EAAW,gBAAkB,EAC9C,cAAe,EAAW,cAAgB,CAC5C,CACF,CAAC,CACH,CACF,CAGA,SAAS,EAAwB,EAAyC,EAA0C,CAClH,IAAM,EAA+B,CAAC,EAChC,EAAoC,CAAC,EAC3C,IAAK,GAAM,CAAC,EAAW,CAAE,SAAQ,0BAA0B,EAAM,QAAQ,EACnE,GAAU,IACZ,EAAmB,KAAK,CAAS,EACjC,EAAS,KAAK,CAAE,SAAQ,mBAAoBC,EAAAA,wBAAwB,CAAM,EAAG,WAAY,CAAoB,CAAC,GAMlH,OAHI,EAAS,OAAS,EACb,CAAC,EAEHC,EAAAA,gCAAgC,EAAU,EAAW,EAAI,CAAC,CAAC,SAAS,CAAE,YAAW,kBAAmB,CACzG,IAAM,EAAY,EAAmB,GAC/B,EAAO,IAAc,IAAA,GAAY,IAAA,GAAY,EAAM,GACzD,OAAO,IAAc,IAAA,IAAa,IAAS,IAAA,GACvC,CAAC,EACD,CAAC,CAAE,GAAG,EAAW,aAAc,EAAW,KAAM,EAAK,IAAK,CAAC,CACjE,CAAC,CACH,CAGA,SAAS,EAA6B,EAAuC,CAC3E,OAAO,IAAI,IAAI,EAAM,IAAK,GAAe,EAAW,IAAI,CAAC,CAAC,CAAC,MAAQ,CACrE,CAEA,SAAS,EAAmB,EAAuC,CACjE,OAAO,EAAM,QAAU,GAAK,IAAI,IAAI,EAAM,IAAK,GAAc,EAAU,YAAY,CAAC,CAAC,CAAC,MAAQ,CAChG,CAOA,SAAS,EAAoB,EAAyC,EAAgC,CACpG,IAAM,EAAyB,CAAC,EAC5B,EAAS,EACb,IAAK,GAAM,CAAE,SAAQ,gBAAgB,EAAO,CAC1C,EAAa,KAAK,CAAM,EAGxB,IAAI,EAAa,GAAQ,QAAU,EACnC,GAAI,CAAC,EACH,IAAK,IAAM,KAAa,EACtB,EAAa,KAAK,IAAI,EAAY,EAAU,aAAa,EAG7D,GAAU,EAAa,EAAe,CACxC,CACA,OAAO,CACT,CAGA,SAAS,EACP,EACA,EACA,EACyB,CACzB,IAAM,EAAmB,EAAO,IAAK,GACnC,EACG,IAAK,GAAmC,CACvC,IAAM,EAAQ,EAAU,iBAAmB,EAAa,EAAU,eAAiB,GAC7E,EAAM,EAAU,eAAiB,EAAa,EAAU,eAAiB,GAC/E,MAAO,CACL,KAAM,EAAU,KAChB,UAAW,EAAU,aACrB,qBAAsB,EAAU,oBAChC,oBAAqB,EAAU,oBAC/B,SAAU,CAAC,CAAE,gBAAiB,EAAO,cAAe,CAAI,CAAC,EACzD,WAAY,EAAU,WACtB,gBAAiB,EACjB,cAAe,EACf,WAAY,EAAU,WACtB,SAAU,EAAU,SACpB,UAAW,EAAU,UACrB,QAAS,EAAU,OACrB,CACF,CAAC,CAAC,CACD,UAAU,EAAM,IAAU,EAAK,gBAAkB,EAAM,eAAe,CAC3E,EACA,OAAOC,EAAAA,oBAAoB,EAAkB,EAAc,CAA4B,CACzF,CAEA,SAAS,EACP,EACA,EACA,EAC6B,CAC7B,IAAM,EAA2C,CAAC,EAG5C,kBAAmB,IAAI,IACvB,EAAiB,IAAI,IACzB,EAAM,KAAK,EAAM,IAAU,CACzB,EAAK,KACL,CAAE,OAAQ,EAAK,OAAQ,gBAAiB,EAAK,gBAAiB,OAAQ,EAAa,IAAU,CAAE,CACjG,CAAC,CACH,EACM,kBAAoB,IAAI,IAC1B,EAAsB,EAC1B,IAAK,IAAM,KAAS,EAAQ,CAI1B,GAAuBC,EAAAA,wBAAwB,CAAK,EACpD,IAAK,IAAM,KAAc,EACvB,EAAuB,EAAY,EAAgB,CAAiB,EAEtE,IAAM,EAAc,EACjB,KAAK,CAAE,OAAM,YAAW,cAAe,CAAE,OAAM,YAAW,SAAQ,EAAE,CAAC,CACrE,UAAU,EAAM,IAAU,EAAK,KAAK,cAAc,EAAM,IAAI,GAAK,EAAK,UAAY,EAAM,SAAS,EAC9F,EAAQ,CAAC,GAAG,IAAI,IAAI,EAAY,KAAK,CAAE,UAAW,CAAI,CAAC,CAAC,EAC9D,IAAK,IAAM,KAAQ,EACjB,EAAiB,IAAI,GAAO,EAAiB,IAAI,CAAI,GAAK,GAAK,CAAC,EAElE,EAAS,KAAK,CAAE,QAAO,cAAa,WAAY,KAAK,IAAI,GAAG,EAAM,KAAK,CAAE,gBAAiB,CAAU,CAAC,CAAE,CAAC,CAC1G,CAOA,OANA,EAAS,MACN,EAAM,IACL,EAAM,WAAa,EAAK,aACvB,EAAK,YAAY,EAAE,EAAE,MAAQ,GAAA,CAAI,cAAc,EAAM,YAAY,EAAE,EAAE,MAAQ,EAAE,IAC/E,EAAK,YAAY,EAAE,EAAE,WAAa,IAAM,EAAM,YAAY,EAAE,EAAE,WAAa,EAChF,EACO,CACL,sBACA,+BAAgC,OAAO,YAAY,CAAgB,EACnE,2BAA4B,OAAO,YACjC,CAAC,GAAG,CAAiB,CAAC,CAAC,KAAK,CAAC,EAAM,KAAW,CAAC,EAAM,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAAC,CAAC,CAC1G,EACA,OAAQ,CACV,CACF,CASA,SAAS,EACP,EACA,EAIA,EACM,CACN,IAAM,EAAW,EAAe,IAAI,EAAW,IAAI,EACnD,GAAI,CAAC,GAAU,OACb,OAEF,IAAI,EAAQ,EAAkB,IAAI,EAAW,IAAI,EAC5C,IACH,kBAAQ,IAAI,IACZ,EAAkB,IAAI,EAAW,KAAM,CAAK,GAE9C,IAAK,IAAM,KAAW,EAAW,SAC/B,EAAA,oBACE,CACE,gBAAiB,EAAQ,gBAAkB,EAAS,OACpD,cAAe,EAAQ,cAAgB,EAAS,MAClD,EACA,EAAS,OACT,EAAS,gBACT,CACF,CAEJ"}
+{"version":3,"file":"crossFileDuplication.cjs","names":["resolveDuplicationOptions","selectMaximalGroups","collectCrossFileNearMissGroups","buildLiteralCountPrefix","collectSequenceWindowCandidates","mergeAdjacentGroups","countRedundantFragments","collectMatchedLines"],"sources":["../src/crossFileDuplication.ts"],"sourcesContent":["import { collectCrossFileNearMissGroups } from './crossFileNearMiss.js';\nimport { selectMaximalGroups, type SelectableRegion } from './duplicateSelection.js';\nimport {\n  buildLiteralCountPrefix,\n  collectMatchedLines,\n  collectSequenceWindowCandidates,\n  countRedundantFragments,\n  mergeAdjacentGroups,\n  resolveDuplicationOptions,\n  type CountedOccurrence,\n  type CrossFileDuplicateCandidate,\n  type CrossFileDuplicationFileData,\n  type SequenceWindowContext,\n  type TokenSegment,\n} from './duplication.js';\nimport type { DuplicationOptions } from './types.js';\n\nexport interface CrossFileDuplicationSourceFile extends Partial<CrossFileDuplicationFileData> {\n  file: string;\n  candidates: CrossFileDuplicateCandidate[];\n}\n\nexport interface CrossFileDuplicateOccurrence {\n  endLine: number;\n  file: string;\n  /**\n   * The code lines carrying the occurrence's matched tokens (see\n   * DuplicateBlockOccurrence.lineNumbers). Absent for a file that supplied only candidates (no\n   * `tokens`), whose matched lines are unknowable.\n   */\n  lineNumbers?: number[];\n  startLine: number;\n  /**\n   * The tokens of its matched runs, or of a near-miss block or its matched cores with the edited\n   * tokens among them; the group's `tokenCount` is the smallest of its occurrences'.\n   */\n  tokenCount: number;\n}\n\nexport interface CrossFileDuplicateBlockGroup {\n  files: string[];\n  occurrences: CrossFileDuplicateOccurrence[];\n  /**\n   * Token count of the smallest occurrence. For exact and gapped groups it is the matched token\n   * count every occurrence shares (gaps are not counted); for near-miss (Type-3) groups it is the\n   * token count of the smallest whole block or set of matched cores, edited tokens included.\n   */\n  tokenCount: number;\n}\n\nexport interface CrossFileDuplicationMetrics {\n  /** Number of redundant copies across all groups, counted per matched fragment like within-file. */\n  duplicateBlockCount: number;\n  /** Groups the file participates in, keyed by the file name passed in. */\n  duplicateBlockGroupCountByFile: Record<string, number>;\n  /**\n   * Per file, the 1-based code lines covered by the tokens of its cross-file occurrences, sorted\n   * ascending: the matched tokens of exact and gapped occurrences, and the lines of a near-miss\n   * block or its matched cores more than half of whose tokens its partners match (like within-file\n   * near-miss coverage). The unmatched gap of a merged\n   * clone and comment/blank lines inside an occurrence's bounding range are excluded (blank rows\n   * inside multi-row tokens only when the file supplied codeLineNumbers). A file that supplied\n   * only candidates (no `tokens`) has no entry — without its token stream the covered lines are\n   * unknowable, and an approximate bounding range would break this field's exactness.\n   */\n  duplicateLineNumbersByFile: Record<string, number[]>;\n  groups: CrossFileDuplicateBlockGroup[];\n}\n\ninterface SelectableCandidate extends CrossFileDuplicateCandidate, SelectableRegion {\n  regionBucket: number;\n  file: string;\n}\n\n/** A cross-file occurrence: a within-file occurrence in the project-wide token index space. */\ninterface CrossFileOccurrence extends CountedOccurrence {\n  file: string;\n  fileIndex: number;\n}\n\n/**\n * Detects code regions duplicated across files. Per-file candidates (whole block subtrees and full\n * container runs, fingerprinted with the same normalization as within-file duplication) are joined\n * by a project-level window index over per-statement fingerprint sequences (CPD-style), so a\n * copy-pasted partial statement run embedded in different surrounding code is matched even though\n * no single file can know it repeats elsewhere. Candidates are grouped by fingerprint, and only\n * maximal, non-overlapping regions whose group spans at least two files are counted. Groups that\n * shrink to a single file during selection are shed — a within-file repeat is already reported by\n * that file's own duplication metrics. A copy nested inside a larger group's region (two files share\n * a whole function, a third file only a block of it) is reported with its group, so the third\n * file's copy still shows what it duplicates. Groups separated by a small token gap within each file then\n * merge into gapped (Type-3) clone groups under `maxGapTokens`, exactly like within-file merging.\n * Finally, blocks of files that supplied `nearMissBlocks` are compared across files for near-miss\n * (Type-3) clones under `minSimilarityPercent` (see crossFileNearMiss.ts).\n */\nexport function measureCrossFileDuplication(\n  files: CrossFileDuplicationSourceFile[],\n  options?: DuplicationOptions\n): CrossFileDuplicationMetrics {\n  const { minTokens, maxGapTokens, minSimilarityPercent } = resolveDuplicationOptions(options);\n  const candidates: SelectableCandidate[] = files.flatMap(({ file, candidates }, fileIndex) =>\n    candidates.map((candidate) => ({ ...candidate, regionBucket: fileIndex, file }))\n  );\n  // Pushed one by one: spreading the project-scale window-candidate array as call arguments\n  // overflows V8's argument limit (~124k) and crashes on Node, though Bun/JSC tolerates it.\n  for (const candidate of collectWindowCandidates(files, minTokens)) {\n    candidates.push(candidate);\n  }\n  const counted = selectMaximalGroups(\n    candidates,\n    spansMultipleFiles,\n    // File index and position break coverage ties deterministically.\n    (left, right) => left.regionBucket - right.regionBucket || left.startIndex - right.startIndex\n  );\n  const tokenOffsets = computeTokenOffsets(files, maxGapTokens);\n  const groups = mergeGapAdjacentGroups([...counted.values()], tokenOffsets, maxGapTokens);\n  for (const group of collectNearMissGroups(files, groups, tokenOffsets, minTokens, minSimilarityPercent)) {\n    groups.push(group);\n  }\n  return summarize(groups, files, tokenOffsets);\n}\n\n/** Near-miss groups among the blocks the exact groups leave unreported, in the project token space. */\nfunction collectNearMissGroups(\n  files: CrossFileDuplicationSourceFile[],\n  exactGroups: CrossFileOccurrence[][],\n  tokenOffsets: number[],\n  minTokens: number,\n  minSimilarityPercent: number\n): CrossFileOccurrence[][] {\n  const reportedSpansByFile: { startTokenIndex: number; endTokenIndex: number }[][] = files.map(() => []);\n  for (const { fileIndex, startTokenIndex, endTokenIndex } of exactGroups.flat()) {\n    const offset = tokenOffsets[fileIndex] ?? 0;\n    reportedSpansByFile[fileIndex]?.push({\n      startTokenIndex: startTokenIndex - offset,\n      endTokenIndex: endTokenIndex - offset,\n    });\n  }\n  return collectCrossFileNearMissGroups(files, reportedSpansByFile, minTokens, minSimilarityPercent).map((group) =>\n    group.map((occurrence) => {\n      const offset = tokenOffsets[occurrence.fileIndex] ?? 0;\n      return {\n        ...occurrence,\n        file: files[occurrence.fileIndex]?.file ?? '',\n        segments: occurrence.segments.map((segment) => shift(segment, offset)),\n        matchedRuns: occurrence.matchedRuns?.map((run) => shift(run, offset)),\n        startTokenIndex: occurrence.startTokenIndex + offset,\n        endTokenIndex: occurrence.endTokenIndex + offset,\n      };\n    })\n  );\n}\n\nfunction shift({ startTokenIndex, endTokenIndex }: TokenSegment, offset: number): TokenSegment {\n  return { startTokenIndex: startTokenIndex + offset, endTokenIndex: endTokenIndex + offset };\n}\n\n/** Repeated sub-windows of sibling statements matched across the whole project's files. */\nfunction collectWindowCandidates(files: CrossFileDuplicationSourceFile[], minTokens: number): SelectableCandidate[] {\n  const fileIndexByContext: number[] = [];\n  const contexts: SequenceWindowContext[] = [];\n  for (const [fileIndex, { tokens, containerStatements }] of files.entries()) {\n    if (tokens && containerStatements) {\n      fileIndexByContext.push(fileIndex);\n      contexts.push({ tokens, literalCountPrefix: buildLiteralCountPrefix(tokens), containers: containerStatements });\n    }\n  }\n  if (contexts.length < 2) {\n    return [];\n  }\n  return collectSequenceWindowCandidates(contexts, minTokens, true).flatMap(({ candidate, contextIndex }) => {\n    const fileIndex = fileIndexByContext[contextIndex];\n    const file = fileIndex === undefined ? undefined : files[fileIndex];\n    return fileIndex === undefined || file === undefined\n      ? []\n      : [{ ...candidate, regionBucket: fileIndex, file: file.file }];\n  });\n}\n\n/** A merged group is reported only while it still covers more than one file. */\nfunction spansMultipleFilesAfterMerge(group: CrossFileOccurrence[]): boolean {\n  return new Set(group.map((occurrence) => occurrence.file)).size >= 2;\n}\n\nfunction spansMultipleFiles(group: SelectableCandidate[]): boolean {\n  return group.length >= 2 && new Set(group.map((candidate) => candidate.regionBucket)).size >= 2;\n}\n\n/**\n * Per-file token offsets that map every file into one project-wide token index space: each file's\n * tokens are offset by more than `maxGapTokens` past the previous file's, so occurrences in\n * different files are never gap-adjacent and merged pairs always stay within one file.\n */\nfunction computeTokenOffsets(files: CrossFileDuplicationSourceFile[], maxGapTokens: number): number[] {\n  const tokenOffsets: number[] = [];\n  let offset = 0;\n  for (const { tokens, candidates } of files) {\n    tokenOffsets.push(offset);\n    // Accumulated in a loop: spreading a project-scale candidate array as call arguments would\n    // overflow V8's argument limit (~124k) and crash on Node.\n    let tokenCount = tokens?.length ?? 0;\n    if (!tokens) {\n      for (const candidate of candidates) {\n        tokenCount = Math.max(tokenCount, candidate.endTokenIndex);\n      }\n    }\n    offset += tokenCount + maxGapTokens + 1;\n  }\n  return tokenOffsets;\n}\n\n/** Reuses the within-file gapped (Type-3) merging in the project-wide token index space. */\nfunction mergeGapAdjacentGroups(\n  groups: SelectableCandidate[][],\n  tokenOffsets: number[],\n  maxGapTokens: number\n): CrossFileOccurrence[][] {\n  const occurrenceGroups = groups.map((group) =>\n    group\n      .map((candidate): CrossFileOccurrence => {\n        const start = candidate.startTokenIndex + (tokenOffsets[candidate.regionBucket] ?? 0);\n        const end = candidate.endTokenIndex + (tokenOffsets[candidate.regionBucket] ?? 0);\n        return {\n          file: candidate.file,\n          fileIndex: candidate.regionBucket,\n          spanCountedElsewhere: candidate.nestedInLargerGroup,\n          nestedInLargerGroup: candidate.nestedInLargerGroup,\n          segments: [{ startTokenIndex: start, endTokenIndex: end }],\n          tokenCount: candidate.tokenCount,\n          startTokenIndex: start,\n          endTokenIndex: end,\n          startIndex: candidate.startIndex,\n          endIndex: candidate.endIndex,\n          startLine: candidate.startLine,\n          endLine: candidate.endLine,\n        };\n      })\n      .toSorted((left, right) => left.startTokenIndex - right.startTokenIndex)\n  );\n  return mergeAdjacentGroups(occurrenceGroups, maxGapTokens, spansMultipleFilesAfterMerge);\n}\n\nfunction summarize(\n  groups: CrossFileOccurrence[][],\n  files: CrossFileDuplicationSourceFile[],\n  tokenOffsets: number[]\n): CrossFileDuplicationMetrics {\n  const reported: CrossFileDuplicateBlockGroup[] = [];\n  // Accumulated in Maps: file names are arbitrary strings, and a plain object would read\n  // inherited properties for names like \"constructor\".\n  const groupCountByFile = new Map<string, number>();\n  const fileDataByName = new Map(\n    files.map((file, index) => [\n      file.file,\n      { tokens: file.tokens, codeLineNumbers: file.codeLineNumbers, offset: tokenOffsets[index] ?? 0 },\n    ])\n  );\n  const lineNumbersByFile = new Map<string, Set<number>>();\n  let duplicateBlockCount = 0;\n  for (const group of groups) {\n    // Mirrors within-file counting: each redundant occurrence contributes one count per matched\n    // fragment, gapped merging consolidates the grouping without halving the count, and spans a\n    // partial merge shares between a retained group and the merged group count once.\n    duplicateBlockCount += countRedundantFragments(group);\n    const occurrences = group\n      .map((occurrence) => {\n        const { file, startLine, endLine, tokenCount } = occurrence;\n        const lineNumbers = collectOccurrenceLines(occurrence, fileDataByName, lineNumbersByFile);\n        return { file, startLine, endLine, ...(lineNumbers && { lineNumbers }), tokenCount };\n      })\n      .toSorted((left, right) => left.file.localeCompare(right.file) || left.startLine - right.startLine);\n    const files = [...new Set(occurrences.map(({ file }) => file))];\n    for (const file of files) {\n      groupCountByFile.set(file, (groupCountByFile.get(file) ?? 0) + 1);\n    }\n    reported.push({ files, occurrences, tokenCount: Math.min(...group.map(({ tokenCount }) => tokenCount)) });\n  }\n  reported.sort(\n    (left, right) =>\n      right.tokenCount - left.tokenCount ||\n      (left.occurrences[0]?.file ?? '').localeCompare(right.occurrences[0]?.file ?? '') ||\n      (left.occurrences[0]?.startLine ?? 0) - (right.occurrences[0]?.startLine ?? 0)\n  );\n  return {\n    duplicateBlockCount,\n    duplicateBlockGroupCountByFile: Object.fromEntries(groupCountByFile),\n    duplicateLineNumbersByFile: Object.fromEntries(\n      [...lineNumbersByFile].map(([file, lines]) => [file, [...lines].toSorted((left, right) => left - right)])\n    ),\n    groups: reported,\n  };\n}\n\n/**\n * Adds the code lines an occurrence's matched tokens cover (see collectMatchedLines) to its file's\n * line set, mapping the project-wide token segments back into the file's own token stream, and\n * returns them in order.\n * A file that supplied only candidates (no token stream) is skipped rather than approximated from\n * the bounding line range, which would include gap and comment/blank lines and break the field's\n * exactness contract.\n */\nfunction collectOccurrenceLines(\n  occurrence: CrossFileOccurrence,\n  fileDataByName: Map<\n    string,\n    { tokens?: CrossFileDuplicationSourceFile['tokens']; codeLineNumbers?: Set<number>; offset: number }\n  >,\n  lineNumbersByFile: Map<string, Set<number>>\n): number[] | undefined {\n  const fileData = fileDataByName.get(occurrence.file);\n  if (!fileData?.tokens) {\n    return undefined;\n  }\n  let fileLines = lineNumbersByFile.get(occurrence.file);\n  if (!fileLines) {\n    fileLines = new Set();\n    lineNumbersByFile.set(occurrence.file, fileLines);\n  }\n  const lines = collectMatchedLines(\n    occurrence.segments.map((segment) => shift(segment, -fileData.offset)),\n    occurrence.matchedRuns?.map((run) => shift(run, -fileData.offset)),\n    fileData.tokens,\n    fileData.codeLineNumbers\n  );\n  for (const line of lines) {\n    fileLines.add(line);\n  }\n  return [...lines].toSorted((left, right) => left - right);\n}\n"],"mappings":"6HA+FA,SAAgB,EACd,EACA,EAC6B,CAC7B,GAAM,CAAE,YAAW,eAAc,wBAAyBA,EAAAA,0BAA0B,CAAO,EACrF,EAAoC,EAAM,SAAS,CAAE,OAAM,cAAc,IAC7E,EAAW,IAAK,IAAe,CAAE,GAAG,EAAW,aAAc,EAAW,MAAK,EAAE,CACjF,EAGA,IAAK,IAAM,KAAa,EAAwB,EAAO,CAAS,EAC9D,EAAW,KAAK,CAAS,EAE3B,IAAM,EAAUC,EAAAA,oBACd,EACA,GAEC,EAAM,IAAU,EAAK,aAAe,EAAM,cAAgB,EAAK,WAAa,EAAM,UACrF,EACM,EAAe,EAAoB,EAAO,CAAY,EACtD,EAAS,EAAuB,CAAC,GAAG,EAAQ,OAAO,CAAC,EAAG,EAAc,CAAY,EACvF,IAAK,IAAM,KAAS,EAAsB,EAAO,EAAQ,EAAc,EAAW,CAAoB,EACpG,EAAO,KAAK,CAAK,EAEnB,OAAO,EAAU,EAAQ,EAAO,CAAY,CAC9C,CAGA,SAAS,EACP,EACA,EACA,EACA,EACA,EACyB,CACzB,IAAM,EAA8E,EAAM,QAAU,CAAC,CAAC,EACtG,IAAK,GAAM,CAAE,YAAW,kBAAiB,mBAAmB,EAAY,KAAK,EAAG,CAC9E,IAAM,EAAS,EAAa,IAAc,EAC1C,EAAoB,EAAU,EAAE,KAAK,CACnC,gBAAiB,EAAkB,EACnC,cAAe,EAAgB,CACjC,CAAC,CACH,CACA,OAAOC,EAAAA,+BAA+B,EAAO,EAAqB,EAAW,CAAoB,CAAC,CAAC,IAAK,GACtG,EAAM,IAAK,GAAe,CACxB,IAAM,EAAS,EAAa,EAAW,YAAc,EACrD,MAAO,CACL,GAAG,EACH,KAAM,EAAM,EAAW,UAAU,EAAE,MAAQ,GAC3C,SAAU,EAAW,SAAS,IAAK,GAAY,EAAM,EAAS,CAAM,CAAC,EACrE,YAAa,EAAW,aAAa,IAAK,GAAQ,EAAM,EAAK,CAAM,CAAC,EACpE,gBAAiB,EAAW,gBAAkB,EAC9C,cAAe,EAAW,cAAgB,CAC5C,CACF,CAAC,CACH,CACF,CAEA,SAAS,EAAM,CAAE,kBAAiB,iBAA+B,EAA8B,CAC7F,MAAO,CAAE,gBAAiB,EAAkB,EAAQ,cAAe,EAAgB,CAAO,CAC5F,CAGA,SAAS,EAAwB,EAAyC,EAA0C,CAClH,IAAM,EAA+B,CAAC,EAChC,EAAoC,CAAC,EAC3C,IAAK,GAAM,CAAC,EAAW,CAAE,SAAQ,0BAA0B,EAAM,QAAQ,EACnE,GAAU,IACZ,EAAmB,KAAK,CAAS,EACjC,EAAS,KAAK,CAAE,SAAQ,mBAAoBC,EAAAA,wBAAwB,CAAM,EAAG,WAAY,CAAoB,CAAC,GAMlH,OAHI,EAAS,OAAS,EACb,CAAC,EAEHC,EAAAA,gCAAgC,EAAU,EAAW,EAAI,CAAC,CAAC,SAAS,CAAE,YAAW,kBAAmB,CACzG,IAAM,EAAY,EAAmB,GAC/B,EAAO,IAAc,IAAA,GAAY,IAAA,GAAY,EAAM,GACzD,OAAO,IAAc,IAAA,IAAa,IAAS,IAAA,GACvC,CAAC,EACD,CAAC,CAAE,GAAG,EAAW,aAAc,EAAW,KAAM,EAAK,IAAK,CAAC,CACjE,CAAC,CACH,CAGA,SAAS,EAA6B,EAAuC,CAC3E,OAAO,IAAI,IAAI,EAAM,IAAK,GAAe,EAAW,IAAI,CAAC,CAAC,CAAC,MAAQ,CACrE,CAEA,SAAS,EAAmB,EAAuC,CACjE,OAAO,EAAM,QAAU,GAAK,IAAI,IAAI,EAAM,IAAK,GAAc,EAAU,YAAY,CAAC,CAAC,CAAC,MAAQ,CAChG,CAOA,SAAS,EAAoB,EAAyC,EAAgC,CACpG,IAAM,EAAyB,CAAC,EAC5B,EAAS,EACb,IAAK,GAAM,CAAE,SAAQ,gBAAgB,EAAO,CAC1C,EAAa,KAAK,CAAM,EAGxB,IAAI,EAAa,GAAQ,QAAU,EACnC,GAAI,CAAC,EACH,IAAK,IAAM,KAAa,EACtB,EAAa,KAAK,IAAI,EAAY,EAAU,aAAa,EAG7D,GAAU,EAAa,EAAe,CACxC,CACA,OAAO,CACT,CAGA,SAAS,EACP,EACA,EACA,EACyB,CACzB,IAAM,EAAmB,EAAO,IAAK,GACnC,EACG,IAAK,GAAmC,CACvC,IAAM,EAAQ,EAAU,iBAAmB,EAAa,EAAU,eAAiB,GAC7E,EAAM,EAAU,eAAiB,EAAa,EAAU,eAAiB,GAC/E,MAAO,CACL,KAAM,EAAU,KAChB,UAAW,EAAU,aACrB,qBAAsB,EAAU,oBAChC,oBAAqB,EAAU,oBAC/B,SAAU,CAAC,CAAE,gBAAiB,EAAO,cAAe,CAAI,CAAC,EACzD,WAAY,EAAU,WACtB,gBAAiB,EACjB,cAAe,EACf,WAAY,EAAU,WACtB,SAAU,EAAU,SACpB,UAAW,EAAU,UACrB,QAAS,EAAU,OACrB,CACF,CAAC,CAAC,CACD,UAAU,EAAM,IAAU,EAAK,gBAAkB,EAAM,eAAe,CAC3E,EACA,OAAOC,EAAAA,oBAAoB,EAAkB,EAAc,CAA4B,CACzF,CAEA,SAAS,EACP,EACA,EACA,EAC6B,CAC7B,IAAM,EAA2C,CAAC,EAG5C,kBAAmB,IAAI,IACvB,EAAiB,IAAI,IACzB,EAAM,KAAK,EAAM,IAAU,CACzB,EAAK,KACL,CAAE,OAAQ,EAAK,OAAQ,gBAAiB,EAAK,gBAAiB,OAAQ,EAAa,IAAU,CAAE,CACjG,CAAC,CACH,EACM,kBAAoB,IAAI,IAC1B,EAAsB,EAC1B,IAAK,IAAM,KAAS,EAAQ,CAI1B,GAAuBC,EAAAA,wBAAwB,CAAK,EACpD,IAAM,EAAc,EACjB,IAAK,GAAe,CACnB,GAAM,CAAE,OAAM,YAAW,UAAS,cAAe,EAC3C,EAAc,EAAuB,EAAY,EAAgB,CAAiB,EACxF,MAAO,CAAE,OAAM,YAAW,UAAS,GAAI,GAAe,CAAE,aAAY,EAAI,YAAW,CACrF,CAAC,CAAC,CACD,UAAU,EAAM,IAAU,EAAK,KAAK,cAAc,EAAM,IAAI,GAAK,EAAK,UAAY,EAAM,SAAS,EAC9F,EAAQ,CAAC,GAAG,IAAI,IAAI,EAAY,KAAK,CAAE,UAAW,CAAI,CAAC,CAAC,EAC9D,IAAK,IAAM,KAAQ,EACjB,EAAiB,IAAI,GAAO,EAAiB,IAAI,CAAI,GAAK,GAAK,CAAC,EAElE,EAAS,KAAK,CAAE,QAAO,cAAa,WAAY,KAAK,IAAI,GAAG,EAAM,KAAK,CAAE,gBAAiB,CAAU,CAAC,CAAE,CAAC,CAC1G,CAOA,OANA,EAAS,MACN,EAAM,IACL,EAAM,WAAa,EAAK,aACvB,EAAK,YAAY,EAAE,EAAE,MAAQ,GAAA,CAAI,cAAc,EAAM,YAAY,EAAE,EAAE,MAAQ,EAAE,IAC/E,EAAK,YAAY,EAAE,EAAE,WAAa,IAAM,EAAM,YAAY,EAAE,EAAE,WAAa,EAChF,EACO,CACL,sBACA,+BAAgC,OAAO,YAAY,CAAgB,EACnE,2BAA4B,OAAO,YACjC,CAAC,GAAG,CAAiB,CAAC,CAAC,KAAK,CAAC,EAAM,KAAW,CAAC,EAAM,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAAC,CAAC,CAC1G,EACA,OAAQ,CACV,CACF,CAUA,SAAS,EACP,EACA,EAIA,EACsB,CACtB,IAAM,EAAW,EAAe,IAAI,EAAW,IAAI,EACnD,GAAI,CAAC,GAAU,OACb,OAEF,IAAI,EAAY,EAAkB,IAAI,EAAW,IAAI,EAChD,IACH,kBAAY,IAAI,IAChB,EAAkB,IAAI,EAAW,KAAM,CAAS,GAElD,IAAM,EAAQC,EAAAA,oBACZ,EAAW,SAAS,IAAK,GAAY,EAAM,EAAS,CAAC,EAAS,MAAM,CAAC,EACrE,EAAW,aAAa,IAAK,GAAQ,EAAM,EAAK,CAAC,EAAS,MAAM,CAAC,EACjE,EAAS,OACT,EAAS,eACX,EACA,IAAK,IAAM,KAAQ,EACjB,EAAU,IAAI,CAAI,EAEpB,MAAO,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAC1D"}

package/dist/crossFileDuplication.d.ts

@@ -7,7 +7,18 @@ export interface CrossFileDuplicationSourceFile extends Partial<CrossFileDuplica
 export interface CrossFileDuplicateOccurrence {
     endLine: number;
     file: string;
+    /**
+     * The code lines carrying the occurrence's matched tokens (see
+     * DuplicateBlockOccurrence.lineNumbers). Absent for a file that supplied only candidates (no
+     * `tokens`), whose matched lines are unknowable.
+     */
+    lineNumbers?: number[];
     startLine: number;
+    /**
+     * The tokens of its matched runs, or of a near-miss block or its matched cores with the edited
+     * tokens among them; the group's `tokenCount` is the smallest of its occurrences'.
+     */
+    tokenCount: number;
 }
 export interface CrossFileDuplicateBlockGroup {
     files: string[];
@@ -26,8 +37,9 @@ export interface CrossFileDuplicationMetrics {
     duplicateBlockGroupCountByFile: Record<string, number>;
     /**
      * Per file, the 1-based code lines covered by the tokens of its cross-file occurrences, sorted
-     * ascending: the matched tokens of exact and gapped occurrences, and every token of a near-miss
-     * block or its matched cores, edited ones included (like within-file near-miss coverage). The unmatched gap of a merged
+     * ascending: the matched tokens of exact and gapped occurrences, and the lines of a near-miss
+     * block or its matched cores more than half of whose tokens its partners match (like within-file
+     * near-miss coverage). The unmatched gap of a merged
      * clone and comment/blank lines inside an occurrence's bounding range are excluded (blank rows
      * inside multi-row tokens only when the file supplied codeLineNumbers). A file that supplied
      * only candidates (no `tokens`) has no entry — without its token stream the covered lines are

package/dist/crossFileDuplication.js

@@ -1,2 +1,2 @@
-import{buildLiteralCountPrefix as e,collectSegmentLines as t,collectSequenceWindowCandidates as n,countRedundantFragments as r,mergeAdjacentGroups as i,resolveDuplicationOptions as a}from"./duplication.js";import{collectCrossFileNearMissGroups as o}from"./crossFileNearMiss.js";import{selectMaximalGroups as s}from"./duplicateSelection.js";function c(e,t){let{minTokens:n,maxGapTokens:r,minSimilarityPercent:i}=a(t),o=e.flatMap(({file:e,candidates:t},n)=>t.map(t=>({...t,regionBucket:n,file:e})));for(let t of u(e,n))o.push(t);let c=s(o,f,(e,t)=>e.regionBucket-t.regionBucket||e.startIndex-t.startIndex),d=p(e,r),g=m([...c.values()],d,r);for(let t of l(e,g,d,n,i))g.push(t);return h(g,e,d)}function l(e,t,n,r,i){let a=e.map(()=>[]);for(let{fileIndex:e,startTokenIndex:r,endTokenIndex:i}of t.flat()){let t=n[e]??0;a[e]?.push({startTokenIndex:r-t,endTokenIndex:i-t})}return o(e,a,r,i).map(t=>t.map(t=>{let r=n[t.fileIndex]??0;return{...t,file:e[t.fileIndex]?.file??``,segments:t.segments.map(e=>({startTokenIndex:e.startTokenIndex+r,endTokenIndex:e.endTokenIndex+r})),startTokenIndex:t.startTokenIndex+r,endTokenIndex:t.endTokenIndex+r}}))}function u(t,r){let i=[],a=[];for(let[n,{tokens:r,containerStatements:o}]of t.entries())r&&o&&(i.push(n),a.push({tokens:r,literalCountPrefix:e(r),containers:o}));return a.length<2?[]:n(a,r,!0).flatMap(({candidate:e,contextIndex:n})=>{let r=i[n],a=r===void 0?void 0:t[r];return r===void 0||a===void 0?[]:[{...e,regionBucket:r,file:a.file}]})}function d(e){return new Set(e.map(e=>e.file)).size>=2}function f(e){return e.length>=2&&new Set(e.map(e=>e.regionBucket)).size>=2}function p(e,t){let n=[],r=0;for(let{tokens:i,candidates:a}of e){n.push(r);let e=i?.length??0;if(!i)for(let t of a)e=Math.max(e,t.endTokenIndex);r+=e+t+1}return n}function m(e,t,n){let r=e.map(e=>e.map(e=>{let n=e.startTokenIndex+(t[e.regionBucket]??0),r=e.endTokenIndex+(t[e.regionBucket]??0);return{file:e.file,fileIndex:e.regionBucket,spanCountedElsewhere:e.nestedInLargerGroup,nestedInLargerGroup:e.nestedInLargerGroup,segments:[{startTokenIndex:n,endTokenIndex:r}],tokenCount:e.tokenCount,startTokenIndex:n,endTokenIndex:r,startIndex:e.startIndex,endIndex:e.endIndex,startLine:e.startLine,endLine:e.endLine}}).toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex));return i(r,n,d)}function h(e,t,n){let i=[],a=/* @__PURE__ */ new Map,o=new Map(t.map((e,t)=>[e.file,{tokens:e.tokens,codeLineNumbers:e.codeLineNumbers,offset:n[t]??0}])),s=/* @__PURE__ */ new Map,c=0;for(let t of e){c+=r(t);for(let e of t)g(e,o,s);let e=t.map(({file:e,startLine:t,endLine:n})=>({file:e,startLine:t,endLine:n})).toSorted((e,t)=>e.file.localeCompare(t.file)||e.startLine-t.startLine),n=[...new Set(e.map(({file:e})=>e))];for(let e of n)a.set(e,(a.get(e)??0)+1);i.push({files:n,occurrences:e,tokenCount:Math.min(...t.map(({tokenCount:e})=>e))})}return i.sort((e,t)=>t.tokenCount-e.tokenCount||(e.occurrences[0]?.file??``).localeCompare(t.occurrences[0]?.file??``)||(e.occurrences[0]?.startLine??0)-(t.occurrences[0]?.startLine??0)),{duplicateBlockCount:c,duplicateBlockGroupCountByFile:Object.fromEntries(a),duplicateLineNumbersByFile:Object.fromEntries([...s].map(([e,t])=>[e,[...t].toSorted((e,t)=>e-t)])),groups:i}}function g(e,n,r){let i=n.get(e.file);if(!i?.tokens)return;let a=r.get(e.file);a||(a=/* @__PURE__ */ new Set,r.set(e.file,a));for(let n of e.segments)t({startTokenIndex:n.startTokenIndex-i.offset,endTokenIndex:n.endTokenIndex-i.offset},i.tokens,i.codeLineNumbers,a)}export{c as measureCrossFileDuplication};
+import{buildLiteralCountPrefix as e,collectMatchedLines as t,collectSequenceWindowCandidates as n,countRedundantFragments as r,mergeAdjacentGroups as i,resolveDuplicationOptions as a}from"./duplication.js";import{collectCrossFileNearMissGroups as o}from"./crossFileNearMiss.js";import{selectMaximalGroups as s}from"./duplicateSelection.js";function c(e,t){let{minTokens:n,maxGapTokens:r,minSimilarityPercent:i}=a(t),o=e.flatMap(({file:e,candidates:t},n)=>t.map(t=>({...t,regionBucket:n,file:e})));for(let t of d(e,n))o.push(t);let c=s(o,p,(e,t)=>e.regionBucket-t.regionBucket||e.startIndex-t.startIndex),u=m(e,r),f=h([...c.values()],u,r);for(let t of l(e,f,u,n,i))f.push(t);return g(f,e,u)}function l(e,t,n,r,i){let a=e.map(()=>[]);for(let{fileIndex:e,startTokenIndex:r,endTokenIndex:i}of t.flat()){let t=n[e]??0;a[e]?.push({startTokenIndex:r-t,endTokenIndex:i-t})}return o(e,a,r,i).map(t=>t.map(t=>{let r=n[t.fileIndex]??0;return{...t,file:e[t.fileIndex]?.file??``,segments:t.segments.map(e=>u(e,r)),matchedRuns:t.matchedRuns?.map(e=>u(e,r)),startTokenIndex:t.startTokenIndex+r,endTokenIndex:t.endTokenIndex+r}}))}function u({startTokenIndex:e,endTokenIndex:t},n){return{startTokenIndex:e+n,endTokenIndex:t+n}}function d(t,r){let i=[],a=[];for(let[n,{tokens:r,containerStatements:o}]of t.entries())r&&o&&(i.push(n),a.push({tokens:r,literalCountPrefix:e(r),containers:o}));return a.length<2?[]:n(a,r,!0).flatMap(({candidate:e,contextIndex:n})=>{let r=i[n],a=r===void 0?void 0:t[r];return r===void 0||a===void 0?[]:[{...e,regionBucket:r,file:a.file}]})}function f(e){return new Set(e.map(e=>e.file)).size>=2}function p(e){return e.length>=2&&new Set(e.map(e=>e.regionBucket)).size>=2}function m(e,t){let n=[],r=0;for(let{tokens:i,candidates:a}of e){n.push(r);let e=i?.length??0;if(!i)for(let t of a)e=Math.max(e,t.endTokenIndex);r+=e+t+1}return n}function h(e,t,n){let r=e.map(e=>e.map(e=>{let n=e.startTokenIndex+(t[e.regionBucket]??0),r=e.endTokenIndex+(t[e.regionBucket]??0);return{file:e.file,fileIndex:e.regionBucket,spanCountedElsewhere:e.nestedInLargerGroup,nestedInLargerGroup:e.nestedInLargerGroup,segments:[{startTokenIndex:n,endTokenIndex:r}],tokenCount:e.tokenCount,startTokenIndex:n,endTokenIndex:r,startIndex:e.startIndex,endIndex:e.endIndex,startLine:e.startLine,endLine:e.endLine}}).toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex));return i(r,n,f)}function g(e,t,n){let i=[],a=/* @__PURE__ */ new Map,o=new Map(t.map((e,t)=>[e.file,{tokens:e.tokens,codeLineNumbers:e.codeLineNumbers,offset:n[t]??0}])),s=/* @__PURE__ */ new Map,c=0;for(let t of e){c+=r(t);let e=t.map(e=>{let{file:t,startLine:n,endLine:r,tokenCount:i}=e,a=_(e,o,s);return{file:t,startLine:n,endLine:r,...a&&{lineNumbers:a},tokenCount:i}}).toSorted((e,t)=>e.file.localeCompare(t.file)||e.startLine-t.startLine),n=[...new Set(e.map(({file:e})=>e))];for(let e of n)a.set(e,(a.get(e)??0)+1);i.push({files:n,occurrences:e,tokenCount:Math.min(...t.map(({tokenCount:e})=>e))})}return i.sort((e,t)=>t.tokenCount-e.tokenCount||(e.occurrences[0]?.file??``).localeCompare(t.occurrences[0]?.file??``)||(e.occurrences[0]?.startLine??0)-(t.occurrences[0]?.startLine??0)),{duplicateBlockCount:c,duplicateBlockGroupCountByFile:Object.fromEntries(a),duplicateLineNumbersByFile:Object.fromEntries([...s].map(([e,t])=>[e,[...t].toSorted((e,t)=>e-t)])),groups:i}}function _(e,n,r){let i=n.get(e.file);if(!i?.tokens)return;let a=r.get(e.file);a||(a=/* @__PURE__ */ new Set,r.set(e.file,a));let o=t(e.segments.map(e=>u(e,-i.offset)),e.matchedRuns?.map(e=>u(e,-i.offset)),i.tokens,i.codeLineNumbers);for(let e of o)a.add(e);return[...o].toSorted((e,t)=>e-t)}export{c as measureCrossFileDuplication};
 //# sourceMappingURL=crossFileDuplication.js.map

package/dist/crossFileDuplication.js.map

@@ -1 +1 @@
-{"version":3,"file":"crossFileDuplication.js","names":[],"sources":["../src/crossFileDuplication.ts"],"sourcesContent":["import { collectCrossFileNearMissGroups } from './crossFileNearMiss.js';\nimport { selectMaximalGroups, type SelectableRegion } from './duplicateSelection.js';\nimport {\n  buildLiteralCountPrefix,\n  collectSegmentLines,\n  collectSequenceWindowCandidates,\n  countRedundantFragments,\n  mergeAdjacentGroups,\n  resolveDuplicationOptions,\n  type CountedOccurrence,\n  type CrossFileDuplicateCandidate,\n  type CrossFileDuplicationFileData,\n  type SequenceWindowContext,\n} from './duplication.js';\nimport type { DuplicationOptions } from './types.js';\n\nexport interface CrossFileDuplicationSourceFile extends Partial<CrossFileDuplicationFileData> {\n  file: string;\n  candidates: CrossFileDuplicateCandidate[];\n}\n\nexport interface CrossFileDuplicateOccurrence {\n  endLine: number;\n  file: string;\n  startLine: number;\n}\n\nexport interface CrossFileDuplicateBlockGroup {\n  files: string[];\n  occurrences: CrossFileDuplicateOccurrence[];\n  /**\n   * Token count of the smallest occurrence. For exact and gapped groups it is the matched token\n   * count every occurrence shares (gaps are not counted); for near-miss (Type-3) groups it is the\n   * token count of the smallest whole block or set of matched cores, edited tokens included.\n   */\n  tokenCount: number;\n}\n\nexport interface CrossFileDuplicationMetrics {\n  /** Number of redundant copies across all groups, counted per matched fragment like within-file. */\n  duplicateBlockCount: number;\n  /** Groups the file participates in, keyed by the file name passed in. */\n  duplicateBlockGroupCountByFile: Record<string, number>;\n  /**\n   * Per file, the 1-based code lines covered by the tokens of its cross-file occurrences, sorted\n   * ascending: the matched tokens of exact and gapped occurrences, and every token of a near-miss\n   * block or its matched cores, edited ones included (like within-file near-miss coverage). The unmatched gap of a merged\n   * clone and comment/blank lines inside an occurrence's bounding range are excluded (blank rows\n   * inside multi-row tokens only when the file supplied codeLineNumbers). A file that supplied\n   * only candidates (no `tokens`) has no entry — without its token stream the covered lines are\n   * unknowable, and an approximate bounding range would break this field's exactness.\n   */\n  duplicateLineNumbersByFile: Record<string, number[]>;\n  groups: CrossFileDuplicateBlockGroup[];\n}\n\ninterface SelectableCandidate extends CrossFileDuplicateCandidate, SelectableRegion {\n  regionBucket: number;\n  file: string;\n}\n\n/** A cross-file occurrence: a within-file occurrence in the project-wide token index space. */\ninterface CrossFileOccurrence extends CountedOccurrence {\n  file: string;\n  fileIndex: number;\n}\n\n/**\n * Detects code regions duplicated across files. Per-file candidates (whole block subtrees and full\n * container runs, fingerprinted with the same normalization as within-file duplication) are joined\n * by a project-level window index over per-statement fingerprint sequences (CPD-style), so a\n * copy-pasted partial statement run embedded in different surrounding code is matched even though\n * no single file can know it repeats elsewhere. Candidates are grouped by fingerprint, and only\n * maximal, non-overlapping regions whose group spans at least two files are counted. Groups that\n * shrink to a single file during selection are shed — a within-file repeat is already reported by\n * that file's own duplication metrics. A copy nested inside a larger group's region (two files share\n * a whole function, a third file only a block of it) is reported with its group, so the third\n * file's copy still shows what it duplicates. Groups separated by a small token gap within each file then\n * merge into gapped (Type-3) clone groups under `maxGapTokens`, exactly like within-file merging.\n * Finally, blocks of files that supplied `nearMissBlocks` are compared across files for near-miss\n * (Type-3) clones under `minSimilarityPercent` (see crossFileNearMiss.ts).\n */\nexport function measureCrossFileDuplication(\n  files: CrossFileDuplicationSourceFile[],\n  options?: DuplicationOptions\n): CrossFileDuplicationMetrics {\n  const { minTokens, maxGapTokens, minSimilarityPercent } = resolveDuplicationOptions(options);\n  const candidates: SelectableCandidate[] = files.flatMap(({ file, candidates }, fileIndex) =>\n    candidates.map((candidate) => ({ ...candidate, regionBucket: fileIndex, file }))\n  );\n  // Pushed one by one: spreading the project-scale window-candidate array as call arguments\n  // overflows V8's argument limit (~124k) and crashes on Node, though Bun/JSC tolerates it.\n  for (const candidate of collectWindowCandidates(files, minTokens)) {\n    candidates.push(candidate);\n  }\n  const counted = selectMaximalGroups(\n    candidates,\n    spansMultipleFiles,\n    // File index and position break coverage ties deterministically.\n    (left, right) => left.regionBucket - right.regionBucket || left.startIndex - right.startIndex\n  );\n  const tokenOffsets = computeTokenOffsets(files, maxGapTokens);\n  const groups = mergeGapAdjacentGroups([...counted.values()], tokenOffsets, maxGapTokens);\n  for (const group of collectNearMissGroups(files, groups, tokenOffsets, minTokens, minSimilarityPercent)) {\n    groups.push(group);\n  }\n  return summarize(groups, files, tokenOffsets);\n}\n\n/** Near-miss groups among the blocks the exact groups leave unreported, in the project token space. */\nfunction collectNearMissGroups(\n  files: CrossFileDuplicationSourceFile[],\n  exactGroups: CrossFileOccurrence[][],\n  tokenOffsets: number[],\n  minTokens: number,\n  minSimilarityPercent: number\n): CrossFileOccurrence[][] {\n  const reportedSpansByFile: { startTokenIndex: number; endTokenIndex: number }[][] = files.map(() => []);\n  for (const { fileIndex, startTokenIndex, endTokenIndex } of exactGroups.flat()) {\n    const offset = tokenOffsets[fileIndex] ?? 0;\n    reportedSpansByFile[fileIndex]?.push({\n      startTokenIndex: startTokenIndex - offset,\n      endTokenIndex: endTokenIndex - offset,\n    });\n  }\n  return collectCrossFileNearMissGroups(files, reportedSpansByFile, minTokens, minSimilarityPercent).map((group) =>\n    group.map((occurrence) => {\n      const offset = tokenOffsets[occurrence.fileIndex] ?? 0;\n      return {\n        ...occurrence,\n        file: files[occurrence.fileIndex]?.file ?? '',\n        segments: occurrence.segments.map((segment) => ({\n          startTokenIndex: segment.startTokenIndex + offset,\n          endTokenIndex: segment.endTokenIndex + offset,\n        })),\n        startTokenIndex: occurrence.startTokenIndex + offset,\n        endTokenIndex: occurrence.endTokenIndex + offset,\n      };\n    })\n  );\n}\n\n/** Repeated sub-windows of sibling statements matched across the whole project's files. */\nfunction collectWindowCandidates(files: CrossFileDuplicationSourceFile[], minTokens: number): SelectableCandidate[] {\n  const fileIndexByContext: number[] = [];\n  const contexts: SequenceWindowContext[] = [];\n  for (const [fileIndex, { tokens, containerStatements }] of files.entries()) {\n    if (tokens && containerStatements) {\n      fileIndexByContext.push(fileIndex);\n      contexts.push({ tokens, literalCountPrefix: buildLiteralCountPrefix(tokens), containers: containerStatements });\n    }\n  }\n  if (contexts.length < 2) {\n    return [];\n  }\n  return collectSequenceWindowCandidates(contexts, minTokens, true).flatMap(({ candidate, contextIndex }) => {\n    const fileIndex = fileIndexByContext[contextIndex];\n    const file = fileIndex === undefined ? undefined : files[fileIndex];\n    return fileIndex === undefined || file === undefined\n      ? []\n      : [{ ...candidate, regionBucket: fileIndex, file: file.file }];\n  });\n}\n\n/** A merged group is reported only while it still covers more than one file. */\nfunction spansMultipleFilesAfterMerge(group: CrossFileOccurrence[]): boolean {\n  return new Set(group.map((occurrence) => occurrence.file)).size >= 2;\n}\n\nfunction spansMultipleFiles(group: SelectableCandidate[]): boolean {\n  return group.length >= 2 && new Set(group.map((candidate) => candidate.regionBucket)).size >= 2;\n}\n\n/**\n * Per-file token offsets that map every file into one project-wide token index space: each file's\n * tokens are offset by more than `maxGapTokens` past the previous file's, so occurrences in\n * different files are never gap-adjacent and merged pairs always stay within one file.\n */\nfunction computeTokenOffsets(files: CrossFileDuplicationSourceFile[], maxGapTokens: number): number[] {\n  const tokenOffsets: number[] = [];\n  let offset = 0;\n  for (const { tokens, candidates } of files) {\n    tokenOffsets.push(offset);\n    // Accumulated in a loop: spreading a project-scale candidate array as call arguments would\n    // overflow V8's argument limit (~124k) and crash on Node.\n    let tokenCount = tokens?.length ?? 0;\n    if (!tokens) {\n      for (const candidate of candidates) {\n        tokenCount = Math.max(tokenCount, candidate.endTokenIndex);\n      }\n    }\n    offset += tokenCount + maxGapTokens + 1;\n  }\n  return tokenOffsets;\n}\n\n/** Reuses the within-file gapped (Type-3) merging in the project-wide token index space. */\nfunction mergeGapAdjacentGroups(\n  groups: SelectableCandidate[][],\n  tokenOffsets: number[],\n  maxGapTokens: number\n): CrossFileOccurrence[][] {\n  const occurrenceGroups = groups.map((group) =>\n    group\n      .map((candidate): CrossFileOccurrence => {\n        const start = candidate.startTokenIndex + (tokenOffsets[candidate.regionBucket] ?? 0);\n        const end = candidate.endTokenIndex + (tokenOffsets[candidate.regionBucket] ?? 0);\n        return {\n          file: candidate.file,\n          fileIndex: candidate.regionBucket,\n          spanCountedElsewhere: candidate.nestedInLargerGroup,\n          nestedInLargerGroup: candidate.nestedInLargerGroup,\n          segments: [{ startTokenIndex: start, endTokenIndex: end }],\n          tokenCount: candidate.tokenCount,\n          startTokenIndex: start,\n          endTokenIndex: end,\n          startIndex: candidate.startIndex,\n          endIndex: candidate.endIndex,\n          startLine: candidate.startLine,\n          endLine: candidate.endLine,\n        };\n      })\n      .toSorted((left, right) => left.startTokenIndex - right.startTokenIndex)\n  );\n  return mergeAdjacentGroups(occurrenceGroups, maxGapTokens, spansMultipleFilesAfterMerge);\n}\n\nfunction summarize(\n  groups: CrossFileOccurrence[][],\n  files: CrossFileDuplicationSourceFile[],\n  tokenOffsets: number[]\n): CrossFileDuplicationMetrics {\n  const reported: CrossFileDuplicateBlockGroup[] = [];\n  // Accumulated in Maps: file names are arbitrary strings, and a plain object would read\n  // inherited properties for names like \"constructor\".\n  const groupCountByFile = new Map<string, number>();\n  const fileDataByName = new Map(\n    files.map((file, index) => [\n      file.file,\n      { tokens: file.tokens, codeLineNumbers: file.codeLineNumbers, offset: tokenOffsets[index] ?? 0 },\n    ])\n  );\n  const lineNumbersByFile = new Map<string, Set<number>>();\n  let duplicateBlockCount = 0;\n  for (const group of groups) {\n    // Mirrors within-file counting: each redundant occurrence contributes one count per matched\n    // fragment, gapped merging consolidates the grouping without halving the count, and spans a\n    // partial merge shares between a retained group and the merged group count once.\n    duplicateBlockCount += countRedundantFragments(group);\n    for (const occurrence of group) {\n      collectOccurrenceLines(occurrence, fileDataByName, lineNumbersByFile);\n    }\n    const occurrences = group\n      .map(({ file, startLine, endLine }) => ({ file, startLine, endLine }))\n      .toSorted((left, right) => left.file.localeCompare(right.file) || left.startLine - right.startLine);\n    const files = [...new Set(occurrences.map(({ file }) => file))];\n    for (const file of files) {\n      groupCountByFile.set(file, (groupCountByFile.get(file) ?? 0) + 1);\n    }\n    reported.push({ files, occurrences, tokenCount: Math.min(...group.map(({ tokenCount }) => tokenCount)) });\n  }\n  reported.sort(\n    (left, right) =>\n      right.tokenCount - left.tokenCount ||\n      (left.occurrences[0]?.file ?? '').localeCompare(right.occurrences[0]?.file ?? '') ||\n      (left.occurrences[0]?.startLine ?? 0) - (right.occurrences[0]?.startLine ?? 0)\n  );\n  return {\n    duplicateBlockCount,\n    duplicateBlockGroupCountByFile: Object.fromEntries(groupCountByFile),\n    duplicateLineNumbersByFile: Object.fromEntries(\n      [...lineNumbersByFile].map(([file, lines]) => [file, [...lines].toSorted((left, right) => left - right)])\n    ),\n    groups: reported,\n  };\n}\n\n/**\n * Adds the code lines an occurrence's segment tokens cover (matched tokens of an exact or gapped\n * occurrence, the whole block or matched cores of a near-miss one) to its file's line set, mapping the\n * project-wide token segments back into the file's own token stream. A file that supplied only\n * candidates (no token stream) is skipped rather than approximated from the bounding line range,\n * which would include gap and comment/blank lines and break the field's exactness contract.\n */\nfunction collectOccurrenceLines(\n  occurrence: CrossFileOccurrence,\n  fileDataByName: Map<\n    string,\n    { tokens?: CrossFileDuplicationSourceFile['tokens']; codeLineNumbers?: Set<number>; offset: number }\n  >,\n  lineNumbersByFile: Map<string, Set<number>>\n): void {\n  const fileData = fileDataByName.get(occurrence.file);\n  if (!fileData?.tokens) {\n    return;\n  }\n  let lines = lineNumbersByFile.get(occurrence.file);\n  if (!lines) {\n    lines = new Set();\n    lineNumbersByFile.set(occurrence.file, lines);\n  }\n  for (const segment of occurrence.segments) {\n    collectSegmentLines(\n      {\n        startTokenIndex: segment.startTokenIndex - fileData.offset,\n        endTokenIndex: segment.endTokenIndex - fileData.offset,\n      },\n      fileData.tokens,\n      fileData.codeLineNumbers,\n      lines\n    );\n  }\n}\n"],"mappings":"oVAkFA,SAAgB,EACd,EACA,EAC6B,CAC7B,GAAM,CAAE,YAAW,eAAc,wBAAyB,EAA0B,CAAO,EACrF,EAAoC,EAAM,SAAS,CAAE,OAAM,cAAc,IAC7E,EAAW,IAAK,IAAe,CAAE,GAAG,EAAW,aAAc,EAAW,MAAK,EAAE,CACjF,EAGA,IAAK,IAAM,KAAa,EAAwB,EAAO,CAAS,EAC9D,EAAW,KAAK,CAAS,EAE3B,IAAM,EAAU,EACd,EACA,GAEC,EAAM,IAAU,EAAK,aAAe,EAAM,cAAgB,EAAK,WAAa,EAAM,UACrF,EACM,EAAe,EAAoB,EAAO,CAAY,EACtD,EAAS,EAAuB,CAAC,GAAG,EAAQ,OAAO,CAAC,EAAG,EAAc,CAAY,EACvF,IAAK,IAAM,KAAS,EAAsB,EAAO,EAAQ,EAAc,EAAW,CAAoB,EACpG,EAAO,KAAK,CAAK,EAEnB,OAAO,EAAU,EAAQ,EAAO,CAAY,CAC9C,CAGA,SAAS,EACP,EACA,EACA,EACA,EACA,EACyB,CACzB,IAAM,EAA8E,EAAM,QAAU,CAAC,CAAC,EACtG,IAAK,GAAM,CAAE,YAAW,kBAAiB,mBAAmB,EAAY,KAAK,EAAG,CAC9E,IAAM,EAAS,EAAa,IAAc,EAC1C,EAAoB,EAAU,EAAE,KAAK,CACnC,gBAAiB,EAAkB,EACnC,cAAe,EAAgB,CACjC,CAAC,CACH,CACA,OAAO,EAA+B,EAAO,EAAqB,EAAW,CAAoB,CAAC,CAAC,IAAK,GACtG,EAAM,IAAK,GAAe,CACxB,IAAM,EAAS,EAAa,EAAW,YAAc,EACrD,MAAO,CACL,GAAG,EACH,KAAM,EAAM,EAAW,UAAU,EAAE,MAAQ,GAC3C,SAAU,EAAW,SAAS,IAAK,IAAa,CAC9C,gBAAiB,EAAQ,gBAAkB,EAC3C,cAAe,EAAQ,cAAgB,CACzC,EAAE,EACF,gBAAiB,EAAW,gBAAkB,EAC9C,cAAe,EAAW,cAAgB,CAC5C,CACF,CAAC,CACH,CACF,CAGA,SAAS,EAAwB,EAAyC,EAA0C,CAClH,IAAM,EAA+B,CAAC,EAChC,EAAoC,CAAC,EAC3C,IAAK,GAAM,CAAC,EAAW,CAAE,SAAQ,0BAA0B,EAAM,QAAQ,EACnE,GAAU,IACZ,EAAmB,KAAK,CAAS,EACjC,EAAS,KAAK,CAAE,SAAQ,mBAAoB,EAAwB,CAAM,EAAG,WAAY,CAAoB,CAAC,GAMlH,OAHI,EAAS,OAAS,EACb,CAAC,EAEH,EAAgC,EAAU,EAAW,EAAI,CAAC,CAAC,SAAS,CAAE,YAAW,kBAAmB,CACzG,IAAM,EAAY,EAAmB,GAC/B,EAAO,IAAc,IAAA,GAAY,IAAA,GAAY,EAAM,GACzD,OAAO,IAAc,IAAA,IAAa,IAAS,IAAA,GACvC,CAAC,EACD,CAAC,CAAE,GAAG,EAAW,aAAc,EAAW,KAAM,EAAK,IAAK,CAAC,CACjE,CAAC,CACH,CAGA,SAAS,EAA6B,EAAuC,CAC3E,OAAO,IAAI,IAAI,EAAM,IAAK,GAAe,EAAW,IAAI,CAAC,CAAC,CAAC,MAAQ,CACrE,CAEA,SAAS,EAAmB,EAAuC,CACjE,OAAO,EAAM,QAAU,GAAK,IAAI,IAAI,EAAM,IAAK,GAAc,EAAU,YAAY,CAAC,CAAC,CAAC,MAAQ,CAChG,CAOA,SAAS,EAAoB,EAAyC,EAAgC,CACpG,IAAM,EAAyB,CAAC,EAC5B,EAAS,EACb,IAAK,GAAM,CAAE,SAAQ,gBAAgB,EAAO,CAC1C,EAAa,KAAK,CAAM,EAGxB,IAAI,EAAa,GAAQ,QAAU,EACnC,GAAI,CAAC,EACH,IAAK,IAAM,KAAa,EACtB,EAAa,KAAK,IAAI,EAAY,EAAU,aAAa,EAG7D,GAAU,EAAa,EAAe,CACxC,CACA,OAAO,CACT,CAGA,SAAS,EACP,EACA,EACA,EACyB,CACzB,IAAM,EAAmB,EAAO,IAAK,GACnC,EACG,IAAK,GAAmC,CACvC,IAAM,EAAQ,EAAU,iBAAmB,EAAa,EAAU,eAAiB,GAC7E,EAAM,EAAU,eAAiB,EAAa,EAAU,eAAiB,GAC/E,MAAO,CACL,KAAM,EAAU,KAChB,UAAW,EAAU,aACrB,qBAAsB,EAAU,oBAChC,oBAAqB,EAAU,oBAC/B,SAAU,CAAC,CAAE,gBAAiB,EAAO,cAAe,CAAI,CAAC,EACzD,WAAY,EAAU,WACtB,gBAAiB,EACjB,cAAe,EACf,WAAY,EAAU,WACtB,SAAU,EAAU,SACpB,UAAW,EAAU,UACrB,QAAS,EAAU,OACrB,CACF,CAAC,CAAC,CACD,UAAU,EAAM,IAAU,EAAK,gBAAkB,EAAM,eAAe,CAC3E,EACA,OAAO,EAAoB,EAAkB,EAAc,CAA4B,CACzF,CAEA,SAAS,EACP,EACA,EACA,EAC6B,CAC7B,IAAM,EAA2C,CAAC,EAG5C,kBAAmB,IAAI,IACvB,EAAiB,IAAI,IACzB,EAAM,KAAK,EAAM,IAAU,CACzB,EAAK,KACL,CAAE,OAAQ,EAAK,OAAQ,gBAAiB,EAAK,gBAAiB,OAAQ,EAAa,IAAU,CAAE,CACjG,CAAC,CACH,EACM,kBAAoB,IAAI,IAC1B,EAAsB,EAC1B,IAAK,IAAM,KAAS,EAAQ,CAI1B,GAAuB,EAAwB,CAAK,EACpD,IAAK,IAAM,KAAc,EACvB,EAAuB,EAAY,EAAgB,CAAiB,EAEtE,IAAM,EAAc,EACjB,KAAK,CAAE,OAAM,YAAW,cAAe,CAAE,OAAM,YAAW,SAAQ,EAAE,CAAC,CACrE,UAAU,EAAM,IAAU,EAAK,KAAK,cAAc,EAAM,IAAI,GAAK,EAAK,UAAY,EAAM,SAAS,EAC9F,EAAQ,CAAC,GAAG,IAAI,IAAI,EAAY,KAAK,CAAE,UAAW,CAAI,CAAC,CAAC,EAC9D,IAAK,IAAM,KAAQ,EACjB,EAAiB,IAAI,GAAO,EAAiB,IAAI,CAAI,GAAK,GAAK,CAAC,EAElE,EAAS,KAAK,CAAE,QAAO,cAAa,WAAY,KAAK,IAAI,GAAG,EAAM,KAAK,CAAE,gBAAiB,CAAU,CAAC,CAAE,CAAC,CAC1G,CAOA,OANA,EAAS,MACN,EAAM,IACL,EAAM,WAAa,EAAK,aACvB,EAAK,YAAY,EAAE,EAAE,MAAQ,GAAA,CAAI,cAAc,EAAM,YAAY,EAAE,EAAE,MAAQ,EAAE,IAC/E,EAAK,YAAY,EAAE,EAAE,WAAa,IAAM,EAAM,YAAY,EAAE,EAAE,WAAa,EAChF,EACO,CACL,sBACA,+BAAgC,OAAO,YAAY,CAAgB,EACnE,2BAA4B,OAAO,YACjC,CAAC,GAAG,CAAiB,CAAC,CAAC,KAAK,CAAC,EAAM,KAAW,CAAC,EAAM,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAAC,CAAC,CAC1G,EACA,OAAQ,CACV,CACF,CASA,SAAS,EACP,EACA,EAIA,EACM,CACN,IAAM,EAAW,EAAe,IAAI,EAAW,IAAI,EACnD,GAAI,CAAC,GAAU,OACb,OAEF,IAAI,EAAQ,EAAkB,IAAI,EAAW,IAAI,EAC5C,IACH,kBAAQ,IAAI,IACZ,EAAkB,IAAI,EAAW,KAAM,CAAK,GAE9C,IAAK,IAAM,KAAW,EAAW,SAC/B,EACE,CACE,gBAAiB,EAAQ,gBAAkB,EAAS,OACpD,cAAe,EAAQ,cAAgB,EAAS,MAClD,EACA,EAAS,OACT,EAAS,gBACT,CACF,CAEJ"}
+{"version":3,"file":"crossFileDuplication.js","names":[],"sources":["../src/crossFileDuplication.ts"],"sourcesContent":["import { collectCrossFileNearMissGroups } from './crossFileNearMiss.js';\nimport { selectMaximalGroups, type SelectableRegion } from './duplicateSelection.js';\nimport {\n  buildLiteralCountPrefix,\n  collectMatchedLines,\n  collectSequenceWindowCandidates,\n  countRedundantFragments,\n  mergeAdjacentGroups,\n  resolveDuplicationOptions,\n  type CountedOccurrence,\n  type CrossFileDuplicateCandidate,\n  type CrossFileDuplicationFileData,\n  type SequenceWindowContext,\n  type TokenSegment,\n} from './duplication.js';\nimport type { DuplicationOptions } from './types.js';\n\nexport interface CrossFileDuplicationSourceFile extends Partial<CrossFileDuplicationFileData> {\n  file: string;\n  candidates: CrossFileDuplicateCandidate[];\n}\n\nexport interface CrossFileDuplicateOccurrence {\n  endLine: number;\n  file: string;\n  /**\n   * The code lines carrying the occurrence's matched tokens (see\n   * DuplicateBlockOccurrence.lineNumbers). Absent for a file that supplied only candidates (no\n   * `tokens`), whose matched lines are unknowable.\n   */\n  lineNumbers?: number[];\n  startLine: number;\n  /**\n   * The tokens of its matched runs, or of a near-miss block or its matched cores with the edited\n   * tokens among them; the group's `tokenCount` is the smallest of its occurrences'.\n   */\n  tokenCount: number;\n}\n\nexport interface CrossFileDuplicateBlockGroup {\n  files: string[];\n  occurrences: CrossFileDuplicateOccurrence[];\n  /**\n   * Token count of the smallest occurrence. For exact and gapped groups it is the matched token\n   * count every occurrence shares (gaps are not counted); for near-miss (Type-3) groups it is the\n   * token count of the smallest whole block or set of matched cores, edited tokens included.\n   */\n  tokenCount: number;\n}\n\nexport interface CrossFileDuplicationMetrics {\n  /** Number of redundant copies across all groups, counted per matched fragment like within-file. */\n  duplicateBlockCount: number;\n  /** Groups the file participates in, keyed by the file name passed in. */\n  duplicateBlockGroupCountByFile: Record<string, number>;\n  /**\n   * Per file, the 1-based code lines covered by the tokens of its cross-file occurrences, sorted\n   * ascending: the matched tokens of exact and gapped occurrences, and the lines of a near-miss\n   * block or its matched cores more than half of whose tokens its partners match (like within-file\n   * near-miss coverage). The unmatched gap of a merged\n   * clone and comment/blank lines inside an occurrence's bounding range are excluded (blank rows\n   * inside multi-row tokens only when the file supplied codeLineNumbers). A file that supplied\n   * only candidates (no `tokens`) has no entry — without its token stream the covered lines are\n   * unknowable, and an approximate bounding range would break this field's exactness.\n   */\n  duplicateLineNumbersByFile: Record<string, number[]>;\n  groups: CrossFileDuplicateBlockGroup[];\n}\n\ninterface SelectableCandidate extends CrossFileDuplicateCandidate, SelectableRegion {\n  regionBucket: number;\n  file: string;\n}\n\n/** A cross-file occurrence: a within-file occurrence in the project-wide token index space. */\ninterface CrossFileOccurrence extends CountedOccurrence {\n  file: string;\n  fileIndex: number;\n}\n\n/**\n * Detects code regions duplicated across files. Per-file candidates (whole block subtrees and full\n * container runs, fingerprinted with the same normalization as within-file duplication) are joined\n * by a project-level window index over per-statement fingerprint sequences (CPD-style), so a\n * copy-pasted partial statement run embedded in different surrounding code is matched even though\n * no single file can know it repeats elsewhere. Candidates are grouped by fingerprint, and only\n * maximal, non-overlapping regions whose group spans at least two files are counted. Groups that\n * shrink to a single file during selection are shed — a within-file repeat is already reported by\n * that file's own duplication metrics. A copy nested inside a larger group's region (two files share\n * a whole function, a third file only a block of it) is reported with its group, so the third\n * file's copy still shows what it duplicates. Groups separated by a small token gap within each file then\n * merge into gapped (Type-3) clone groups under `maxGapTokens`, exactly like within-file merging.\n * Finally, blocks of files that supplied `nearMissBlocks` are compared across files for near-miss\n * (Type-3) clones under `minSimilarityPercent` (see crossFileNearMiss.ts).\n */\nexport function measureCrossFileDuplication(\n  files: CrossFileDuplicationSourceFile[],\n  options?: DuplicationOptions\n): CrossFileDuplicationMetrics {\n  const { minTokens, maxGapTokens, minSimilarityPercent } = resolveDuplicationOptions(options);\n  const candidates: SelectableCandidate[] = files.flatMap(({ file, candidates }, fileIndex) =>\n    candidates.map((candidate) => ({ ...candidate, regionBucket: fileIndex, file }))\n  );\n  // Pushed one by one: spreading the project-scale window-candidate array as call arguments\n  // overflows V8's argument limit (~124k) and crashes on Node, though Bun/JSC tolerates it.\n  for (const candidate of collectWindowCandidates(files, minTokens)) {\n    candidates.push(candidate);\n  }\n  const counted = selectMaximalGroups(\n    candidates,\n    spansMultipleFiles,\n    // File index and position break coverage ties deterministically.\n    (left, right) => left.regionBucket - right.regionBucket || left.startIndex - right.startIndex\n  );\n  const tokenOffsets = computeTokenOffsets(files, maxGapTokens);\n  const groups = mergeGapAdjacentGroups([...counted.values()], tokenOffsets, maxGapTokens);\n  for (const group of collectNearMissGroups(files, groups, tokenOffsets, minTokens, minSimilarityPercent)) {\n    groups.push(group);\n  }\n  return summarize(groups, files, tokenOffsets);\n}\n\n/** Near-miss groups among the blocks the exact groups leave unreported, in the project token space. */\nfunction collectNearMissGroups(\n  files: CrossFileDuplicationSourceFile[],\n  exactGroups: CrossFileOccurrence[][],\n  tokenOffsets: number[],\n  minTokens: number,\n  minSimilarityPercent: number\n): CrossFileOccurrence[][] {\n  const reportedSpansByFile: { startTokenIndex: number; endTokenIndex: number }[][] = files.map(() => []);\n  for (const { fileIndex, startTokenIndex, endTokenIndex } of exactGroups.flat()) {\n    const offset = tokenOffsets[fileIndex] ?? 0;\n    reportedSpansByFile[fileIndex]?.push({\n      startTokenIndex: startTokenIndex - offset,\n      endTokenIndex: endTokenIndex - offset,\n    });\n  }\n  return collectCrossFileNearMissGroups(files, reportedSpansByFile, minTokens, minSimilarityPercent).map((group) =>\n    group.map((occurrence) => {\n      const offset = tokenOffsets[occurrence.fileIndex] ?? 0;\n      return {\n        ...occurrence,\n        file: files[occurrence.fileIndex]?.file ?? '',\n        segments: occurrence.segments.map((segment) => shift(segment, offset)),\n        matchedRuns: occurrence.matchedRuns?.map((run) => shift(run, offset)),\n        startTokenIndex: occurrence.startTokenIndex + offset,\n        endTokenIndex: occurrence.endTokenIndex + offset,\n      };\n    })\n  );\n}\n\nfunction shift({ startTokenIndex, endTokenIndex }: TokenSegment, offset: number): TokenSegment {\n  return { startTokenIndex: startTokenIndex + offset, endTokenIndex: endTokenIndex + offset };\n}\n\n/** Repeated sub-windows of sibling statements matched across the whole project's files. */\nfunction collectWindowCandidates(files: CrossFileDuplicationSourceFile[], minTokens: number): SelectableCandidate[] {\n  const fileIndexByContext: number[] = [];\n  const contexts: SequenceWindowContext[] = [];\n  for (const [fileIndex, { tokens, containerStatements }] of files.entries()) {\n    if (tokens && containerStatements) {\n      fileIndexByContext.push(fileIndex);\n      contexts.push({ tokens, literalCountPrefix: buildLiteralCountPrefix(tokens), containers: containerStatements });\n    }\n  }\n  if (contexts.length < 2) {\n    return [];\n  }\n  return collectSequenceWindowCandidates(contexts, minTokens, true).flatMap(({ candidate, contextIndex }) => {\n    const fileIndex = fileIndexByContext[contextIndex];\n    const file = fileIndex === undefined ? undefined : files[fileIndex];\n    return fileIndex === undefined || file === undefined\n      ? []\n      : [{ ...candidate, regionBucket: fileIndex, file: file.file }];\n  });\n}\n\n/** A merged group is reported only while it still covers more than one file. */\nfunction spansMultipleFilesAfterMerge(group: CrossFileOccurrence[]): boolean {\n  return new Set(group.map((occurrence) => occurrence.file)).size >= 2;\n}\n\nfunction spansMultipleFiles(group: SelectableCandidate[]): boolean {\n  return group.length >= 2 && new Set(group.map((candidate) => candidate.regionBucket)).size >= 2;\n}\n\n/**\n * Per-file token offsets that map every file into one project-wide token index space: each file's\n * tokens are offset by more than `maxGapTokens` past the previous file's, so occurrences in\n * different files are never gap-adjacent and merged pairs always stay within one file.\n */\nfunction computeTokenOffsets(files: CrossFileDuplicationSourceFile[], maxGapTokens: number): number[] {\n  const tokenOffsets: number[] = [];\n  let offset = 0;\n  for (const { tokens, candidates } of files) {\n    tokenOffsets.push(offset);\n    // Accumulated in a loop: spreading a project-scale candidate array as call arguments would\n    // overflow V8's argument limit (~124k) and crash on Node.\n    let tokenCount = tokens?.length ?? 0;\n    if (!tokens) {\n      for (const candidate of candidates) {\n        tokenCount = Math.max(tokenCount, candidate.endTokenIndex);\n      }\n    }\n    offset += tokenCount + maxGapTokens + 1;\n  }\n  return tokenOffsets;\n}\n\n/** Reuses the within-file gapped (Type-3) merging in the project-wide token index space. */\nfunction mergeGapAdjacentGroups(\n  groups: SelectableCandidate[][],\n  tokenOffsets: number[],\n  maxGapTokens: number\n): CrossFileOccurrence[][] {\n  const occurrenceGroups = groups.map((group) =>\n    group\n      .map((candidate): CrossFileOccurrence => {\n        const start = candidate.startTokenIndex + (tokenOffsets[candidate.regionBucket] ?? 0);\n        const end = candidate.endTokenIndex + (tokenOffsets[candidate.regionBucket] ?? 0);\n        return {\n          file: candidate.file,\n          fileIndex: candidate.regionBucket,\n          spanCountedElsewhere: candidate.nestedInLargerGroup,\n          nestedInLargerGroup: candidate.nestedInLargerGroup,\n          segments: [{ startTokenIndex: start, endTokenIndex: end }],\n          tokenCount: candidate.tokenCount,\n          startTokenIndex: start,\n          endTokenIndex: end,\n          startIndex: candidate.startIndex,\n          endIndex: candidate.endIndex,\n          startLine: candidate.startLine,\n          endLine: candidate.endLine,\n        };\n      })\n      .toSorted((left, right) => left.startTokenIndex - right.startTokenIndex)\n  );\n  return mergeAdjacentGroups(occurrenceGroups, maxGapTokens, spansMultipleFilesAfterMerge);\n}\n\nfunction summarize(\n  groups: CrossFileOccurrence[][],\n  files: CrossFileDuplicationSourceFile[],\n  tokenOffsets: number[]\n): CrossFileDuplicationMetrics {\n  const reported: CrossFileDuplicateBlockGroup[] = [];\n  // Accumulated in Maps: file names are arbitrary strings, and a plain object would read\n  // inherited properties for names like \"constructor\".\n  const groupCountByFile = new Map<string, number>();\n  const fileDataByName = new Map(\n    files.map((file, index) => [\n      file.file,\n      { tokens: file.tokens, codeLineNumbers: file.codeLineNumbers, offset: tokenOffsets[index] ?? 0 },\n    ])\n  );\n  const lineNumbersByFile = new Map<string, Set<number>>();\n  let duplicateBlockCount = 0;\n  for (const group of groups) {\n    // Mirrors within-file counting: each redundant occurrence contributes one count per matched\n    // fragment, gapped merging consolidates the grouping without halving the count, and spans a\n    // partial merge shares between a retained group and the merged group count once.\n    duplicateBlockCount += countRedundantFragments(group);\n    const occurrences = group\n      .map((occurrence) => {\n        const { file, startLine, endLine, tokenCount } = occurrence;\n        const lineNumbers = collectOccurrenceLines(occurrence, fileDataByName, lineNumbersByFile);\n        return { file, startLine, endLine, ...(lineNumbers && { lineNumbers }), tokenCount };\n      })\n      .toSorted((left, right) => left.file.localeCompare(right.file) || left.startLine - right.startLine);\n    const files = [...new Set(occurrences.map(({ file }) => file))];\n    for (const file of files) {\n      groupCountByFile.set(file, (groupCountByFile.get(file) ?? 0) + 1);\n    }\n    reported.push({ files, occurrences, tokenCount: Math.min(...group.map(({ tokenCount }) => tokenCount)) });\n  }\n  reported.sort(\n    (left, right) =>\n      right.tokenCount - left.tokenCount ||\n      (left.occurrences[0]?.file ?? '').localeCompare(right.occurrences[0]?.file ?? '') ||\n      (left.occurrences[0]?.startLine ?? 0) - (right.occurrences[0]?.startLine ?? 0)\n  );\n  return {\n    duplicateBlockCount,\n    duplicateBlockGroupCountByFile: Object.fromEntries(groupCountByFile),\n    duplicateLineNumbersByFile: Object.fromEntries(\n      [...lineNumbersByFile].map(([file, lines]) => [file, [...lines].toSorted((left, right) => left - right)])\n    ),\n    groups: reported,\n  };\n}\n\n/**\n * Adds the code lines an occurrence's matched tokens cover (see collectMatchedLines) to its file's\n * line set, mapping the project-wide token segments back into the file's own token stream, and\n * returns them in order.\n * A file that supplied only candidates (no token stream) is skipped rather than approximated from\n * the bounding line range, which would include gap and comment/blank lines and break the field's\n * exactness contract.\n */\nfunction collectOccurrenceLines(\n  occurrence: CrossFileOccurrence,\n  fileDataByName: Map<\n    string,\n    { tokens?: CrossFileDuplicationSourceFile['tokens']; codeLineNumbers?: Set<number>; offset: number }\n  >,\n  lineNumbersByFile: Map<string, Set<number>>\n): number[] | undefined {\n  const fileData = fileDataByName.get(occurrence.file);\n  if (!fileData?.tokens) {\n    return undefined;\n  }\n  let fileLines = lineNumbersByFile.get(occurrence.file);\n  if (!fileLines) {\n    fileLines = new Set();\n    lineNumbersByFile.set(occurrence.file, fileLines);\n  }\n  const lines = collectMatchedLines(\n    occurrence.segments.map((segment) => shift(segment, -fileData.offset)),\n    occurrence.matchedRuns?.map((run) => shift(run, -fileData.offset)),\n    fileData.tokens,\n    fileData.codeLineNumbers\n  );\n  for (const line of lines) {\n    fileLines.add(line);\n  }\n  return [...lines].toSorted((left, right) => left - right);\n}\n"],"mappings":"oVA+FA,SAAgB,EACd,EACA,EAC6B,CAC7B,GAAM,CAAE,YAAW,eAAc,wBAAyB,EAA0B,CAAO,EACrF,EAAoC,EAAM,SAAS,CAAE,OAAM,cAAc,IAC7E,EAAW,IAAK,IAAe,CAAE,GAAG,EAAW,aAAc,EAAW,MAAK,EAAE,CACjF,EAGA,IAAK,IAAM,KAAa,EAAwB,EAAO,CAAS,EAC9D,EAAW,KAAK,CAAS,EAE3B,IAAM,EAAU,EACd,EACA,GAEC,EAAM,IAAU,EAAK,aAAe,EAAM,cAAgB,EAAK,WAAa,EAAM,UACrF,EACM,EAAe,EAAoB,EAAO,CAAY,EACtD,EAAS,EAAuB,CAAC,GAAG,EAAQ,OAAO,CAAC,EAAG,EAAc,CAAY,EACvF,IAAK,IAAM,KAAS,EAAsB,EAAO,EAAQ,EAAc,EAAW,CAAoB,EACpG,EAAO,KAAK,CAAK,EAEnB,OAAO,EAAU,EAAQ,EAAO,CAAY,CAC9C,CAGA,SAAS,EACP,EACA,EACA,EACA,EACA,EACyB,CACzB,IAAM,EAA8E,EAAM,QAAU,CAAC,CAAC,EACtG,IAAK,GAAM,CAAE,YAAW,kBAAiB,mBAAmB,EAAY,KAAK,EAAG,CAC9E,IAAM,EAAS,EAAa,IAAc,EAC1C,EAAoB,EAAU,EAAE,KAAK,CACnC,gBAAiB,EAAkB,EACnC,cAAe,EAAgB,CACjC,CAAC,CACH,CACA,OAAO,EAA+B,EAAO,EAAqB,EAAW,CAAoB,CAAC,CAAC,IAAK,GACtG,EAAM,IAAK,GAAe,CACxB,IAAM,EAAS,EAAa,EAAW,YAAc,EACrD,MAAO,CACL,GAAG,EACH,KAAM,EAAM,EAAW,UAAU,EAAE,MAAQ,GAC3C,SAAU,EAAW,SAAS,IAAK,GAAY,EAAM,EAAS,CAAM,CAAC,EACrE,YAAa,EAAW,aAAa,IAAK,GAAQ,EAAM,EAAK,CAAM,CAAC,EACpE,gBAAiB,EAAW,gBAAkB,EAC9C,cAAe,EAAW,cAAgB,CAC5C,CACF,CAAC,CACH,CACF,CAEA,SAAS,EAAM,CAAE,kBAAiB,iBAA+B,EAA8B,CAC7F,MAAO,CAAE,gBAAiB,EAAkB,EAAQ,cAAe,EAAgB,CAAO,CAC5F,CAGA,SAAS,EAAwB,EAAyC,EAA0C,CAClH,IAAM,EAA+B,CAAC,EAChC,EAAoC,CAAC,EAC3C,IAAK,GAAM,CAAC,EAAW,CAAE,SAAQ,0BAA0B,EAAM,QAAQ,EACnE,GAAU,IACZ,EAAmB,KAAK,CAAS,EACjC,EAAS,KAAK,CAAE,SAAQ,mBAAoB,EAAwB,CAAM,EAAG,WAAY,CAAoB,CAAC,GAMlH,OAHI,EAAS,OAAS,EACb,CAAC,EAEH,EAAgC,EAAU,EAAW,EAAI,CAAC,CAAC,SAAS,CAAE,YAAW,kBAAmB,CACzG,IAAM,EAAY,EAAmB,GAC/B,EAAO,IAAc,IAAA,GAAY,IAAA,GAAY,EAAM,GACzD,OAAO,IAAc,IAAA,IAAa,IAAS,IAAA,GACvC,CAAC,EACD,CAAC,CAAE,GAAG,EAAW,aAAc,EAAW,KAAM,EAAK,IAAK,CAAC,CACjE,CAAC,CACH,CAGA,SAAS,EAA6B,EAAuC,CAC3E,OAAO,IAAI,IAAI,EAAM,IAAK,GAAe,EAAW,IAAI,CAAC,CAAC,CAAC,MAAQ,CACrE,CAEA,SAAS,EAAmB,EAAuC,CACjE,OAAO,EAAM,QAAU,GAAK,IAAI,IAAI,EAAM,IAAK,GAAc,EAAU,YAAY,CAAC,CAAC,CAAC,MAAQ,CAChG,CAOA,SAAS,EAAoB,EAAyC,EAAgC,CACpG,IAAM,EAAyB,CAAC,EAC5B,EAAS,EACb,IAAK,GAAM,CAAE,SAAQ,gBAAgB,EAAO,CAC1C,EAAa,KAAK,CAAM,EAGxB,IAAI,EAAa,GAAQ,QAAU,EACnC,GAAI,CAAC,EACH,IAAK,IAAM,KAAa,EACtB,EAAa,KAAK,IAAI,EAAY,EAAU,aAAa,EAG7D,GAAU,EAAa,EAAe,CACxC,CACA,OAAO,CACT,CAGA,SAAS,EACP,EACA,EACA,EACyB,CACzB,IAAM,EAAmB,EAAO,IAAK,GACnC,EACG,IAAK,GAAmC,CACvC,IAAM,EAAQ,EAAU,iBAAmB,EAAa,EAAU,eAAiB,GAC7E,EAAM,EAAU,eAAiB,EAAa,EAAU,eAAiB,GAC/E,MAAO,CACL,KAAM,EAAU,KAChB,UAAW,EAAU,aACrB,qBAAsB,EAAU,oBAChC,oBAAqB,EAAU,oBAC/B,SAAU,CAAC,CAAE,gBAAiB,EAAO,cAAe,CAAI,CAAC,EACzD,WAAY,EAAU,WACtB,gBAAiB,EACjB,cAAe,EACf,WAAY,EAAU,WACtB,SAAU,EAAU,SACpB,UAAW,EAAU,UACrB,QAAS,EAAU,OACrB,CACF,CAAC,CAAC,CACD,UAAU,EAAM,IAAU,EAAK,gBAAkB,EAAM,eAAe,CAC3E,EACA,OAAO,EAAoB,EAAkB,EAAc,CAA4B,CACzF,CAEA,SAAS,EACP,EACA,EACA,EAC6B,CAC7B,IAAM,EAA2C,CAAC,EAG5C,kBAAmB,IAAI,IACvB,EAAiB,IAAI,IACzB,EAAM,KAAK,EAAM,IAAU,CACzB,EAAK,KACL,CAAE,OAAQ,EAAK,OAAQ,gBAAiB,EAAK,gBAAiB,OAAQ,EAAa,IAAU,CAAE,CACjG,CAAC,CACH,EACM,kBAAoB,IAAI,IAC1B,EAAsB,EAC1B,IAAK,IAAM,KAAS,EAAQ,CAI1B,GAAuB,EAAwB,CAAK,EACpD,IAAM,EAAc,EACjB,IAAK,GAAe,CACnB,GAAM,CAAE,OAAM,YAAW,UAAS,cAAe,EAC3C,EAAc,EAAuB,EAAY,EAAgB,CAAiB,EACxF,MAAO,CAAE,OAAM,YAAW,UAAS,GAAI,GAAe,CAAE,aAAY,EAAI,YAAW,CACrF,CAAC,CAAC,CACD,UAAU,EAAM,IAAU,EAAK,KAAK,cAAc,EAAM,IAAI,GAAK,EAAK,UAAY,EAAM,SAAS,EAC9F,EAAQ,CAAC,GAAG,IAAI,IAAI,EAAY,KAAK,CAAE,UAAW,CAAI,CAAC,CAAC,EAC9D,IAAK,IAAM,KAAQ,EACjB,EAAiB,IAAI,GAAO,EAAiB,IAAI,CAAI,GAAK,GAAK,CAAC,EAElE,EAAS,KAAK,CAAE,QAAO,cAAa,WAAY,KAAK,IAAI,GAAG,EAAM,KAAK,CAAE,gBAAiB,CAAU,CAAC,CAAE,CAAC,CAC1G,CAOA,OANA,EAAS,MACN,EAAM,IACL,EAAM,WAAa,EAAK,aACvB,EAAK,YAAY,EAAE,EAAE,MAAQ,GAAA,CAAI,cAAc,EAAM,YAAY,EAAE,EAAE,MAAQ,EAAE,IAC/E,EAAK,YAAY,EAAE,EAAE,WAAa,IAAM,EAAM,YAAY,EAAE,EAAE,WAAa,EAChF,EACO,CACL,sBACA,+BAAgC,OAAO,YAAY,CAAgB,EACnE,2BAA4B,OAAO,YACjC,CAAC,GAAG,CAAiB,CAAC,CAAC,KAAK,CAAC,EAAM,KAAW,CAAC,EAAM,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAAC,CAAC,CAC1G,EACA,OAAQ,CACV,CACF,CAUA,SAAS,EACP,EACA,EAIA,EACsB,CACtB,IAAM,EAAW,EAAe,IAAI,EAAW,IAAI,EACnD,GAAI,CAAC,GAAU,OACb,OAEF,IAAI,EAAY,EAAkB,IAAI,EAAW,IAAI,EAChD,IACH,kBAAY,IAAI,IAChB,EAAkB,IAAI,EAAW,KAAM,CAAS,GAElD,IAAM,EAAQ,EACZ,EAAW,SAAS,IAAK,GAAY,EAAM,EAAS,CAAC,EAAS,MAAM,CAAC,EACrE,EAAW,aAAa,IAAK,GAAQ,EAAM,EAAK,CAAC,EAAS,MAAM,CAAC,EACjE,EAAS,OACT,EAAS,eACX,EACA,IAAK,IAAM,KAAQ,EACjB,EAAU,IAAI,CAAI,EAEpB,MAAO,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAC1D"}

package/dist/crossFileNearMiss.cjs

@@ -1,2 +1,2 @@
-"use strict";const e=require("./duplication.cjs");function t(e,t,c,l){if(l>=100)return[];let u=g(e),d=s(u,c,l),f=t.map(i),p=t.map(r),m=u.map(({fileIndex:e,range:t})=>p[e]?.(t)??!1),h=u.map(({fileIndex:e,range:t})=>f[e]?.(t)??!1),_=[];o(u,m,l,(e,t)=>{let n=u[e],r=u[t],i=n&&r&&d(n,r,t);if(i){if(i.kind===`whole`)h[e]&&h[t]||_.push([e,void 0,t,void 0]);else for(let[n,r]of i.cores)_.push([e,n,t,r])}});let v=u.map(()=>!1),y=u.map(()=>[]);for(let[e,t,n,r]of _)for(let[i,a]of[[e,t],[n,r]])a?y[i]?.push(a):v[i]=!0;let b=[],x=[];for(let e of u.keys()){x.push(b.length);let t=v[e]?[]:n(y[e]??[]);t.length===0&&b.push({blockIndex:e,core:void 0});for(let n of t)b.push({blockIndex:e,core:n})}let S=(e,t)=>{let n=x[e]??0;if(!t||v[e])return n;for(let r=n;b[r]?.blockIndex===e;r+=1){let e=b[r]?.core;if(e&&e[0]<=t[0]&&t[1]<=e[1])return r}throw Error(`every local core lies in one of its block's merged cores`)},C=b.map((e,t)=>t),w=e=>{let t=e;for(;C[t]!==t;)t=C[t]??t;for(let n=e;C[n]!==t;){let e=C[n]??t;C[n]=t,n=e}return t},T=b.map(({blockIndex:e,core:t})=>{let n=u[e],[r,i]=t??[n?.range.startTokenIndex??0,n?.range.endTokenIndex??0];return f[n?.fileIndex??0]?.({startTokenIndex:r,endTokenIndex:i})??!1});for(let[e,t,n,r]of _){let i=S(e,t),a=S(n,r);if(T[i]&&T[a])continue;let o=w(i),s=w(a);C[Math.max(o,s)]=Math.min(o,s)}let E=/* @__PURE__ */ new Map;for(let e of b.keys()){let t=w(e),n=E.get(t)??[];n.push(e),E.set(t,n)}let D=[];for(let t of E.values()){if(t.length<2||t.every(e=>T[e]))continue;let n=/* @__PURE__ */ new Map;for(let e of t){let{blockIndex:t=0,core:r}=b[e]??{},i=n.get(t)??{cores:[],anchor:!1};i.cores.push(r),i.anchor||=T[e]??!1,n.set(t,i)}D.push([...n].flatMap(([t,{cores:n,anchor:r}])=>{let i=u[t];return i?[a(i,e,n,r)]:[]}))}return D}function n(e){let t=[];for(let[n,r]of e.toSorted((e,t)=>e[0]-t[0])){let e=t.at(-1);e&&n<e[1]?e[1]=Math.max(e[1],r):t.push([n,r])}return t}function r(e){let t=[];for(let{startTokenIndex:n,endTokenIndex:r}of e.toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex)){let e=t.at(-1);e&&n<=e[1]?e[1]=Math.max(e[1],r):t.push([n,r])}return e=>{let n=0,r=t.length;for(;n<r;){let i=n+r>>>1;(t[i]?.[0]??0)<=e.startTokenIndex?n=i+1:r=i}return(t[n-1]?.[1]??-1)>=e.endTokenIndex}}function i(e){let t=e.toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex),n=new Int32Array(t.length),r=-1;for(let[e,i]of t.entries())r=Math.max(r,i.endTokenIndex),n[e]=r;return e=>{let r=0,i=t.length;for(;r<i;){let n=r+i>>>1;(t[n]?.startTokenIndex??0)<e.endTokenIndex?r=n+1:i=n}return r>0&&(n[r-1]??-1)>e.startTokenIndex}}function a({fileIndex:e,range:t},n,r,i){let a=r.includes(void 0),o=r.map(e=>e??[t.startTokenIndex,t.endTokenIndex]).toSorted((e,t)=>e[0]-t[0]).map(([e,t])=>({startTokenIndex:e,endTokenIndex:t})),s=o[0]?.startTokenIndex??t.startTokenIndex,c=o.at(-1)?.endTokenIndex??t.endTokenIndex,l=n[e]?.tokens;return{fileIndex:e,spanCountedElsewhere:i||void 0,segments:o,tokenCount:o.reduce((e,t)=>e+t.endTokenIndex-t.startTokenIndex,0),startTokenIndex:s,endTokenIndex:c,startIndex:t.startIndex,endIndex:t.endIndex,startLine:a?t.startLine:(l?.[s]?.startRow??0)+1,endLine:a?t.endLine:(l?.[c-1]?.endRow??0)+1}}function o(e,t,n,r){let i=/* @__PURE__ */ new Map;for(let t of e)for(let e of t.ngrams)i.set(e,(i.get(e)??0)+1);for(let t of e)t.ngrams=t.ngrams.filter(e=>(i.get(e)??0)<=1e3);let a=Int32Array.from(e,e=>e.fileIndex),o=Uint8Array.from(t,Number),s=Int32Array.from(e,e=>e.sequence.length),c=Int32Array.from(e,e=>e.ngrams.length),l=[...e.keys()].toSorted((e,t)=>(s[e]??0)-(s[t]??0)),u=/* @__PURE__ */ new Map,d=new Int32Array(e.length),f=[];for(let t of l){let i=a[t],l=o[t]===1,p=Math.min(Math.ceil((s[t]??0)/3),Math.ceil(n*(s[t]??0)/100)),m=e[t]?.ngrams??[];for(let e of m){let n=u.get(e);if(!n){u.set(e,[t]);continue}for(let e=n.length-1;e>=0;--e){let t=n[e]??0;if((s[t]??0)<p)break;a[t]===i||l&&o[t]===1||(d[t]===0&&f.push(t),d[t]=(d[t]??0)+1)}n.push(t)}for(let e of f){let n=d[e]??0;d[e]=0,n*100>=10*Math.min(c[e]??0,m.length)&&r(e,t)}f.length=0}}function s(t,n,r){let i=/* @__PURE__ */ new Map;for(let e of t)for(let t of e.contentCounts.keys())i.set(t,(i.get(t)??0)+1);let a=e=>Math.min(31-Math.clz32(Math.floor((t.length+1)/e))+1,3),o=/* @__PURE__ */ new Map;for(let[e,t]of i)o.set(e,a(t));let s=e=>{let t=Int32Array.from(e.keys()).toSorted(),n=Int32Array.from(t,t=>(e.get(t)??0)*(o.get(t)??0)),r=0;for(let e of n)r+=e;return{symbols:t,weightedCounts:n,total:r}},u=new Map(t.map(e=>[e,s(e.contentCounts)])),d=-1,g,_=(t,n,r)=>((d!==n||!g)&&(d=n,g=e.createLcsLengthCounter(t.sequence)),g(r)),b=(t,i)=>{let a=[];for(let e=0,n=0;e<t.uniqueNgrams.length&&n<i.uniqueNgrams.length;){let r=t.uniqueNgrams[e]??0,o=i.uniqueNgrams[n]??0;r===o&&a.push([t.uniqueNgramOffsets[e]??0,i.uniqueNgramOffsets[n]??0]),r<=o&&(e+=1),o<=r&&(n+=1)}a.sort((e,t)=>e[0]-t[0]);let o=a.filter(([e,t],n)=>{let r=a[n-1],i=a[n+1];return r?.[0]===e-1&&r[1]===t-1||i?.[0]===e+1&&i[1]===t+1}),l=t.range.startTokenIndex,u=i.range.startTokenIndex,d=[];for(let a of p(f(o))){let[o,f]=a[0]??[0,0],[p,h]=a.at(-1)??[0,0],g=p+5,_=h+5,b=g-o,x=_-f,S=Math.min(b,x),C=r*Math.max(b,x);S>=n&&S*100>=C&&m(a)*100>=50*S&&c(s(v(t.symbols,t.isContent,o,g)),s(v(i.symbols,i.isContent,f,_)))&&e.lcsLength(y(t.symbols.subarray(o,g)),y(i.symbols.subarray(f,_)))*100>=C&&d.push([[l+o,l+g],[u+f,u+_]])}return d.length>0?{kind:`local`,cores:d}:void 0};return(e,t,n)=>{let i=r*Math.max(e.sequence.length,t.sequence.length);return Math.min(e.sequence.length,t.sequence.length)*100>=i&&c(u.get(e),u.get(t))&&(h(e.sortedSequence,t.sortedSequence)*100>=i&&_(t,n,e.sequence)*100>=i||l(e,t,i))?{kind:`whole`}:b(e,t)}}function c(e,t){if(!e||!t)return!1;let n=0;for(let r=0,i=0;r<e.symbols.length&&i<t.symbols.length;){let a=e.symbols[r]??0,o=t.symbols[i]??0;a===o&&(n+=Math.min(e.weightedCounts[r]??0,t.weightedCounts[i]??0)),a<=o&&(r+=1),o<=a&&(i+=1)}return n*100>50*Math.max(e.total,t.total)}function l(t,n,r){return t.canonicalSequence!==void 0&&n.canonicalSequence!==void 0&&e.lcsLength(t.canonicalSequence,n.canonicalSequence)*100>=r}function u(e,t,n){if(t.length<2)return;let r=[],i=0;for(let[a,o]of t){let t=a-n;i<t&&r.push(y(e.subarray(i,t))),r.push(y(e.subarray(t,o-n))),i=o-n}i<e.length&&r.push(y(e.subarray(i))),r.sort(d);let a=new Int32Array(e.length),o=0;for(let e of r)a.set(e,o),o+=e.length;return a}function d(e,t){for(let n=0;n<Math.min(e.length,t.length);n+=1){let r=(e[n]??0)-(t[n]??0);if(r!==0)return r}return e.length-t.length}function f(e){let t=[],n=[];for(let[r,[,i]]of e.entries()){let a=0,o=t.length;for(;a<o;){let n=a+o>>>1;(e[t[n]??0]?.[1]??0)<i?a=n+1:o=n}n.push(a>0?t[a-1]??-1:-1),t[a]=r}let r=[];for(let i=t.at(-1)??-1;i>=0;i=n[i]??-1){let t=e[i];t&&r.push(t)}return r.toReversed()}function p(e){let t=[];for(let[n,r]of e.entries()){let i=e[n-1];i!==void 0&&r[0]-(i[0]+5)<=30&&r[1]-(i[1]+5)<=30?t.at(-1)?.push(r):t.push([r])}return t}function m(e){let t=5;for(let n=1;n<e.length;n+=1)t+=Math.min((e[n]?.[0]??0)-(e[n-1]?.[0]??0),5);return t}function h(e,t){let n=0,r=0,i=0;for(;r<e.length&&i<t.length;){let a=e[r]??0,o=t[i]??0;a===o?(n+=1,r+=1,i+=1):a<o?r+=1:i+=1}return n}function g(e){let t=/* @__PURE__ */ new Map,n=[];for(let[r,{tokens:i,containerStatements:a,nearMissBlocks:o}]of e.entries()){if(!i||!o?.length)continue;let e=new Int32Array(i.length),s=new Uint8Array(i.length),c=/* @__PURE__ */ new Map;for(let[n,r]of i.entries()){if(r.kind===`id`){let t=c.get(r.text);t===void 0&&(t=c.size,c.set(r.text,t)),e[n]=-(t+1);continue}let i=b(r),a=t.get(i);a===void 0&&(a=t.size,t.set(i,a)),e[n]=a,s[n]=r.isName||r.literalHash!==void 0?1:0}let l=_(a??[]);for(let t of o){let{startTokenIndex:i,endTokenIndex:a}=t,o=e.subarray(i,a),c=s.subarray(i,a),d=y(o),f=x(o),p=/* @__PURE__ */ new Map;for(let e of f)p.set(e,(p.get(e)??0)+1);let m=f.keys().filter(e=>p.get(f[e]??0)===1).toArray().toSorted((e,t)=>(f[e]??0)-(f[t]??0));n.push({fileIndex:r,range:t,symbols:o,isContent:c,sequence:d,sortedSequence:d.toSorted(),ngrams:Int32Array.from(p.keys()),uniqueNgrams:Int32Array.from(m,e=>f[e]??0),uniqueNgramOffsets:Int32Array.from(m),contentCounts:v(o,c,0,o.length),canonicalSequence:u(o,l(i,a),i)})}}return n}function _(e){let t=e.flat().filter(e=>e.startTokenIndex<e.endTokenIndex).map(e=>[e.startTokenIndex,e.endTokenIndex]).toSorted((e,t)=>e[0]-t[0]||t[1]-e[1]);return(e,n)=>{let r=0,i=t.length;for(;r<i;){let n=r+i>>>1;(t[n]?.[0]??0)<e?r=n+1:i=n}let a=[];for(let i=r;i<t.length;i+=1){let r=t[i];if(!r||r[0]>=n)break;let o=a.at(-1),s=o!==void 0&&r[0]<o[1];r[1]<=n&&(r[0]!==e||r[1]!==n)&&!s&&a.push(r)}return a}}function v(e,t,n,r){let i=/* @__PURE__ */ new Map;for(let a=n;a<r;a+=1)if(t[a]===1){let t=e[a]??0;i.set(t,(i.get(t)??0)+1)}return i}function y(e){let t=/* @__PURE__ */ new Map;return e.map(e=>{if(e>=0)return e;let n=t.get(e);return n===void 0&&(n=t.size,t.set(e,n)),-(n+1)})}function b(e){let t=e.textHash^Math.imul(e.literalHash??0,2654435761),n=e.textHash2^Math.imul(e.literalHash2??0,2246822507);return(t>>>0)*2097152+(n>>>11)}function x(e){let t=new Int32Array(Math.max(e.length-5+1,0));for(let n=0;n<t.length;n+=1){let r=5381;for(let t=0;t<5;t+=1){let i=e[n+t]??0;r=Math.imul(r,31)+(i<0?-1:i)|0}t[n]=r}return t}exports.collectCrossFileNearMissGroups=t;
+"use strict";const e=require("./duplication.cjs");function t(e,t,i,a){if(a>=100)return[];let o=x(e),s=p(o,i,a),m=t.map(u),h=t.map(l),g=o.map(({fileIndex:e,range:t})=>h[e]?.(t)??!1),_=o.map(({fileIndex:e,range:t})=>m[e]?.(t)??!1),v=({fileIndex:e},[t,n])=>m[e]?.({startTokenIndex:t,endTokenIndex:n})??!1,y=[];f(o,g,a,(e,t)=>{let n=o[e],r=o[t],i=n&&r&&s(n,r,t);if(i){if(!i.cores){_[e]&&_[t]||y.push({left:e,right:t,reordered:i.reordered});return}for(let{pair:a,reordered:o}of i.cores)v(n,a[0])&&v(r,a[1])||y.push({left:e,right:t,cores:a,reordered:o})}});let b=o.map(()=>!1),S=o.map(()=>[]);for(let e of y)for(let[t,r]of n(e))r?S[t]?.push(r):b[t]=!0;let C=[],w=[];for(let e of o.keys()){w.push(C.length);let t=b[e]?[]:c(S[e]??[]);t.length===0&&C.push({blockIndex:e,core:void 0});for(let n of t)C.push({blockIndex:e,core:n})}let T=(e,t)=>{let n=w[e]??0;if(!t||b[e])return n;for(let r=n;C[r]?.blockIndex===e;r+=1){let e=C[r]?.core;if(e&&e[0]<=t[0]&&t[1]<=e[1])return r}throw Error(`every local core lies in one of its block's merged cores`)},E=C.map((e,t)=>t),D=e=>{let t=e;for(;E[t]!==t;)t=E[t]??t;for(let n=e;E[n]!==t;){let e=E[n]??t;E[n]=t,n=e}return t},O=C.map(({blockIndex:e,core:t})=>{let n=o[e],[r,i]=t??[n?.range.startTokenIndex??0,n?.range.endTokenIndex??0];return m[n?.fileIndex??0]?.({startTokenIndex:r,endTokenIndex:i})??!1}),k=y.map(e=>n(e).map(([e,t])=>T(e,t)));for(let[e=0,t=0]of k){let n=D(e),r=D(t);E[Math.max(n,r)]=Math.min(n,r)}let A=r(o,y,k,C.length),j=/* @__PURE__ */ new Map;for(let e of C.keys()){let t=D(e),n=j.get(t)??[];n.push(e),j.set(t,n)}let M=[];for(let t of j.values()){if(t.length<2||t.every(e=>O[e]))continue;let n=/* @__PURE__ */ new Map;for(let e of t){let{blockIndex:t=0,core:r}=C[e]??{},i=n.get(t)??{cores:[],anchor:!1,matchedRuns:[]};i.cores.push(r),i.anchor||=O[e]??!1,i.matchedRuns.push(...A[e]??[]),n.set(t,i)}M.push([...n].flatMap(([t,n])=>{let r=o[t];return r?[d(r,e,n)]:[]}))}return M}function n({left:e,right:t,cores:n}){return[[e,n?.[0]],[t,n?.[1]]]}function r(e,t,n,r){let a=Array.from({length:r},()=>[]);for(let[r,o]of t.entries()){let t=e[o.left],s=e[o.right];if(t&&s){let[e,c]=i(t,s,o);a[n[r]?.[0]??0]?.push(...e),a[n[r]?.[1]??0]?.push(...c)}}return a}function i(t,n,{cores:r,reordered:i}){let c=s(t,r?.[0],i),l=s(n,r?.[1],i);if(g(a(t,r?.[0]),a(n,r?.[1]))<=0){let[t,n]=e.markLcs(c.sequence,l.sequence);return[o(c.positions,t),o(l.positions,n)]}let[u,d]=e.markLcs(l.sequence,c.sequence);return[o(c.positions,d),o(l.positions,u)]}function a(e,t){if(!t)return e.writtenContent;let{sequence:n,positions:r}=s(e,t,!1);return Float64Array.from(n,(t,n)=>t<0?t:e.fileTokenKeys[r[n]??0]??0)}function o(e,t){let n=[];for(let r of e.filter((e,n)=>t[n]===1).toSorted()){let e=n.at(-1);e?.[1]===r?e[1]=r+1:n.push([r,r+1])}return n}function s(e,t,n){let r=e.range.startTokenIndex;return t?{sequence:w(e.symbols.subarray(t[0]-r,t[1]-r)),positions:Int32Array.from({length:t[1]-t[0]},(e,n)=>t[0]+n)}:n&&e.canonical?{sequence:e.canonical.sequence,positions:e.canonical.offsets.map(e=>r+e)}:{sequence:e.sequence,positions:e.sequence.map((e,t)=>r+t)}}function c(e){let t=[];for(let[n,r]of e.toSorted((e,t)=>e[0]-t[0])){let e=t.at(-1);e&&n<e[1]?e[1]=Math.max(e[1],r):t.push([n,r])}return t}function l(e){let t=[];for(let{startTokenIndex:n,endTokenIndex:r}of e.toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex)){let e=t.at(-1);e&&n<=e[1]?e[1]=Math.max(e[1],r):t.push([n,r])}return e=>{let n=0,r=t.length;for(;n<r;){let i=n+r>>>1;(t[i]?.[0]??0)<=e.startTokenIndex?n=i+1:r=i}return(t[n-1]?.[1]??-1)>=e.endTokenIndex}}function u(e){let t=e.toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex),n=new Int32Array(t.length),r=-1;for(let[e,i]of t.entries())r=Math.max(r,i.endTokenIndex),n[e]=r;return e=>{let r=0,i=t.length;for(;r<i;){let n=r+i>>>1;(t[n]?.startTokenIndex??0)<e.endTokenIndex?r=n+1:i=n}return r>0&&(n[r-1]??-1)>e.startTokenIndex}}function d({fileIndex:e,range:t},n,{cores:r,anchor:i,matchedRuns:a}){let o=r.includes(void 0),s=r.map(e=>e??[t.startTokenIndex,t.endTokenIndex]).toSorted((e,t)=>e[0]-t[0]).map(([e,t])=>({startTokenIndex:e,endTokenIndex:t})),l=s[0]?.startTokenIndex??t.startTokenIndex,u=s.at(-1)?.endTokenIndex??t.endTokenIndex,d=n[e]?.tokens;return{fileIndex:e,spanCountedElsewhere:i||void 0,segments:s,matchedRuns:c(a).map(([e,t])=>({startTokenIndex:e,endTokenIndex:t})),tokenCount:s.reduce((e,t)=>e+t.endTokenIndex-t.startTokenIndex,0),startTokenIndex:l,endTokenIndex:u,startIndex:t.startIndex,endIndex:t.endIndex,startLine:o?t.startLine:(d?.[l]?.startRow??0)+1,endLine:o?t.endLine:(d?.[u-1]?.endRow??0)+1}}function f(e,t,n,r){let i=/* @__PURE__ */ new Map;for(let t of e)for(let e of t.ngrams)i.set(e,(i.get(e)??0)+1);for(let t of e)t.ngrams=t.ngrams.filter(e=>(i.get(e)??0)<=1e3);let a=Int32Array.from(e,e=>e.fileIndex),o=Uint8Array.from(t,Number),s=Int32Array.from(e,e=>e.sequence.length),c=Int32Array.from(e,e=>e.ngrams.length),l=[...e.keys()].toSorted((e,t)=>(s[e]??0)-(s[t]??0)),u=/* @__PURE__ */ new Map,d=new Int32Array(e.length),f=[];for(let t of l){let i=a[t],l=o[t]===1,p=Math.min(Math.ceil((s[t]??0)/3),Math.ceil(n*(s[t]??0)/100)),m=e[t]?.ngrams??[];for(let e of m){let n=u.get(e);if(!n){u.set(e,[t]);continue}for(let e=n.length-1;e>=0;--e){let t=n[e]??0;if((s[t]??0)<p)break;a[t]===i||l&&o[t]===1||(d[t]===0&&f.push(t),d[t]=(d[t]??0)+1)}n.push(t)}for(let e of f){let n=d[e]??0;d[e]=0,n*100>=10*Math.min(c[e]??0,m.length)&&r(e,t)}f.length=0}}function p(t,n,r){let i=/* @__PURE__ */ new Map;for(let e of t)for(let t of e.contentCounts.keys())i.set(t,(i.get(t)??0)+1);let a=e=>Math.min(31-Math.clz32(Math.floor((t.length+1)/e))+1,3),o=/* @__PURE__ */ new Map;for(let[e,t]of i)o.set(e,a(t));let s=e=>{let t=Int32Array.from(e.keys()).toSorted(),n=Int32Array.from(t,t=>(e.get(t)??0)*(o.get(t)??0)),r=0;for(let e of n)r+=e;return{symbols:t,weightedCounts:n,total:r}},c=new Map(t.map(e=>[e,s(e.contentCounts)])),l=-1,u,d=(t,n,r)=>((l!==n||!u)&&(l=n,u=e.createLcsLengthCounter(t.sequence)),u(r)),f=(t,i)=>{let a=[];for(let e=0,n=0;e<t.uniqueNgrams.length&&n<i.uniqueNgrams.length;){let r=t.uniqueNgrams[e]??0,o=i.uniqueNgrams[n]??0;r===o&&a.push([t.uniqueNgramOffsets[e]??0,i.uniqueNgramOffsets[n]??0]),r<=o&&(e+=1),o<=r&&(n+=1)}a.sort((e,t)=>e[0]-t[0]);let o=a.filter(([e,t],n)=>{let r=a[n-1],i=a[n+1];return r?.[0]===e-1&&r[1]===t-1||i?.[0]===e+1&&i[1]===t+1}),c=t.range.startTokenIndex,l=i.range.startTokenIndex,u=[];for(let a of v(_(o))){let[o,d]=a[0]??[0,0],[f,p]=a.at(-1)??[0,0],h=f+5,g=p+5,_=h-o,v=g-d,b=Math.min(_,v),x=r*Math.max(_,v);b<n||b*100<x||y(a)*100<50*b||!m(s(C(t.symbols,t.isContent,o,h)),s(C(i.symbols,i.isContent,d,g)))||e.lcsLength(w(t.symbols.subarray(o,h)),w(i.symbols.subarray(d,g)))*100>=x&&u.push({pair:[[c+o,c+h],[l+d,l+g]],reordered:!1})}return u.length>0?{cores:u}:void 0};return(t,n,i)=>{let a=r*Math.max(t.sequence.length,n.sequence.length);if(Math.min(t.sequence.length,n.sequence.length)*100>=a&&m(c.get(t),c.get(n))){if(b(t.sortedSequence,n.sortedSequence)*100>=a&&d(n,i,t.sequence)*100>=a)return{reordered:!1};if(t.canonical&&n.canonical&&e.lcsLength(t.canonical.sequence,n.canonical.sequence)*100>=a)return{reordered:!0}}if(g(n.writtenContent,t.writtenContent)<0){let e=f(n,t);return e?.cores&&{cores:e.cores.map(({pair:[e,t],...n})=>({...n,pair:[t,e]}))}}return f(t,n)}}function m(e,t){if(!e||!t)return!1;let n=0;for(let r=0,i=0;r<e.symbols.length&&i<t.symbols.length;){let a=e.symbols[r]??0,o=t.symbols[i]??0;a===o&&(n+=Math.min(e.weightedCounts[r]??0,t.weightedCounts[i]??0)),a<=o&&(r+=1),o<=a&&(i+=1)}return n*100>50*Math.max(e.total,t.total)}function h(e,t,n,r){if(n.length<2)return;let i=[],a=(n,r)=>{let a=w(e.subarray(n,r)),o=Float64Array.from(a,(e,r)=>e<0?e:t[n+r]??0);i.push({unit:a,content:o,start:n})},o=0;for(let[e,t]of n){let n=e-r;o<n&&a(o,n),a(n,t-r),o=t-r}o<e.length&&a(o,e.length),i.sort((e,t)=>g(e.content,t.content)||e.start-t.start);let s=new Int32Array(e.length),c=new Int32Array(e.length),l=0;for(let{unit:e,start:t}of i){s.set(e,l);for(let n=0;n<e.length;n+=1)c[l+n]=t+n;l+=e.length}return{sequence:s,offsets:c}}function g(e,t){for(let n=0;n<Math.min(e.length,t.length);n+=1){let r=(e[n]??0)-(t[n]??0);if(r!==0)return r}return e.length-t.length}function _(e){let t=[],n=[];for(let[r,[,i]]of e.entries()){let a=0,o=t.length;for(;a<o;){let n=a+o>>>1;(e[t[n]??0]?.[1]??0)<i?a=n+1:o=n}n.push(a>0?t[a-1]??-1:-1),t[a]=r}let r=[];for(let i=t.at(-1)??-1;i>=0;i=n[i]??-1){let t=e[i];t&&r.push(t)}return r.toReversed()}function v(e){let t=[];for(let[n,r]of e.entries()){let i=e[n-1];i!==void 0&&r[0]-(i[0]+5)<=30&&r[1]-(i[1]+5)<=30?t.at(-1)?.push(r):t.push([r])}return t}function y(e){let t=5;for(let n=1;n<e.length;n+=1)t+=Math.min((e[n]?.[0]??0)-(e[n-1]?.[0]??0),5);return t}function b(e,t){let n=0,r=0,i=0;for(;r<e.length&&i<t.length;){let a=e[r]??0,o=t[i]??0;a===o?(n+=1,r+=1,i+=1):a<o?r+=1:i+=1}return n}function x(e){let t=/* @__PURE__ */ new Map,n=[];for(let[r,{tokens:i,containerStatements:a,nearMissBlocks:o}]of e.entries()){if(!i||!o?.length)continue;let e=new Int32Array(i.length),s=new Uint8Array(i.length),c=new Float64Array(i.length),l=/* @__PURE__ */ new Map;for(let[n,r]of i.entries()){if(r.kind===`id`){let t=l.get(r.text);t===void 0&&(t=l.size,l.set(r.text,t)),e[n]=-(t+1);continue}let i=T(r);c[n]=i;let a=t.get(i);a===void 0&&(a=t.size,t.set(i,a)),e[n]=a,s[n]=r.isName||r.literalHash!==void 0?1:0}let u=S(a??[]);for(let t of o){let{startTokenIndex:i,endTokenIndex:a}=t,o=e.subarray(i,a),l=s.subarray(i,a),d=w(o),f=E(o,c.subarray(i,a)),p=/* @__PURE__ */ new Map;for(let e of f)p.set(e,(p.get(e)??0)+1);let m=f.keys().filter(e=>p.get(f[e]??0)===1).toArray().toSorted((e,t)=>(f[e]??0)-(f[t]??0));n.push({fileIndex:r,range:t,symbols:o,isContent:l,fileTokenKeys:c,sequence:d,writtenContent:Float64Array.from(d,(e,t)=>e<0?e:c[i+t]??0),sortedSequence:d.toSorted(),ngrams:Int32Array.from(p.keys()),uniqueNgrams:Int32Array.from(m,e=>f[e]??0),uniqueNgramOffsets:Int32Array.from(m),contentCounts:C(o,l,0,o.length),canonical:h(o,c.subarray(i,a),u(i,a),i)})}}return n}function S(e){let t=e.flat().filter(e=>e.startTokenIndex<e.endTokenIndex).map(e=>[e.startTokenIndex,e.endTokenIndex]).toSorted((e,t)=>e[0]-t[0]||t[1]-e[1]);return(e,n)=>{let r=0,i=t.length;for(;r<i;){let n=r+i>>>1;(t[n]?.[0]??0)<e?r=n+1:i=n}let a=[];for(let i=r;i<t.length;i+=1){let r=t[i];if(!r||r[0]>=n)break;let o=a.at(-1),s=o!==void 0&&r[0]<o[1];r[1]<=n&&(r[0]!==e||r[1]!==n)&&!s&&a.push(r)}return a}}function C(e,t,n,r){let i=/* @__PURE__ */ new Map;for(let a=n;a<r;a+=1)if(t[a]===1){let t=e[a]??0;i.set(t,(i.get(t)??0)+1)}return i}function w(e){let t=/* @__PURE__ */ new Map;return e.map(e=>{if(e>=0)return e;let n=t.get(e);return n===void 0&&(n=t.size,t.set(e,n)),-(n+1)})}function T(e){let t=e.textHash^Math.imul(e.literalHash??0,2654435761),n=e.textHash2^Math.imul(e.literalHash2??0,2246822507);return(t>>>0)*2097152+(n>>>11)}function E(e,t){let n=new Int32Array(Math.max(e.length-5+1,0));for(let r=0;r<n.length;r+=1){let i=5381;for(let n=0;n<5;n+=1){let a=t[r+n]??0,o=(e[r+n]??0)<0?-1:a^Math.floor(a/4294967296);i=Math.imul(i,31)+o|0}n[r]=i}return n}exports.collectCrossFileNearMissGroups=t;
 //# sourceMappingURL=crossFileNearMiss.cjs.map