code-gauge 10.2.1 → 10.2.3

package/README.md

@@ -445,7 +445,8 @@ The `duplication` section tunes how clones are detected:
   `require`s carry no tokens, since every module must spell out its own), with adjacent matches around a small edit merged into gapped (Type-3) clone groups and
   near-miss (Type-3) clones matched by token-LCS similarity (tolerating reordered statements and
   copies embedded in added code), each copy with its line span, its matched lines, and its matched token count,
-  plus duplicated line count and ratio
+  plus duplicated line count and ratio; code standing on its own that copies only a part of larger
+  copies is listed with that part of each of them
 - Cross-file duplication (via `measureCrossFileDuplication`): copy-pasted blocks shared between
   files, matched with the same normalization (exact, gapped, and near-miss clones) and reported as
   groups with the file location, the matched lines, and the matched token count of each copy
@@ -510,7 +511,7 @@ does not skip generated files (e.g. `.d.ts`, `.min.js`) or test files.
 ### Cloudflare Workers
 
 Workers cannot load native addons, so the package's `workerd` export runs the same API on a
-WebAssembly build of the Rust engine (about 2.3 MB gzipped). Wrangler resolves that export
+WebAssembly build of the Rust engine (about 3.5 MB gzipped). Wrangler resolves that export
 automatically, and no `nodejs_compat` flag is needed:
 
 ```ts

package/THIRD-PARTY-NOTICES.txt

@@ -7859,39 +7859,6 @@ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR
 IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
 DEALINGS IN THE SOFTWARE.
 
-====================================================================================================
-tree-sitter-go 0.25.0 — MIT — https://github.com/tree-sitter/tree-sitter-go
-====================================================================================================
------ LICENSE -----
-The MIT License (MIT)
-
-Copyright (c) 2014 Max Brunsfeld
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
-
-====================================================================================================
-tree-sitter-java 0.23.5 — MIT — https://github.com/tree-sitter/tree-sitter-java
-====================================================================================================
-This crate's published sources bundle no root license file; the upstream notice is:
-Copyright (c) 2017 Ayman Nadeem
-The canonical text of the MIT license is reproduced in the appendix at the end of this document.
-
 ====================================================================================================
 tree-sitter-language 0.1.8 — MIT — https://github.com/tree-sitter/tree-sitter
 ====================================================================================================
@@ -7918,39 +7885,6 @@ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 SOFTWARE.
 
-====================================================================================================
-tree-sitter-python 0.25.0 — MIT — https://github.com/tree-sitter/tree-sitter-python
-====================================================================================================
------ LICENSE -----
-The MIT License (MIT)
-
-Copyright (c) 2016 Max Brunsfeld
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
-
-====================================================================================================
-tree-sitter-ruby 0.23.1 — MIT — https://github.com/tree-sitter/tree-sitter-ruby
-====================================================================================================
-This crate's published sources bundle no root license file; the upstream notice is:
-Copyright (c) 2016 Rob Rix
-The canonical text of the MIT license is reproduced in the appendix at the end of this document.
-
 ====================================================================================================
 unicode-ident 1.0.24 — (MIT OR Apache-2.0) AND Unicode-3.0 — https://github.com/dtolnay/unicode-ident
 ====================================================================================================
@@ -9150,6 +9084,60 @@ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 SOFTWARE.
 
+====================================================================================================
+willbooster-tree-sitter-go 1.0.0 — MIT — https://github.com/WillBooster/tree-sitter-go
+====================================================================================================
+----- LICENSE -----
+The MIT License (MIT)
+
+Copyright (c) 2014 Max Brunsfeld
+Copyright (c) 2026 WillBooster Inc.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+====================================================================================================
+willbooster-tree-sitter-java 1.0.0 — MIT — https://github.com/WillBooster/tree-sitter-java
+====================================================================================================
+----- LICENSE -----
+MIT License
+
+Copyright (c) 2017 Ayman Nadeem
+Copyright (c) 2026 WillBooster Inc.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
 ====================================================================================================
 willbooster-tree-sitter-javascript 4.0.2 — MIT — https://github.com/WillBooster/tree-sitter-javascript
 ====================================================================================================
@@ -9204,6 +9192,60 @@ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 SOFTWARE.
 
+====================================================================================================
+willbooster-tree-sitter-python 1.0.1 — MIT — https://github.com/WillBooster/tree-sitter-python
+====================================================================================================
+----- LICENSE -----
+The MIT License (MIT)
+
+Copyright (c) 2016 Max Brunsfeld
+Copyright (c) 2026 WillBooster Inc.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+====================================================================================================
+willbooster-tree-sitter-ruby 1.0.0 — MIT — https://github.com/WillBooster/tree-sitter-ruby
+====================================================================================================
+----- LICENSE -----
+The MIT License (MIT)
+
+Copyright (c) 2016 Rob Rix
+Copyright (c) 2026 WillBooster Inc.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
 ====================================================================================================
 willbooster-tree-sitter-rust 2.0.2 — MIT — https://github.com/WillBooster/tree-sitter-rust
 ====================================================================================================

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:a,maxGapTokens:o,minSimilarityPercent:s}=e.resolveDuplicationOptions(r),l=t.flatMap(({file:e,candidates:t},n)=>t.map(t=>({...t,regionBucket:n,file:e})));for(let e of u(t,a))l.push(e);let d=n.selectMaximalGroups(l,f,(e,t)=>e.regionBucket-t.regionBucket||e.startIndex-t.startIndex),g=p(t,o),_=m([...d.values()],g,o);return h(i(_,c(t,_,g,a,s),a),t,g)}function i(e,t,n){let r=/* @__PURE__ */ new Map;for(let[t,n]of e.entries())for(let{fileIndex:e}of n){let n=r.get(e)??/* @__PURE__ */ new Set;n.add(t),r.set(e,n)}let i=/* @__PURE__ */ new Set,s=/* @__PURE__ */ new Set,c=t.map(t=>{let c=e=>{let r=t.filter(t=>t.segments.some(t=>e.segments.some(e=>a(e,t)))),[i]=r;if(r.length===1&&i)return Math.max(0,i.startTokenIndex-e.startTokenIndex)+Math.max(0,e.endTokenIndex-i.endTokenIndex)<n?i:void 0},l=/* @__PURE__ */ new Map,u=/* @__PURE__ */ new Set;for(let e of t)for(let t of e.spanCountedElsewhere?r.get(e.fileIndex)??[]:[])u.add(t);for(let t of[...u].toSorted((e,t)=>e-t)){let n=e[t]??[],r=n.map(c);if(!(i.has(t)||r.includes(void 0)||new Set(r).size!==r.length)){i.add(t);for(let[e,t]of r.entries()){let r=n[e];if(t&&r){let e=l.get(t)??[];e.push(r),l.set(t,e)}}}}return t.map(e=>{let t=l.get(e);if(!t)return e;let n=o([e,...t]);return s.add(n),n})});for(let t of s)for(let n of r.get(t.fileIndex)??[])if(!i.has(n)&&e[n]?.some(e=>!e.spanCountedElsewhere&&t.segments.every(t=>e.segments.some(e=>e.startTokenIndex<=t.startTokenIndex&&t.endTokenIndex<=e.endTokenIndex)))){t.spanCountedElsewhere=!0;break}return[...e.filter((e,t)=>!i.has(t)),...c]}function a(e,t){return e.startTokenIndex<t.endTokenIndex&&t.startTokenIndex<e.endTokenIndex}function o(e){let t=s(e.flatMap(e=>e.segments));return{...e[0],spanCountedElsewhere:void 0,nestedInLargerGroup:void 0,segments:t,matchedRuns:s(e.flatMap(e=>e.matchedRuns??e.segments)),exactSegments:s(e.flatMap(e=>e.matchedRuns?e.exactSegments??[]:e.segments)),tokenCount:t.reduce((e,t)=>e+t.endTokenIndex-t.startTokenIndex,0),startTokenIndex:Math.min(...e.map(e=>e.startTokenIndex)),endTokenIndex:Math.max(...e.map(e=>e.endTokenIndex)),startIndex:Math.min(...e.map(e=>e.startIndex)),endIndex:Math.max(...e.map(e=>e.endIndex)),startLine:Math.min(...e.map(e=>e.startLine)),endLine:Math.max(...e.map(e=>e.endLine))}}function s(e){let t=[];for(let n of e.toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex)){let e=t.at(-1);e&&n.startTokenIndex<e.endTokenIndex?e.endTokenIndex=Math.max(e.endTokenIndex,n.endTokenIndex):t.push({...n})}return t}function c(e,n,r,i,a){let o=e.map(()=>[]);for(let{fileIndex:e,segments:t}of n.flat()){let n=r[e]??0;for(let r of t)o[e]?.push(l(r,-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=>l(e,n)),matchedRuns:t.matchedRuns?.map(e=>l(e,n)),startTokenIndex:t.startTokenIndex+n,endTokenIndex:t.endTokenIndex+n}}))}function l({startTokenIndex:e,endTokenIndex:t},n){return{startTokenIndex:e+n,endTokenIndex:t+n}}function u(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 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(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,d)}function h(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=g(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 g(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));let o=e.collectMatchedLines(t.segments.map(e=>l(e,-i.offset)),t.matchedRuns?.map(e=>l(e,-i.offset)),i.tokens,i.codeLineNumbers,t.exactSegments?.map(e=>l(e,-i.offset)));for(let e of o)a.add(e);return[...o].toSorted((e,t)=>e-t)}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:o,minSimilarityPercent:s}=e.resolveDuplicationOptions(r),l=t.flatMap(({file:e,candidates:t},n)=>t.map(t=>({...t,regionBucket:n,file:e})));for(let e of u(t,a))l.push(e);let d=n.selectMaximalGroups(l,f,(e,t)=>e.regionBucket-t.regionBucket||e.startIndex-t.startIndex),g=p(t,o),_=m([...d.values()],g,o);return h(i(_,c(t,_,g,a,s),a),t,g)}function i(e,t,n){let r=/* @__PURE__ */ new Map;for(let[t,n]of e.entries())for(let{fileIndex:e}of n){let n=r.get(e)??/* @__PURE__ */ new Set;n.add(t),r.set(e,n)}let i=/* @__PURE__ */ new Set,s=/* @__PURE__ */ new Set,c=t.map(t=>{let c=e=>{let r=t.filter(t=>t.segments.some(t=>e.segments.some(e=>a(e,t)))),[i]=r;if(r.length===1&&i)return Math.max(0,i.startTokenIndex-e.startTokenIndex)+Math.max(0,e.endTokenIndex-i.endTokenIndex)<n?i:void 0},l=/* @__PURE__ */ new Map,u=/* @__PURE__ */ new Set;for(let e of t)for(let t of e.spanCountedElsewhere?r.get(e.fileIndex)??[]:[])u.add(t);for(let t of[...u].toSorted((e,t)=>e-t)){let n=e[t]??[],r=n.map(c);if(!(i.has(t)||r.includes(void 0)||new Set(r).size!==r.length)){i.add(t);for(let[e,t]of r.entries()){let r=n[e];if(t&&r){let e=l.get(t)??[];e.push(r),l.set(t,e)}}}}return t.map(e=>{let t=l.get(e);if(!t)return e;let n=o([e,...t]);return s.add(n),n})});for(let t of s)for(let n of r.get(t.fileIndex)??[])if(!i.has(n)&&e[n]?.some(e=>!e.spanCountedElsewhere&&t.segments.every(t=>e.segments.some(e=>e.startTokenIndex<=t.startTokenIndex&&t.endTokenIndex<=e.endTokenIndex)))){t.spanCountedElsewhere=!0;break}return[...e.filter((e,t)=>!i.has(t)),...c]}function a(e,t){return e.startTokenIndex<t.endTokenIndex&&t.startTokenIndex<e.endTokenIndex}function o(e){let t=s(e.flatMap(e=>e.segments));return{...e[0],spanCountedElsewhere:void 0,nestedInLargerGroup:void 0,segments:t,matchedRuns:s(e.flatMap(e=>e.matchedRuns??e.segments)),exactSegments:s(e.flatMap(e=>e.matchedRuns?e.exactSegments??[]:e.segments)),tokenCount:t.reduce((e,t)=>e+t.endTokenIndex-t.startTokenIndex,0),startTokenIndex:Math.min(...e.map(e=>e.startTokenIndex)),endTokenIndex:Math.max(...e.map(e=>e.endTokenIndex)),startIndex:Math.min(...e.map(e=>e.startIndex)),endIndex:Math.max(...e.map(e=>e.endIndex)),startLine:Math.min(...e.map(e=>e.startLine)),endLine:Math.max(...e.map(e=>e.endLine))}}function s(e){let t=[];for(let n of e.toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex)){let e=t.at(-1);e&&n.startTokenIndex<e.endTokenIndex?e.endTokenIndex=Math.max(e.endTokenIndex,n.endTokenIndex):t.push({...n})}return t}function c(e,n,r,i,a){let o=t=>{let n=e.map(()=>[]);for(let{fileIndex:e,segments:i}of t.flat()){let t=r[e]??0;for(let r of i)n[e]?.push(l(r,-t))}return n};return t.collectCrossFileNearMissGroups(e,o(n),o(n.filter(e=>!e.some(e=>e.nestedInLargerGroup))),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=>l(e,n)),matchedRuns:t.matchedRuns?.map(e=>l(e,n)),startTokenIndex:t.startTokenIndex+n,endTokenIndex:t.endTokenIndex+n}}))}function l({startTokenIndex:e,endTokenIndex:t},n){return{startTokenIndex:e+n,endTokenIndex:t+n}}function u(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 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(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,d)}function h(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=g(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 g(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));let o=e.collectMatchedLines(t.segments.map(e=>l(e,-i.offset)),t.matchedRuns?.map(e=>l(e,-i.offset)),i.tokens,i.codeLineNumbers,t.exactSegments?.map(e=>l(e,-i.offset)));for(let e of o)a.add(e);return[...o].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","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 or which a part of it\n   * matched exactly reaches (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 exactGroups = mergeGapAdjacentGroups([...counted.values()], tokenOffsets, maxGapTokens);\n  const nearMissGroups = collectNearMissGroups(files, exactGroups, tokenOffsets, minTokens, minSimilarityPercent);\n  return summarize(takeOverExactGroups(exactGroups, nearMissGroups, minTokens), files, tokenOffsets);\n}\n\n/**\n * Lets a near-miss group take over the exact groups standing for the same copies, as within a\n * file (collect_near_miss_groups in native/src/duplication.rs): every occurrence of such a group\n * overlaps exactly one copy of the near-miss group (one overlapping none reports content the copies\n * do not share, and one reaching into two would make one copy take over lines of the next), exceeds\n * its span by less than `minTokens` (a copy is listed at the size it matched, not at that of a\n * clone larger by a reportable part, which may run into code no copy matched), and no two overlap\n * the same copy (those repeat within the copy, not between the copies).\n * Each copy then becomes one occurrence with the exact fragments it holds, so the copies are\n * listed once.\n */\nfunction takeOverExactGroups(\n  exactGroups: CrossFileOccurrence[][],\n  nearMissGroups: CrossFileOccurrence[][],\n  minTokens: number\n): CrossFileOccurrence[][] {\n  const exactGroupIndexesByFile = new Map<number, Set<number>>();\n  for (const [groupIndex, group] of exactGroups.entries()) {\n    for (const { fileIndex } of group) {\n      const indexes = exactGroupIndexesByFile.get(fileIndex) ?? new Set();\n      indexes.add(groupIndex);\n      exactGroupIndexesByFile.set(fileIndex, indexes);\n    }\n  }\n  const taken = new Set<number>();\n  const coalesced = new Set<CrossFileOccurrence>();\n  const merged = nearMissGroups.map((copies) => {\n    const copyOf = (occurrence: CrossFileOccurrence): CrossFileOccurrence | undefined => {\n      const overlapped = copies.filter((copy) =>\n        copy.segments.some((segment) => occurrence.segments.some((matched) => overlap(matched, segment)))\n      );\n      const [copy] = overlapped;\n      if (overlapped.length !== 1 || !copy) {\n        return undefined;\n      }\n      const excess =\n        Math.max(0, copy.startTokenIndex - occurrence.startTokenIndex) +\n        Math.max(0, occurrence.endTokenIndex - copy.endTokenIndex);\n      return excess < minTokens ? copy : undefined;\n    };\n    const fragmentsByCopy = new Map<CrossFileOccurrence, CrossFileOccurrence[]>();\n    const candidateIndexes = new Set<number>();\n    for (const copy of copies) {\n      for (const groupIndex of copy.spanCountedElsewhere ? (exactGroupIndexesByFile.get(copy.fileIndex) ?? []) : []) {\n        candidateIndexes.add(groupIndex);\n      }\n    }\n    for (const groupIndex of [...candidateIndexes].toSorted((left, right) => left - right)) {\n      const group = exactGroups[groupIndex] ?? [];\n      const owners = group.map(copyOf);\n      if (taken.has(groupIndex) || owners.includes(undefined) || new Set(owners).size !== owners.length) {\n        continue;\n      }\n      taken.add(groupIndex);\n      for (const [index, owner] of owners.entries()) {\n        const fragment = group[index];\n        if (owner && fragment) {\n          const fragments = fragmentsByCopy.get(owner) ?? [];\n          fragments.push(fragment);\n          fragmentsByCopy.set(owner, fragments);\n        }\n      }\n    }\n    return copies.map((copy): CrossFileOccurrence => {\n      const fragments = fragmentsByCopy.get(copy);\n      if (!fragments) {\n        return copy;\n      }\n      const whole = coalesceOccurrences([copy, ...fragments]);\n      coalesced.add(whole);\n      return whole;\n    });\n  });\n  // A copy that took its fragments over counts them itself, unless a group left in place still\n  // counts all of it: a larger clone whose matched segments enclose those of the copy, which its\n  // fragments were nested in. A smaller run left in place inside the copy counts its own repeats,\n  // not the copy.\n  for (const copy of coalesced) {\n    for (const groupIndex of exactGroupIndexesByFile.get(copy.fileIndex) ?? []) {\n      const enclosed =\n        !taken.has(groupIndex) &&\n        exactGroups[groupIndex]?.some(\n          (occurrence) =>\n            !occurrence.spanCountedElsewhere &&\n            copy.segments.every((segment) =>\n              occurrence.segments.some(\n                (matched) =>\n                  matched.startTokenIndex <= segment.startTokenIndex && segment.endTokenIndex <= matched.endTokenIndex\n              )\n            )\n        );\n      if (enclosed) {\n        copy.spanCountedElsewhere = true;\n        break;\n      }\n    }\n  }\n  return [...exactGroups.filter((_, groupIndex) => !taken.has(groupIndex)), ...merged];\n}\n\nfunction overlap(left: TokenSegment, right: TokenSegment): boolean {\n  return left.startTokenIndex < right.endTokenIndex && right.startTokenIndex < left.endTokenIndex;\n}\n\n/** The parts of one copy as one occurrence: an exact part is matched throughout, a near-miss part where its partners match it. */\nfunction coalesceOccurrences(parts: [CrossFileOccurrence, ...CrossFileOccurrence[]]): CrossFileOccurrence {\n  const segments = mergeSegments(parts.flatMap((part) => part.segments));\n  return {\n    ...parts[0],\n    spanCountedElsewhere: undefined,\n    nestedInLargerGroup: undefined,\n    segments,\n    matchedRuns: mergeSegments(parts.flatMap((part) => part.matchedRuns ?? part.segments)),\n    exactSegments: mergeSegments(\n      parts.flatMap((part) => (part.matchedRuns ? (part.exactSegments ?? []) : part.segments))\n    ),\n    tokenCount: segments.reduce((sum, segment) => sum + segment.endTokenIndex - segment.startTokenIndex, 0),\n    startTokenIndex: Math.min(...parts.map((part) => part.startTokenIndex)),\n    endTokenIndex: Math.max(...parts.map((part) => part.endTokenIndex)),\n    startIndex: Math.min(...parts.map((part) => part.startIndex)),\n    endIndex: Math.max(...parts.map((part) => part.endIndex)),\n    startLine: Math.min(...parts.map((part) => part.startLine)),\n    endLine: Math.max(...parts.map((part) => part.endLine)),\n  };\n}\n\n/** The unions of overlapping segments, in position order. */\nfunction mergeSegments(segments: TokenSegment[]): TokenSegment[] {\n  const merged: TokenSegment[] = [];\n  for (const segment of segments.toSorted((left, right) => left.startTokenIndex - right.startTokenIndex)) {\n    const last = merged.at(-1);\n    if (last && segment.startTokenIndex < last.endTokenIndex) {\n      last.endTokenIndex = Math.max(last.endTokenIndex, segment.endTokenIndex);\n    } else {\n      merged.push({ ...segment });\n    }\n  }\n  return merged;\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  // Segment by segment: the gap of a gapped clone is not reported content.\n  for (const { fileIndex, segments } of exactGroups.flat()) {\n    const offset = tokenOffsets[fileIndex] ?? 0;\n    for (const segment of segments) {\n      reportedSpansByFile[fileIndex]?.push(shift(segment, -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    occurrence.exactSegments?.map((segment) => shift(segment, -fileData.offset))\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,EAAc,EAAuB,CAAC,GAAG,EAAQ,OAAO,CAAC,EAAG,EAAc,CAAY,EAE5F,OAAO,EAAU,EAAoB,EADd,EAAsB,EAAO,EAAa,EAAc,EAAW,CAC3B,EAAG,CAAS,EAAG,EAAO,CAAY,CACnG,CAaA,SAAS,EACP,EACA,EACA,EACyB,CACzB,IAAM,kBAA0B,IAAI,IACpC,IAAK,GAAM,CAAC,EAAY,KAAU,EAAY,QAAQ,EACpD,IAAK,GAAM,CAAE,eAAe,EAAO,CACjC,IAAM,EAAU,EAAwB,IAAI,CAAS,mBAAK,IAAI,IAC9D,EAAQ,IAAI,CAAU,EACtB,EAAwB,IAAI,EAAW,CAAO,CAChD,CAEF,IAAM,kBAAQ,IAAI,IACZ,kBAAY,IAAI,IAChB,EAAS,EAAe,IAAK,GAAW,CAC5C,IAAM,EAAU,GAAqE,CACnF,IAAM,EAAa,EAAO,OAAQ,GAChC,EAAK,SAAS,KAAM,GAAY,EAAW,SAAS,KAAM,GAAY,EAAQ,EAAS,CAAO,CAAC,CAAC,CAClG,EACM,CAAC,GAAQ,EACX,KAAW,SAAW,GAAM,EAMhC,OAFE,KAAK,IAAI,EAAG,EAAK,gBAAkB,EAAW,eAAe,EAC7D,KAAK,IAAI,EAAG,EAAW,cAAgB,EAAK,aAAa,EAC3C,EAAY,EAAO,IAAA,EACrC,EACM,kBAAkB,IAAI,IACtB,kBAAmB,IAAI,IAC7B,IAAK,IAAM,KAAQ,EACjB,IAAK,IAAM,KAAc,EAAK,qBAAwB,EAAwB,IAAI,EAAK,SAAS,GAAK,CAAC,EAAK,CAAC,EAC1G,EAAiB,IAAI,CAAU,EAGnC,IAAK,IAAM,IAAc,CAAC,GAAG,CAAgB,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,EAAG,CACtF,IAAM,EAAQ,EAAY,IAAe,CAAC,EACpC,EAAS,EAAM,IAAI,CAAM,EAC3B,OAAM,IAAI,CAAU,GAAK,EAAO,SAAS,IAAA,EAAS,GAAK,IAAI,IAAI,CAAM,CAAC,CAAC,OAAS,EAAO,QAG3F,GAAM,IAAI,CAAU,EACpB,IAAK,GAAM,CAAC,EAAO,KAAU,EAAO,QAAQ,EAAG,CAC7C,IAAM,EAAW,EAAM,GACvB,GAAI,GAAS,EAAU,CACrB,IAAM,EAAY,EAAgB,IAAI,CAAK,GAAK,CAAC,EACjD,EAAU,KAAK,CAAQ,EACvB,EAAgB,IAAI,EAAO,CAAS,CACtC,CACF,CARoB,CAStB,CACA,OAAO,EAAO,IAAK,GAA8B,CAC/C,IAAM,EAAY,EAAgB,IAAI,CAAI,EAC1C,GAAI,CAAC,EACH,OAAO,EAET,IAAM,EAAQ,EAAoB,CAAC,EAAM,GAAG,CAAS,CAAC,EAEtD,OADA,EAAU,IAAI,CAAK,EACZ,CACT,CAAC,CACH,CAAC,EAKD,IAAK,IAAM,KAAQ,EACjB,IAAK,IAAM,KAAc,EAAwB,IAAI,EAAK,SAAS,GAAK,CAAC,EAavE,GAXE,CAAC,EAAM,IAAI,CAAU,GACrB,EAAY,EAAW,EAAE,KACtB,GACC,CAAC,EAAW,sBACZ,EAAK,SAAS,MAAO,GACnB,EAAW,SAAS,KACjB,GACC,EAAQ,iBAAmB,EAAQ,iBAAmB,EAAQ,eAAiB,EAAQ,aAC3F,CACF,CACJ,EACY,CACZ,EAAK,qBAAuB,GAC5B,KACF,CAGJ,MAAO,CAAC,GAAG,EAAY,QAAQ,EAAG,IAAe,CAAC,EAAM,IAAI,CAAU,CAAC,EAAG,GAAG,CAAM,CACrF,CAEA,SAAS,EAAQ,EAAoB,EAA8B,CACjE,OAAO,EAAK,gBAAkB,EAAM,eAAiB,EAAM,gBAAkB,EAAK,aACpF,CAGA,SAAS,EAAoB,EAA6E,CACxG,IAAM,EAAW,EAAc,EAAM,QAAS,GAAS,EAAK,QAAQ,CAAC,EACrE,MAAO,CACL,GAAG,EAAM,GACT,qBAAsB,IAAA,GACtB,oBAAqB,IAAA,GACrB,WACA,YAAa,EAAc,EAAM,QAAS,GAAS,EAAK,aAAe,EAAK,QAAQ,CAAC,EACrF,cAAe,EACb,EAAM,QAAS,GAAU,EAAK,YAAe,EAAK,eAAiB,CAAC,EAAK,EAAK,QAAS,CACzF,EACA,WAAY,EAAS,QAAQ,EAAK,IAAY,EAAM,EAAQ,cAAgB,EAAQ,gBAAiB,CAAC,EACtG,gBAAiB,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,eAAe,CAAC,EACtE,cAAe,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,aAAa,CAAC,EAClE,WAAY,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,UAAU,CAAC,EAC5D,SAAU,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,QAAQ,CAAC,EACxD,UAAW,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,SAAS,CAAC,EAC1D,QAAS,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,OAAO,CAAC,CACxD,CACF,CAGA,SAAS,EAAc,EAA0C,CAC/D,IAAM,EAAyB,CAAC,EAChC,IAAK,IAAM,KAAW,EAAS,UAAU,EAAM,IAAU,EAAK,gBAAkB,EAAM,eAAe,EAAG,CACtG,IAAM,EAAO,EAAO,GAAG,EAAE,EACrB,GAAQ,EAAQ,gBAAkB,EAAK,cACzC,EAAK,cAAgB,KAAK,IAAI,EAAK,cAAe,EAAQ,aAAa,EAEvE,EAAO,KAAK,CAAE,GAAG,CAAQ,CAAC,CAE9B,CACA,OAAO,CACT,CAGA,SAAS,EACP,EACA,EACA,EACA,EACA,EACyB,CACzB,IAAM,EAA8E,EAAM,QAAU,CAAC,CAAC,EAEtG,IAAK,GAAM,CAAE,YAAW,cAAc,EAAY,KAAK,EAAG,CACxD,IAAM,EAAS,EAAa,IAAc,EAC1C,IAAK,IAAM,KAAW,EACpB,EAAoB,EAAU,EAAE,KAAK,EAAM,EAAS,CAAC,CAAM,CAAC,CAEhE,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,gBACT,EAAW,eAAe,IAAK,GAAY,EAAM,EAAS,CAAC,EAAS,MAAM,CAAC,CAC7E,EACA,IAAK,IAAM,KAAQ,EACjB,EAAU,IAAI,CAAI,EAEpB,MAAO,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAC1D"}
+{"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 or which a part of it\n   * matched exactly reaches (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 exactGroups = mergeGapAdjacentGroups([...counted.values()], tokenOffsets, maxGapTokens);\n  const nearMissGroups = collectNearMissGroups(files, exactGroups, tokenOffsets, minTokens, minSimilarityPercent);\n  return summarize(takeOverExactGroups(exactGroups, nearMissGroups, minTokens), files, tokenOffsets);\n}\n\n/**\n * Lets a near-miss group take over the exact groups standing for the same copies, as within a\n * file (collect_near_miss_groups in native/src/duplication.rs): every occurrence of such a group\n * overlaps exactly one copy of the near-miss group (one overlapping none reports content the copies\n * do not share, and one reaching into two would make one copy take over lines of the next), exceeds\n * its span by less than `minTokens` (a copy is listed at the size it matched, not at that of a\n * clone larger by a reportable part, which may run into code no copy matched), and no two overlap\n * the same copy (those repeat within the copy, not between the copies).\n * Each copy then becomes one occurrence with the exact fragments it holds, so the copies are\n * listed once.\n */\nfunction takeOverExactGroups(\n  exactGroups: CrossFileOccurrence[][],\n  nearMissGroups: CrossFileOccurrence[][],\n  minTokens: number\n): CrossFileOccurrence[][] {\n  const exactGroupIndexesByFile = new Map<number, Set<number>>();\n  for (const [groupIndex, group] of exactGroups.entries()) {\n    for (const { fileIndex } of group) {\n      const indexes = exactGroupIndexesByFile.get(fileIndex) ?? new Set();\n      indexes.add(groupIndex);\n      exactGroupIndexesByFile.set(fileIndex, indexes);\n    }\n  }\n  const taken = new Set<number>();\n  const coalesced = new Set<CrossFileOccurrence>();\n  const merged = nearMissGroups.map((copies) => {\n    const copyOf = (occurrence: CrossFileOccurrence): CrossFileOccurrence | undefined => {\n      const overlapped = copies.filter((copy) =>\n        copy.segments.some((segment) => occurrence.segments.some((matched) => overlap(matched, segment)))\n      );\n      const [copy] = overlapped;\n      if (overlapped.length !== 1 || !copy) {\n        return undefined;\n      }\n      const excess =\n        Math.max(0, copy.startTokenIndex - occurrence.startTokenIndex) +\n        Math.max(0, occurrence.endTokenIndex - copy.endTokenIndex);\n      return excess < minTokens ? copy : undefined;\n    };\n    const fragmentsByCopy = new Map<CrossFileOccurrence, CrossFileOccurrence[]>();\n    const candidateIndexes = new Set<number>();\n    for (const copy of copies) {\n      for (const groupIndex of copy.spanCountedElsewhere ? (exactGroupIndexesByFile.get(copy.fileIndex) ?? []) : []) {\n        candidateIndexes.add(groupIndex);\n      }\n    }\n    for (const groupIndex of [...candidateIndexes].toSorted((left, right) => left - right)) {\n      const group = exactGroups[groupIndex] ?? [];\n      const owners = group.map(copyOf);\n      if (taken.has(groupIndex) || owners.includes(undefined) || new Set(owners).size !== owners.length) {\n        continue;\n      }\n      taken.add(groupIndex);\n      for (const [index, owner] of owners.entries()) {\n        const fragment = group[index];\n        if (owner && fragment) {\n          const fragments = fragmentsByCopy.get(owner) ?? [];\n          fragments.push(fragment);\n          fragmentsByCopy.set(owner, fragments);\n        }\n      }\n    }\n    return copies.map((copy): CrossFileOccurrence => {\n      const fragments = fragmentsByCopy.get(copy);\n      if (!fragments) {\n        return copy;\n      }\n      const whole = coalesceOccurrences([copy, ...fragments]);\n      coalesced.add(whole);\n      return whole;\n    });\n  });\n  // A copy that took its fragments over counts them itself, unless a group left in place still\n  // counts all of it: a larger clone whose matched segments enclose those of the copy, which its\n  // fragments were nested in. A smaller run left in place inside the copy counts its own repeats,\n  // not the copy.\n  for (const copy of coalesced) {\n    for (const groupIndex of exactGroupIndexesByFile.get(copy.fileIndex) ?? []) {\n      const enclosed =\n        !taken.has(groupIndex) &&\n        exactGroups[groupIndex]?.some(\n          (occurrence) =>\n            !occurrence.spanCountedElsewhere &&\n            copy.segments.every((segment) =>\n              occurrence.segments.some(\n                (matched) =>\n                  matched.startTokenIndex <= segment.startTokenIndex && segment.endTokenIndex <= matched.endTokenIndex\n              )\n            )\n        );\n      if (enclosed) {\n        copy.spanCountedElsewhere = true;\n        break;\n      }\n    }\n  }\n  return [...exactGroups.filter((_, groupIndex) => !taken.has(groupIndex)), ...merged];\n}\n\nfunction overlap(left: TokenSegment, right: TokenSegment): boolean {\n  return left.startTokenIndex < right.endTokenIndex && right.startTokenIndex < left.endTokenIndex;\n}\n\n/** The parts of one copy as one occurrence: an exact part is matched throughout, a near-miss part where its partners match it. */\nfunction coalesceOccurrences(parts: [CrossFileOccurrence, ...CrossFileOccurrence[]]): CrossFileOccurrence {\n  const segments = mergeSegments(parts.flatMap((part) => part.segments));\n  return {\n    ...parts[0],\n    spanCountedElsewhere: undefined,\n    nestedInLargerGroup: undefined,\n    segments,\n    matchedRuns: mergeSegments(parts.flatMap((part) => part.matchedRuns ?? part.segments)),\n    exactSegments: mergeSegments(\n      parts.flatMap((part) => (part.matchedRuns ? (part.exactSegments ?? []) : part.segments))\n    ),\n    tokenCount: segments.reduce((sum, segment) => sum + segment.endTokenIndex - segment.startTokenIndex, 0),\n    startTokenIndex: Math.min(...parts.map((part) => part.startTokenIndex)),\n    endTokenIndex: Math.max(...parts.map((part) => part.endTokenIndex)),\n    startIndex: Math.min(...parts.map((part) => part.startIndex)),\n    endIndex: Math.max(...parts.map((part) => part.endIndex)),\n    startLine: Math.min(...parts.map((part) => part.startLine)),\n    endLine: Math.max(...parts.map((part) => part.endLine)),\n  };\n}\n\n/** The unions of overlapping segments, in position order. */\nfunction mergeSegments(segments: TokenSegment[]): TokenSegment[] {\n  const merged: TokenSegment[] = [];\n  for (const segment of segments.toSorted((left, right) => left.startTokenIndex - right.startTokenIndex)) {\n    const last = merged.at(-1);\n    if (last && segment.startTokenIndex < last.endTokenIndex) {\n      last.endTokenIndex = Math.max(last.endTokenIndex, segment.endTokenIndex);\n    } else {\n      merged.push({ ...segment });\n    }\n  }\n  return merged;\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  // Segment by segment: the gap of a gapped clone is not reported content.\n  const spansByFile = (groups: CrossFileOccurrence[][]): TokenSegment[][] => {\n    const spans: TokenSegment[][] = files.map(() => []);\n    for (const { fileIndex, segments } of groups.flat()) {\n      const offset = tokenOffsets[fileIndex] ?? 0;\n      for (const segment of segments) {\n        spans[fileIndex]?.push(shift(segment, -offset));\n      }\n    }\n    return spans;\n  };\n  return collectCrossFileNearMissGroups(\n    files,\n    spansByFile(exactGroups),\n    spansByFile(exactGroups.filter((group) => !group.some((occurrence) => occurrence.nestedInLargerGroup))),\n    minTokens,\n    minSimilarityPercent\n  ).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    occurrence.exactSegments?.map((segment) => shift(segment, -fileData.offset))\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,EAAc,EAAuB,CAAC,GAAG,EAAQ,OAAO,CAAC,EAAG,EAAc,CAAY,EAE5F,OAAO,EAAU,EAAoB,EADd,EAAsB,EAAO,EAAa,EAAc,EAAW,CAC3B,EAAG,CAAS,EAAG,EAAO,CAAY,CACnG,CAaA,SAAS,EACP,EACA,EACA,EACyB,CACzB,IAAM,kBAA0B,IAAI,IACpC,IAAK,GAAM,CAAC,EAAY,KAAU,EAAY,QAAQ,EACpD,IAAK,GAAM,CAAE,eAAe,EAAO,CACjC,IAAM,EAAU,EAAwB,IAAI,CAAS,mBAAK,IAAI,IAC9D,EAAQ,IAAI,CAAU,EACtB,EAAwB,IAAI,EAAW,CAAO,CAChD,CAEF,IAAM,kBAAQ,IAAI,IACZ,kBAAY,IAAI,IAChB,EAAS,EAAe,IAAK,GAAW,CAC5C,IAAM,EAAU,GAAqE,CACnF,IAAM,EAAa,EAAO,OAAQ,GAChC,EAAK,SAAS,KAAM,GAAY,EAAW,SAAS,KAAM,GAAY,EAAQ,EAAS,CAAO,CAAC,CAAC,CAClG,EACM,CAAC,GAAQ,EACX,KAAW,SAAW,GAAM,EAMhC,OAFE,KAAK,IAAI,EAAG,EAAK,gBAAkB,EAAW,eAAe,EAC7D,KAAK,IAAI,EAAG,EAAW,cAAgB,EAAK,aAAa,EAC3C,EAAY,EAAO,IAAA,EACrC,EACM,kBAAkB,IAAI,IACtB,kBAAmB,IAAI,IAC7B,IAAK,IAAM,KAAQ,EACjB,IAAK,IAAM,KAAc,EAAK,qBAAwB,EAAwB,IAAI,EAAK,SAAS,GAAK,CAAC,EAAK,CAAC,EAC1G,EAAiB,IAAI,CAAU,EAGnC,IAAK,IAAM,IAAc,CAAC,GAAG,CAAgB,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,EAAG,CACtF,IAAM,EAAQ,EAAY,IAAe,CAAC,EACpC,EAAS,EAAM,IAAI,CAAM,EAC3B,OAAM,IAAI,CAAU,GAAK,EAAO,SAAS,IAAA,EAAS,GAAK,IAAI,IAAI,CAAM,CAAC,CAAC,OAAS,EAAO,QAG3F,GAAM,IAAI,CAAU,EACpB,IAAK,GAAM,CAAC,EAAO,KAAU,EAAO,QAAQ,EAAG,CAC7C,IAAM,EAAW,EAAM,GACvB,GAAI,GAAS,EAAU,CACrB,IAAM,EAAY,EAAgB,IAAI,CAAK,GAAK,CAAC,EACjD,EAAU,KAAK,CAAQ,EACvB,EAAgB,IAAI,EAAO,CAAS,CACtC,CACF,CARoB,CAStB,CACA,OAAO,EAAO,IAAK,GAA8B,CAC/C,IAAM,EAAY,EAAgB,IAAI,CAAI,EAC1C,GAAI,CAAC,EACH,OAAO,EAET,IAAM,EAAQ,EAAoB,CAAC,EAAM,GAAG,CAAS,CAAC,EAEtD,OADA,EAAU,IAAI,CAAK,EACZ,CACT,CAAC,CACH,CAAC,EAKD,IAAK,IAAM,KAAQ,EACjB,IAAK,IAAM,KAAc,EAAwB,IAAI,EAAK,SAAS,GAAK,CAAC,EAavE,GAXE,CAAC,EAAM,IAAI,CAAU,GACrB,EAAY,EAAW,EAAE,KACtB,GACC,CAAC,EAAW,sBACZ,EAAK,SAAS,MAAO,GACnB,EAAW,SAAS,KACjB,GACC,EAAQ,iBAAmB,EAAQ,iBAAmB,EAAQ,eAAiB,EAAQ,aAC3F,CACF,CACJ,EACY,CACZ,EAAK,qBAAuB,GAC5B,KACF,CAGJ,MAAO,CAAC,GAAG,EAAY,QAAQ,EAAG,IAAe,CAAC,EAAM,IAAI,CAAU,CAAC,EAAG,GAAG,CAAM,CACrF,CAEA,SAAS,EAAQ,EAAoB,EAA8B,CACjE,OAAO,EAAK,gBAAkB,EAAM,eAAiB,EAAM,gBAAkB,EAAK,aACpF,CAGA,SAAS,EAAoB,EAA6E,CACxG,IAAM,EAAW,EAAc,EAAM,QAAS,GAAS,EAAK,QAAQ,CAAC,EACrE,MAAO,CACL,GAAG,EAAM,GACT,qBAAsB,IAAA,GACtB,oBAAqB,IAAA,GACrB,WACA,YAAa,EAAc,EAAM,QAAS,GAAS,EAAK,aAAe,EAAK,QAAQ,CAAC,EACrF,cAAe,EACb,EAAM,QAAS,GAAU,EAAK,YAAe,EAAK,eAAiB,CAAC,EAAK,EAAK,QAAS,CACzF,EACA,WAAY,EAAS,QAAQ,EAAK,IAAY,EAAM,EAAQ,cAAgB,EAAQ,gBAAiB,CAAC,EACtG,gBAAiB,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,eAAe,CAAC,EACtE,cAAe,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,aAAa,CAAC,EAClE,WAAY,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,UAAU,CAAC,EAC5D,SAAU,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,QAAQ,CAAC,EACxD,UAAW,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,SAAS,CAAC,EAC1D,QAAS,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,OAAO,CAAC,CACxD,CACF,CAGA,SAAS,EAAc,EAA0C,CAC/D,IAAM,EAAyB,CAAC,EAChC,IAAK,IAAM,KAAW,EAAS,UAAU,EAAM,IAAU,EAAK,gBAAkB,EAAM,eAAe,EAAG,CACtG,IAAM,EAAO,EAAO,GAAG,EAAE,EACrB,GAAQ,EAAQ,gBAAkB,EAAK,cACzC,EAAK,cAAgB,KAAK,IAAI,EAAK,cAAe,EAAQ,aAAa,EAEvE,EAAO,KAAK,CAAE,GAAG,CAAQ,CAAC,CAE9B,CACA,OAAO,CACT,CAGA,SAAS,EACP,EACA,EACA,EACA,EACA,EACyB,CAEzB,IAAM,EAAe,GAAsD,CACzE,IAAM,EAA0B,EAAM,QAAU,CAAC,CAAC,EAClD,IAAK,GAAM,CAAE,YAAW,cAAc,EAAO,KAAK,EAAG,CACnD,IAAM,EAAS,EAAa,IAAc,EAC1C,IAAK,IAAM,KAAW,EACpB,EAAM,EAAU,EAAE,KAAK,EAAM,EAAS,CAAC,CAAM,CAAC,CAElD,CACA,OAAO,CACT,EACA,OAAOC,EAAAA,+BACL,EACA,EAAY,CAAW,EACvB,EAAY,EAAY,OAAQ,GAAU,CAAC,EAAM,KAAM,GAAe,EAAW,mBAAmB,CAAC,CAAC,EACtG,EACA,CACF,CAAC,CAAC,IAAK,GACL,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,gBACT,EAAW,eAAe,IAAK,GAAY,EAAM,EAAS,CAAC,EAAS,MAAM,CAAC,CAC7E,EACA,IAAK,IAAM,KAAQ,EACjB,EAAU,IAAI,CAAI,EAEpB,MAAO,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAC1D"}

package/dist/crossFileDuplication.js

@@ -1,2 +1,2 @@
-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 h(e,n))o.push(t);let c=s(o,_,(e,t)=>e.regionBucket-t.regionBucket||e.startIndex-t.startIndex),u=v(e,r),d=y([...c.values()],u,r);return b(l(d,p(e,d,u,n,i),n),e,u)}function l(e,t,n){let r=/* @__PURE__ */ new Map;for(let[t,n]of e.entries())for(let{fileIndex:e}of n){let n=r.get(e)??/* @__PURE__ */ new Set;n.add(t),r.set(e,n)}let i=/* @__PURE__ */ new Set,a=/* @__PURE__ */ new Set,o=t.map(t=>{let o=e=>{let r=t.filter(t=>t.segments.some(t=>e.segments.some(e=>u(e,t)))),[i]=r;if(r.length===1&&i)return Math.max(0,i.startTokenIndex-e.startTokenIndex)+Math.max(0,e.endTokenIndex-i.endTokenIndex)<n?i:void 0},s=/* @__PURE__ */ new Map,c=/* @__PURE__ */ new Set;for(let e of t)for(let t of e.spanCountedElsewhere?r.get(e.fileIndex)??[]:[])c.add(t);for(let t of[...c].toSorted((e,t)=>e-t)){let n=e[t]??[],r=n.map(o);if(!(i.has(t)||r.includes(void 0)||new Set(r).size!==r.length)){i.add(t);for(let[e,t]of r.entries()){let r=n[e];if(t&&r){let e=s.get(t)??[];e.push(r),s.set(t,e)}}}}return t.map(e=>{let t=s.get(e);if(!t)return e;let n=d([e,...t]);return a.add(n),n})});for(let t of a)for(let n of r.get(t.fileIndex)??[])if(!i.has(n)&&e[n]?.some(e=>!e.spanCountedElsewhere&&t.segments.every(t=>e.segments.some(e=>e.startTokenIndex<=t.startTokenIndex&&t.endTokenIndex<=e.endTokenIndex)))){t.spanCountedElsewhere=!0;break}return[...e.filter((e,t)=>!i.has(t)),...o]}function u(e,t){return e.startTokenIndex<t.endTokenIndex&&t.startTokenIndex<e.endTokenIndex}function d(e){let t=f(e.flatMap(e=>e.segments));return{...e[0],spanCountedElsewhere:void 0,nestedInLargerGroup:void 0,segments:t,matchedRuns:f(e.flatMap(e=>e.matchedRuns??e.segments)),exactSegments:f(e.flatMap(e=>e.matchedRuns?e.exactSegments??[]:e.segments)),tokenCount:t.reduce((e,t)=>e+t.endTokenIndex-t.startTokenIndex,0),startTokenIndex:Math.min(...e.map(e=>e.startTokenIndex)),endTokenIndex:Math.max(...e.map(e=>e.endTokenIndex)),startIndex:Math.min(...e.map(e=>e.startIndex)),endIndex:Math.max(...e.map(e=>e.endIndex)),startLine:Math.min(...e.map(e=>e.startLine)),endLine:Math.max(...e.map(e=>e.endLine))}}function f(e){let t=[];for(let n of e.toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex)){let e=t.at(-1);e&&n.startTokenIndex<e.endTokenIndex?e.endTokenIndex=Math.max(e.endTokenIndex,n.endTokenIndex):t.push({...n})}return t}function p(e,t,n,r,i){let a=e.map(()=>[]);for(let{fileIndex:e,segments:r}of t.flat()){let t=n[e]??0;for(let n of r)a[e]?.push(m(n,-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=>m(e,r)),matchedRuns:t.matchedRuns?.map(e=>m(e,r)),startTokenIndex:t.startTokenIndex+r,endTokenIndex:t.endTokenIndex+r}}))}function m({startTokenIndex:e,endTokenIndex:t},n){return{startTokenIndex:e+n,endTokenIndex:t+n}}function h(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 g(e){return new Set(e.map(e=>e.file)).size>=2}function _(e){return e.length>=2&&new Set(e.map(e=>e.regionBucket)).size>=2}function v(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 y(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,g)}function b(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=x(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 x(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=>m(e,-i.offset)),e.matchedRuns?.map(e=>m(e,-i.offset)),i.tokens,i.codeLineNumbers,e.exactSegments?.map(e=>m(e,-i.offset)));for(let e of o)a.add(e);return[...o].toSorted((e,t)=>e-t)}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 h(e,n))o.push(t);let c=s(o,_,(e,t)=>e.regionBucket-t.regionBucket||e.startIndex-t.startIndex),u=v(e,r),d=y([...c.values()],u,r);return b(l(d,p(e,d,u,n,i),n),e,u)}function l(e,t,n){let r=/* @__PURE__ */ new Map;for(let[t,n]of e.entries())for(let{fileIndex:e}of n){let n=r.get(e)??/* @__PURE__ */ new Set;n.add(t),r.set(e,n)}let i=/* @__PURE__ */ new Set,a=/* @__PURE__ */ new Set,o=t.map(t=>{let o=e=>{let r=t.filter(t=>t.segments.some(t=>e.segments.some(e=>u(e,t)))),[i]=r;if(r.length===1&&i)return Math.max(0,i.startTokenIndex-e.startTokenIndex)+Math.max(0,e.endTokenIndex-i.endTokenIndex)<n?i:void 0},s=/* @__PURE__ */ new Map,c=/* @__PURE__ */ new Set;for(let e of t)for(let t of e.spanCountedElsewhere?r.get(e.fileIndex)??[]:[])c.add(t);for(let t of[...c].toSorted((e,t)=>e-t)){let n=e[t]??[],r=n.map(o);if(!(i.has(t)||r.includes(void 0)||new Set(r).size!==r.length)){i.add(t);for(let[e,t]of r.entries()){let r=n[e];if(t&&r){let e=s.get(t)??[];e.push(r),s.set(t,e)}}}}return t.map(e=>{let t=s.get(e);if(!t)return e;let n=d([e,...t]);return a.add(n),n})});for(let t of a)for(let n of r.get(t.fileIndex)??[])if(!i.has(n)&&e[n]?.some(e=>!e.spanCountedElsewhere&&t.segments.every(t=>e.segments.some(e=>e.startTokenIndex<=t.startTokenIndex&&t.endTokenIndex<=e.endTokenIndex)))){t.spanCountedElsewhere=!0;break}return[...e.filter((e,t)=>!i.has(t)),...o]}function u(e,t){return e.startTokenIndex<t.endTokenIndex&&t.startTokenIndex<e.endTokenIndex}function d(e){let t=f(e.flatMap(e=>e.segments));return{...e[0],spanCountedElsewhere:void 0,nestedInLargerGroup:void 0,segments:t,matchedRuns:f(e.flatMap(e=>e.matchedRuns??e.segments)),exactSegments:f(e.flatMap(e=>e.matchedRuns?e.exactSegments??[]:e.segments)),tokenCount:t.reduce((e,t)=>e+t.endTokenIndex-t.startTokenIndex,0),startTokenIndex:Math.min(...e.map(e=>e.startTokenIndex)),endTokenIndex:Math.max(...e.map(e=>e.endTokenIndex)),startIndex:Math.min(...e.map(e=>e.startIndex)),endIndex:Math.max(...e.map(e=>e.endIndex)),startLine:Math.min(...e.map(e=>e.startLine)),endLine:Math.max(...e.map(e=>e.endLine))}}function f(e){let t=[];for(let n of e.toSorted((e,t)=>e.startTokenIndex-t.startTokenIndex)){let e=t.at(-1);e&&n.startTokenIndex<e.endTokenIndex?e.endTokenIndex=Math.max(e.endTokenIndex,n.endTokenIndex):t.push({...n})}return t}function p(e,t,n,r,i){let a=t=>{let r=e.map(()=>[]);for(let{fileIndex:e,segments:i}of t.flat()){let t=n[e]??0;for(let n of i)r[e]?.push(m(n,-t))}return r};return o(e,a(t),a(t.filter(e=>!e.some(e=>e.nestedInLargerGroup))),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=>m(e,r)),matchedRuns:t.matchedRuns?.map(e=>m(e,r)),startTokenIndex:t.startTokenIndex+r,endTokenIndex:t.endTokenIndex+r}}))}function m({startTokenIndex:e,endTokenIndex:t},n){return{startTokenIndex:e+n,endTokenIndex:t+n}}function h(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 g(e){return new Set(e.map(e=>e.file)).size>=2}function _(e){return e.length>=2&&new Set(e.map(e=>e.regionBucket)).size>=2}function v(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 y(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,g)}function b(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=x(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 x(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=>m(e,-i.offset)),e.matchedRuns?.map(e=>m(e,-i.offset)),i.tokens,i.codeLineNumbers,e.exactSegments?.map(e=>m(e,-i.offset)));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  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 or which a part of it\n   * matched exactly reaches (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 exactGroups = mergeGapAdjacentGroups([...counted.values()], tokenOffsets, maxGapTokens);\n  const nearMissGroups = collectNearMissGroups(files, exactGroups, tokenOffsets, minTokens, minSimilarityPercent);\n  return summarize(takeOverExactGroups(exactGroups, nearMissGroups, minTokens), files, tokenOffsets);\n}\n\n/**\n * Lets a near-miss group take over the exact groups standing for the same copies, as within a\n * file (collect_near_miss_groups in native/src/duplication.rs): every occurrence of such a group\n * overlaps exactly one copy of the near-miss group (one overlapping none reports content the copies\n * do not share, and one reaching into two would make one copy take over lines of the next), exceeds\n * its span by less than `minTokens` (a copy is listed at the size it matched, not at that of a\n * clone larger by a reportable part, which may run into code no copy matched), and no two overlap\n * the same copy (those repeat within the copy, not between the copies).\n * Each copy then becomes one occurrence with the exact fragments it holds, so the copies are\n * listed once.\n */\nfunction takeOverExactGroups(\n  exactGroups: CrossFileOccurrence[][],\n  nearMissGroups: CrossFileOccurrence[][],\n  minTokens: number\n): CrossFileOccurrence[][] {\n  const exactGroupIndexesByFile = new Map<number, Set<number>>();\n  for (const [groupIndex, group] of exactGroups.entries()) {\n    for (const { fileIndex } of group) {\n      const indexes = exactGroupIndexesByFile.get(fileIndex) ?? new Set();\n      indexes.add(groupIndex);\n      exactGroupIndexesByFile.set(fileIndex, indexes);\n    }\n  }\n  const taken = new Set<number>();\n  const coalesced = new Set<CrossFileOccurrence>();\n  const merged = nearMissGroups.map((copies) => {\n    const copyOf = (occurrence: CrossFileOccurrence): CrossFileOccurrence | undefined => {\n      const overlapped = copies.filter((copy) =>\n        copy.segments.some((segment) => occurrence.segments.some((matched) => overlap(matched, segment)))\n      );\n      const [copy] = overlapped;\n      if (overlapped.length !== 1 || !copy) {\n        return undefined;\n      }\n      const excess =\n        Math.max(0, copy.startTokenIndex - occurrence.startTokenIndex) +\n        Math.max(0, occurrence.endTokenIndex - copy.endTokenIndex);\n      return excess < minTokens ? copy : undefined;\n    };\n    const fragmentsByCopy = new Map<CrossFileOccurrence, CrossFileOccurrence[]>();\n    const candidateIndexes = new Set<number>();\n    for (const copy of copies) {\n      for (const groupIndex of copy.spanCountedElsewhere ? (exactGroupIndexesByFile.get(copy.fileIndex) ?? []) : []) {\n        candidateIndexes.add(groupIndex);\n      }\n    }\n    for (const groupIndex of [...candidateIndexes].toSorted((left, right) => left - right)) {\n      const group = exactGroups[groupIndex] ?? [];\n      const owners = group.map(copyOf);\n      if (taken.has(groupIndex) || owners.includes(undefined) || new Set(owners).size !== owners.length) {\n        continue;\n      }\n      taken.add(groupIndex);\n      for (const [index, owner] of owners.entries()) {\n        const fragment = group[index];\n        if (owner && fragment) {\n          const fragments = fragmentsByCopy.get(owner) ?? [];\n          fragments.push(fragment);\n          fragmentsByCopy.set(owner, fragments);\n        }\n      }\n    }\n    return copies.map((copy): CrossFileOccurrence => {\n      const fragments = fragmentsByCopy.get(copy);\n      if (!fragments) {\n        return copy;\n      }\n      const whole = coalesceOccurrences([copy, ...fragments]);\n      coalesced.add(whole);\n      return whole;\n    });\n  });\n  // A copy that took its fragments over counts them itself, unless a group left in place still\n  // counts all of it: a larger clone whose matched segments enclose those of the copy, which its\n  // fragments were nested in. A smaller run left in place inside the copy counts its own repeats,\n  // not the copy.\n  for (const copy of coalesced) {\n    for (const groupIndex of exactGroupIndexesByFile.get(copy.fileIndex) ?? []) {\n      const enclosed =\n        !taken.has(groupIndex) &&\n        exactGroups[groupIndex]?.some(\n          (occurrence) =>\n            !occurrence.spanCountedElsewhere &&\n            copy.segments.every((segment) =>\n              occurrence.segments.some(\n                (matched) =>\n                  matched.startTokenIndex <= segment.startTokenIndex && segment.endTokenIndex <= matched.endTokenIndex\n              )\n            )\n        );\n      if (enclosed) {\n        copy.spanCountedElsewhere = true;\n        break;\n      }\n    }\n  }\n  return [...exactGroups.filter((_, groupIndex) => !taken.has(groupIndex)), ...merged];\n}\n\nfunction overlap(left: TokenSegment, right: TokenSegment): boolean {\n  return left.startTokenIndex < right.endTokenIndex && right.startTokenIndex < left.endTokenIndex;\n}\n\n/** The parts of one copy as one occurrence: an exact part is matched throughout, a near-miss part where its partners match it. */\nfunction coalesceOccurrences(parts: [CrossFileOccurrence, ...CrossFileOccurrence[]]): CrossFileOccurrence {\n  const segments = mergeSegments(parts.flatMap((part) => part.segments));\n  return {\n    ...parts[0],\n    spanCountedElsewhere: undefined,\n    nestedInLargerGroup: undefined,\n    segments,\n    matchedRuns: mergeSegments(parts.flatMap((part) => part.matchedRuns ?? part.segments)),\n    exactSegments: mergeSegments(\n      parts.flatMap((part) => (part.matchedRuns ? (part.exactSegments ?? []) : part.segments))\n    ),\n    tokenCount: segments.reduce((sum, segment) => sum + segment.endTokenIndex - segment.startTokenIndex, 0),\n    startTokenIndex: Math.min(...parts.map((part) => part.startTokenIndex)),\n    endTokenIndex: Math.max(...parts.map((part) => part.endTokenIndex)),\n    startIndex: Math.min(...parts.map((part) => part.startIndex)),\n    endIndex: Math.max(...parts.map((part) => part.endIndex)),\n    startLine: Math.min(...parts.map((part) => part.startLine)),\n    endLine: Math.max(...parts.map((part) => part.endLine)),\n  };\n}\n\n/** The unions of overlapping segments, in position order. */\nfunction mergeSegments(segments: TokenSegment[]): TokenSegment[] {\n  const merged: TokenSegment[] = [];\n  for (const segment of segments.toSorted((left, right) => left.startTokenIndex - right.startTokenIndex)) {\n    const last = merged.at(-1);\n    if (last && segment.startTokenIndex < last.endTokenIndex) {\n      last.endTokenIndex = Math.max(last.endTokenIndex, segment.endTokenIndex);\n    } else {\n      merged.push({ ...segment });\n    }\n  }\n  return merged;\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  // Segment by segment: the gap of a gapped clone is not reported content.\n  for (const { fileIndex, segments } of exactGroups.flat()) {\n    const offset = tokenOffsets[fileIndex] ?? 0;\n    for (const segment of segments) {\n      reportedSpansByFile[fileIndex]?.push(shift(segment, -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    occurrence.exactSegments?.map((segment) => shift(segment, -fileData.offset))\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,EAAc,EAAuB,CAAC,GAAG,EAAQ,OAAO,CAAC,EAAG,EAAc,CAAY,EAE5F,OAAO,EAAU,EAAoB,EADd,EAAsB,EAAO,EAAa,EAAc,EAAW,CAC3B,EAAG,CAAS,EAAG,EAAO,CAAY,CACnG,CAaA,SAAS,EACP,EACA,EACA,EACyB,CACzB,IAAM,kBAA0B,IAAI,IACpC,IAAK,GAAM,CAAC,EAAY,KAAU,EAAY,QAAQ,EACpD,IAAK,GAAM,CAAE,eAAe,EAAO,CACjC,IAAM,EAAU,EAAwB,IAAI,CAAS,mBAAK,IAAI,IAC9D,EAAQ,IAAI,CAAU,EACtB,EAAwB,IAAI,EAAW,CAAO,CAChD,CAEF,IAAM,kBAAQ,IAAI,IACZ,kBAAY,IAAI,IAChB,EAAS,EAAe,IAAK,GAAW,CAC5C,IAAM,EAAU,GAAqE,CACnF,IAAM,EAAa,EAAO,OAAQ,GAChC,EAAK,SAAS,KAAM,GAAY,EAAW,SAAS,KAAM,GAAY,EAAQ,EAAS,CAAO,CAAC,CAAC,CAClG,EACM,CAAC,GAAQ,EACX,KAAW,SAAW,GAAM,EAMhC,OAFE,KAAK,IAAI,EAAG,EAAK,gBAAkB,EAAW,eAAe,EAC7D,KAAK,IAAI,EAAG,EAAW,cAAgB,EAAK,aAAa,EAC3C,EAAY,EAAO,IAAA,EACrC,EACM,kBAAkB,IAAI,IACtB,kBAAmB,IAAI,IAC7B,IAAK,IAAM,KAAQ,EACjB,IAAK,IAAM,KAAc,EAAK,qBAAwB,EAAwB,IAAI,EAAK,SAAS,GAAK,CAAC,EAAK,CAAC,EAC1G,EAAiB,IAAI,CAAU,EAGnC,IAAK,IAAM,IAAc,CAAC,GAAG,CAAgB,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,EAAG,CACtF,IAAM,EAAQ,EAAY,IAAe,CAAC,EACpC,EAAS,EAAM,IAAI,CAAM,EAC3B,OAAM,IAAI,CAAU,GAAK,EAAO,SAAS,IAAA,EAAS,GAAK,IAAI,IAAI,CAAM,CAAC,CAAC,OAAS,EAAO,QAG3F,GAAM,IAAI,CAAU,EACpB,IAAK,GAAM,CAAC,EAAO,KAAU,EAAO,QAAQ,EAAG,CAC7C,IAAM,EAAW,EAAM,GACvB,GAAI,GAAS,EAAU,CACrB,IAAM,EAAY,EAAgB,IAAI,CAAK,GAAK,CAAC,EACjD,EAAU,KAAK,CAAQ,EACvB,EAAgB,IAAI,EAAO,CAAS,CACtC,CACF,CARoB,CAStB,CACA,OAAO,EAAO,IAAK,GAA8B,CAC/C,IAAM,EAAY,EAAgB,IAAI,CAAI,EAC1C,GAAI,CAAC,EACH,OAAO,EAET,IAAM,EAAQ,EAAoB,CAAC,EAAM,GAAG,CAAS,CAAC,EAEtD,OADA,EAAU,IAAI,CAAK,EACZ,CACT,CAAC,CACH,CAAC,EAKD,IAAK,IAAM,KAAQ,EACjB,IAAK,IAAM,KAAc,EAAwB,IAAI,EAAK,SAAS,GAAK,CAAC,EAavE,GAXE,CAAC,EAAM,IAAI,CAAU,GACrB,EAAY,EAAW,EAAE,KACtB,GACC,CAAC,EAAW,sBACZ,EAAK,SAAS,MAAO,GACnB,EAAW,SAAS,KACjB,GACC,EAAQ,iBAAmB,EAAQ,iBAAmB,EAAQ,eAAiB,EAAQ,aAC3F,CACF,CACJ,EACY,CACZ,EAAK,qBAAuB,GAC5B,KACF,CAGJ,MAAO,CAAC,GAAG,EAAY,QAAQ,EAAG,IAAe,CAAC,EAAM,IAAI,CAAU,CAAC,EAAG,GAAG,CAAM,CACrF,CAEA,SAAS,EAAQ,EAAoB,EAA8B,CACjE,OAAO,EAAK,gBAAkB,EAAM,eAAiB,EAAM,gBAAkB,EAAK,aACpF,CAGA,SAAS,EAAoB,EAA6E,CACxG,IAAM,EAAW,EAAc,EAAM,QAAS,GAAS,EAAK,QAAQ,CAAC,EACrE,MAAO,CACL,GAAG,EAAM,GACT,qBAAsB,IAAA,GACtB,oBAAqB,IAAA,GACrB,WACA,YAAa,EAAc,EAAM,QAAS,GAAS,EAAK,aAAe,EAAK,QAAQ,CAAC,EACrF,cAAe,EACb,EAAM,QAAS,GAAU,EAAK,YAAe,EAAK,eAAiB,CAAC,EAAK,EAAK,QAAS,CACzF,EACA,WAAY,EAAS,QAAQ,EAAK,IAAY,EAAM,EAAQ,cAAgB,EAAQ,gBAAiB,CAAC,EACtG,gBAAiB,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,eAAe,CAAC,EACtE,cAAe,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,aAAa,CAAC,EAClE,WAAY,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,UAAU,CAAC,EAC5D,SAAU,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,QAAQ,CAAC,EACxD,UAAW,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,SAAS,CAAC,EAC1D,QAAS,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,OAAO,CAAC,CACxD,CACF,CAGA,SAAS,EAAc,EAA0C,CAC/D,IAAM,EAAyB,CAAC,EAChC,IAAK,IAAM,KAAW,EAAS,UAAU,EAAM,IAAU,EAAK,gBAAkB,EAAM,eAAe,EAAG,CACtG,IAAM,EAAO,EAAO,GAAG,EAAE,EACrB,GAAQ,EAAQ,gBAAkB,EAAK,cACzC,EAAK,cAAgB,KAAK,IAAI,EAAK,cAAe,EAAQ,aAAa,EAEvE,EAAO,KAAK,CAAE,GAAG,CAAQ,CAAC,CAE9B,CACA,OAAO,CACT,CAGA,SAAS,EACP,EACA,EACA,EACA,EACA,EACyB,CACzB,IAAM,EAA8E,EAAM,QAAU,CAAC,CAAC,EAEtG,IAAK,GAAM,CAAE,YAAW,cAAc,EAAY,KAAK,EAAG,CACxD,IAAM,EAAS,EAAa,IAAc,EAC1C,IAAK,IAAM,KAAW,EACpB,EAAoB,EAAU,EAAE,KAAK,EAAM,EAAS,CAAC,CAAM,CAAC,CAEhE,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,gBACT,EAAW,eAAe,IAAK,GAAY,EAAM,EAAS,CAAC,EAAS,MAAM,CAAC,CAC7E,EACA,IAAK,IAAM,KAAQ,EACjB,EAAU,IAAI,CAAI,EAEpB,MAAO,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAC1D"}
+{"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 or which a part of it\n   * matched exactly reaches (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 exactGroups = mergeGapAdjacentGroups([...counted.values()], tokenOffsets, maxGapTokens);\n  const nearMissGroups = collectNearMissGroups(files, exactGroups, tokenOffsets, minTokens, minSimilarityPercent);\n  return summarize(takeOverExactGroups(exactGroups, nearMissGroups, minTokens), files, tokenOffsets);\n}\n\n/**\n * Lets a near-miss group take over the exact groups standing for the same copies, as within a\n * file (collect_near_miss_groups in native/src/duplication.rs): every occurrence of such a group\n * overlaps exactly one copy of the near-miss group (one overlapping none reports content the copies\n * do not share, and one reaching into two would make one copy take over lines of the next), exceeds\n * its span by less than `minTokens` (a copy is listed at the size it matched, not at that of a\n * clone larger by a reportable part, which may run into code no copy matched), and no two overlap\n * the same copy (those repeat within the copy, not between the copies).\n * Each copy then becomes one occurrence with the exact fragments it holds, so the copies are\n * listed once.\n */\nfunction takeOverExactGroups(\n  exactGroups: CrossFileOccurrence[][],\n  nearMissGroups: CrossFileOccurrence[][],\n  minTokens: number\n): CrossFileOccurrence[][] {\n  const exactGroupIndexesByFile = new Map<number, Set<number>>();\n  for (const [groupIndex, group] of exactGroups.entries()) {\n    for (const { fileIndex } of group) {\n      const indexes = exactGroupIndexesByFile.get(fileIndex) ?? new Set();\n      indexes.add(groupIndex);\n      exactGroupIndexesByFile.set(fileIndex, indexes);\n    }\n  }\n  const taken = new Set<number>();\n  const coalesced = new Set<CrossFileOccurrence>();\n  const merged = nearMissGroups.map((copies) => {\n    const copyOf = (occurrence: CrossFileOccurrence): CrossFileOccurrence | undefined => {\n      const overlapped = copies.filter((copy) =>\n        copy.segments.some((segment) => occurrence.segments.some((matched) => overlap(matched, segment)))\n      );\n      const [copy] = overlapped;\n      if (overlapped.length !== 1 || !copy) {\n        return undefined;\n      }\n      const excess =\n        Math.max(0, copy.startTokenIndex - occurrence.startTokenIndex) +\n        Math.max(0, occurrence.endTokenIndex - copy.endTokenIndex);\n      return excess < minTokens ? copy : undefined;\n    };\n    const fragmentsByCopy = new Map<CrossFileOccurrence, CrossFileOccurrence[]>();\n    const candidateIndexes = new Set<number>();\n    for (const copy of copies) {\n      for (const groupIndex of copy.spanCountedElsewhere ? (exactGroupIndexesByFile.get(copy.fileIndex) ?? []) : []) {\n        candidateIndexes.add(groupIndex);\n      }\n    }\n    for (const groupIndex of [...candidateIndexes].toSorted((left, right) => left - right)) {\n      const group = exactGroups[groupIndex] ?? [];\n      const owners = group.map(copyOf);\n      if (taken.has(groupIndex) || owners.includes(undefined) || new Set(owners).size !== owners.length) {\n        continue;\n      }\n      taken.add(groupIndex);\n      for (const [index, owner] of owners.entries()) {\n        const fragment = group[index];\n        if (owner && fragment) {\n          const fragments = fragmentsByCopy.get(owner) ?? [];\n          fragments.push(fragment);\n          fragmentsByCopy.set(owner, fragments);\n        }\n      }\n    }\n    return copies.map((copy): CrossFileOccurrence => {\n      const fragments = fragmentsByCopy.get(copy);\n      if (!fragments) {\n        return copy;\n      }\n      const whole = coalesceOccurrences([copy, ...fragments]);\n      coalesced.add(whole);\n      return whole;\n    });\n  });\n  // A copy that took its fragments over counts them itself, unless a group left in place still\n  // counts all of it: a larger clone whose matched segments enclose those of the copy, which its\n  // fragments were nested in. A smaller run left in place inside the copy counts its own repeats,\n  // not the copy.\n  for (const copy of coalesced) {\n    for (const groupIndex of exactGroupIndexesByFile.get(copy.fileIndex) ?? []) {\n      const enclosed =\n        !taken.has(groupIndex) &&\n        exactGroups[groupIndex]?.some(\n          (occurrence) =>\n            !occurrence.spanCountedElsewhere &&\n            copy.segments.every((segment) =>\n              occurrence.segments.some(\n                (matched) =>\n                  matched.startTokenIndex <= segment.startTokenIndex && segment.endTokenIndex <= matched.endTokenIndex\n              )\n            )\n        );\n      if (enclosed) {\n        copy.spanCountedElsewhere = true;\n        break;\n      }\n    }\n  }\n  return [...exactGroups.filter((_, groupIndex) => !taken.has(groupIndex)), ...merged];\n}\n\nfunction overlap(left: TokenSegment, right: TokenSegment): boolean {\n  return left.startTokenIndex < right.endTokenIndex && right.startTokenIndex < left.endTokenIndex;\n}\n\n/** The parts of one copy as one occurrence: an exact part is matched throughout, a near-miss part where its partners match it. */\nfunction coalesceOccurrences(parts: [CrossFileOccurrence, ...CrossFileOccurrence[]]): CrossFileOccurrence {\n  const segments = mergeSegments(parts.flatMap((part) => part.segments));\n  return {\n    ...parts[0],\n    spanCountedElsewhere: undefined,\n    nestedInLargerGroup: undefined,\n    segments,\n    matchedRuns: mergeSegments(parts.flatMap((part) => part.matchedRuns ?? part.segments)),\n    exactSegments: mergeSegments(\n      parts.flatMap((part) => (part.matchedRuns ? (part.exactSegments ?? []) : part.segments))\n    ),\n    tokenCount: segments.reduce((sum, segment) => sum + segment.endTokenIndex - segment.startTokenIndex, 0),\n    startTokenIndex: Math.min(...parts.map((part) => part.startTokenIndex)),\n    endTokenIndex: Math.max(...parts.map((part) => part.endTokenIndex)),\n    startIndex: Math.min(...parts.map((part) => part.startIndex)),\n    endIndex: Math.max(...parts.map((part) => part.endIndex)),\n    startLine: Math.min(...parts.map((part) => part.startLine)),\n    endLine: Math.max(...parts.map((part) => part.endLine)),\n  };\n}\n\n/** The unions of overlapping segments, in position order. */\nfunction mergeSegments(segments: TokenSegment[]): TokenSegment[] {\n  const merged: TokenSegment[] = [];\n  for (const segment of segments.toSorted((left, right) => left.startTokenIndex - right.startTokenIndex)) {\n    const last = merged.at(-1);\n    if (last && segment.startTokenIndex < last.endTokenIndex) {\n      last.endTokenIndex = Math.max(last.endTokenIndex, segment.endTokenIndex);\n    } else {\n      merged.push({ ...segment });\n    }\n  }\n  return merged;\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  // Segment by segment: the gap of a gapped clone is not reported content.\n  const spansByFile = (groups: CrossFileOccurrence[][]): TokenSegment[][] => {\n    const spans: TokenSegment[][] = files.map(() => []);\n    for (const { fileIndex, segments } of groups.flat()) {\n      const offset = tokenOffsets[fileIndex] ?? 0;\n      for (const segment of segments) {\n        spans[fileIndex]?.push(shift(segment, -offset));\n      }\n    }\n    return spans;\n  };\n  return collectCrossFileNearMissGroups(\n    files,\n    spansByFile(exactGroups),\n    spansByFile(exactGroups.filter((group) => !group.some((occurrence) => occurrence.nestedInLargerGroup))),\n    minTokens,\n    minSimilarityPercent\n  ).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    occurrence.exactSegments?.map((segment) => shift(segment, -fileData.offset))\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,EAAc,EAAuB,CAAC,GAAG,EAAQ,OAAO,CAAC,EAAG,EAAc,CAAY,EAE5F,OAAO,EAAU,EAAoB,EADd,EAAsB,EAAO,EAAa,EAAc,EAAW,CAC3B,EAAG,CAAS,EAAG,EAAO,CAAY,CACnG,CAaA,SAAS,EACP,EACA,EACA,EACyB,CACzB,IAAM,kBAA0B,IAAI,IACpC,IAAK,GAAM,CAAC,EAAY,KAAU,EAAY,QAAQ,EACpD,IAAK,GAAM,CAAE,eAAe,EAAO,CACjC,IAAM,EAAU,EAAwB,IAAI,CAAS,mBAAK,IAAI,IAC9D,EAAQ,IAAI,CAAU,EACtB,EAAwB,IAAI,EAAW,CAAO,CAChD,CAEF,IAAM,kBAAQ,IAAI,IACZ,kBAAY,IAAI,IAChB,EAAS,EAAe,IAAK,GAAW,CAC5C,IAAM,EAAU,GAAqE,CACnF,IAAM,EAAa,EAAO,OAAQ,GAChC,EAAK,SAAS,KAAM,GAAY,EAAW,SAAS,KAAM,GAAY,EAAQ,EAAS,CAAO,CAAC,CAAC,CAClG,EACM,CAAC,GAAQ,EACX,KAAW,SAAW,GAAM,EAMhC,OAFE,KAAK,IAAI,EAAG,EAAK,gBAAkB,EAAW,eAAe,EAC7D,KAAK,IAAI,EAAG,EAAW,cAAgB,EAAK,aAAa,EAC3C,EAAY,EAAO,IAAA,EACrC,EACM,kBAAkB,IAAI,IACtB,kBAAmB,IAAI,IAC7B,IAAK,IAAM,KAAQ,EACjB,IAAK,IAAM,KAAc,EAAK,qBAAwB,EAAwB,IAAI,EAAK,SAAS,GAAK,CAAC,EAAK,CAAC,EAC1G,EAAiB,IAAI,CAAU,EAGnC,IAAK,IAAM,IAAc,CAAC,GAAG,CAAgB,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,EAAG,CACtF,IAAM,EAAQ,EAAY,IAAe,CAAC,EACpC,EAAS,EAAM,IAAI,CAAM,EAC3B,OAAM,IAAI,CAAU,GAAK,EAAO,SAAS,IAAA,EAAS,GAAK,IAAI,IAAI,CAAM,CAAC,CAAC,OAAS,EAAO,QAG3F,GAAM,IAAI,CAAU,EACpB,IAAK,GAAM,CAAC,EAAO,KAAU,EAAO,QAAQ,EAAG,CAC7C,IAAM,EAAW,EAAM,GACvB,GAAI,GAAS,EAAU,CACrB,IAAM,EAAY,EAAgB,IAAI,CAAK,GAAK,CAAC,EACjD,EAAU,KAAK,CAAQ,EACvB,EAAgB,IAAI,EAAO,CAAS,CACtC,CACF,CARoB,CAStB,CACA,OAAO,EAAO,IAAK,GAA8B,CAC/C,IAAM,EAAY,EAAgB,IAAI,CAAI,EAC1C,GAAI,CAAC,EACH,OAAO,EAET,IAAM,EAAQ,EAAoB,CAAC,EAAM,GAAG,CAAS,CAAC,EAEtD,OADA,EAAU,IAAI,CAAK,EACZ,CACT,CAAC,CACH,CAAC,EAKD,IAAK,IAAM,KAAQ,EACjB,IAAK,IAAM,KAAc,EAAwB,IAAI,EAAK,SAAS,GAAK,CAAC,EAavE,GAXE,CAAC,EAAM,IAAI,CAAU,GACrB,EAAY,EAAW,EAAE,KACtB,GACC,CAAC,EAAW,sBACZ,EAAK,SAAS,MAAO,GACnB,EAAW,SAAS,KACjB,GACC,EAAQ,iBAAmB,EAAQ,iBAAmB,EAAQ,eAAiB,EAAQ,aAC3F,CACF,CACJ,EACY,CACZ,EAAK,qBAAuB,GAC5B,KACF,CAGJ,MAAO,CAAC,GAAG,EAAY,QAAQ,EAAG,IAAe,CAAC,EAAM,IAAI,CAAU,CAAC,EAAG,GAAG,CAAM,CACrF,CAEA,SAAS,EAAQ,EAAoB,EAA8B,CACjE,OAAO,EAAK,gBAAkB,EAAM,eAAiB,EAAM,gBAAkB,EAAK,aACpF,CAGA,SAAS,EAAoB,EAA6E,CACxG,IAAM,EAAW,EAAc,EAAM,QAAS,GAAS,EAAK,QAAQ,CAAC,EACrE,MAAO,CACL,GAAG,EAAM,GACT,qBAAsB,IAAA,GACtB,oBAAqB,IAAA,GACrB,WACA,YAAa,EAAc,EAAM,QAAS,GAAS,EAAK,aAAe,EAAK,QAAQ,CAAC,EACrF,cAAe,EACb,EAAM,QAAS,GAAU,EAAK,YAAe,EAAK,eAAiB,CAAC,EAAK,EAAK,QAAS,CACzF,EACA,WAAY,EAAS,QAAQ,EAAK,IAAY,EAAM,EAAQ,cAAgB,EAAQ,gBAAiB,CAAC,EACtG,gBAAiB,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,eAAe,CAAC,EACtE,cAAe,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,aAAa,CAAC,EAClE,WAAY,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,UAAU,CAAC,EAC5D,SAAU,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,QAAQ,CAAC,EACxD,UAAW,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,SAAS,CAAC,EAC1D,QAAS,KAAK,IAAI,GAAG,EAAM,IAAK,GAAS,EAAK,OAAO,CAAC,CACxD,CACF,CAGA,SAAS,EAAc,EAA0C,CAC/D,IAAM,EAAyB,CAAC,EAChC,IAAK,IAAM,KAAW,EAAS,UAAU,EAAM,IAAU,EAAK,gBAAkB,EAAM,eAAe,EAAG,CACtG,IAAM,EAAO,EAAO,GAAG,EAAE,EACrB,GAAQ,EAAQ,gBAAkB,EAAK,cACzC,EAAK,cAAgB,KAAK,IAAI,EAAK,cAAe,EAAQ,aAAa,EAEvE,EAAO,KAAK,CAAE,GAAG,CAAQ,CAAC,CAE9B,CACA,OAAO,CACT,CAGA,SAAS,EACP,EACA,EACA,EACA,EACA,EACyB,CAEzB,IAAM,EAAe,GAAsD,CACzE,IAAM,EAA0B,EAAM,QAAU,CAAC,CAAC,EAClD,IAAK,GAAM,CAAE,YAAW,cAAc,EAAO,KAAK,EAAG,CACnD,IAAM,EAAS,EAAa,IAAc,EAC1C,IAAK,IAAM,KAAW,EACpB,EAAM,EAAU,EAAE,KAAK,EAAM,EAAS,CAAC,CAAM,CAAC,CAElD,CACA,OAAO,CACT,EACA,OAAO,EACL,EACA,EAAY,CAAW,EACvB,EAAY,EAAY,OAAQ,GAAU,CAAC,EAAM,KAAM,GAAe,EAAW,mBAAmB,CAAC,CAAC,EACtG,EACA,CACF,CAAC,CAAC,IAAK,GACL,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,gBACT,EAAW,eAAe,IAAK,GAAY,EAAM,EAAS,CAAC,EAAS,MAAM,CAAC,CAC7E,EACA,IAAK,IAAM,KAAQ,EACjB,EAAU,IAAI,CAAI,EAEpB,MAAO,CAAC,GAAG,CAAK,CAAC,CAAC,UAAU,EAAM,IAAU,EAAO,CAAK,CAC1D"}