@opensip-cli/graph 0.1.10 → 0.1.12

package/dist/index.js

@@ -1,9 +1,9 @@
-// @fitness-ignore-file module-coupling-fan-out -- Public barrel: re-exports the surface of each pipeline stage; fan-out is the contract of this file
 /**
  * @opensip-cli/graph — public barrel.
  *
- * The graph tool implements a strict six-stage pipeline:
- * discover → inventory → edges → indexes → rules → render.
+ * The graph tool implements a strict seven-stage pipeline (the canonical
+ * `GRAPH_STAGES` execution order):
+ * discover → parse → walk → resolve → index → features → rules.
  */
 // Side-effect import: surfaces the `scope.graph` augmentation on
 // @opensip-cli/core's RunScope interface (D7 — tool subscopes via
@@ -20,6 +20,12 @@ export { GRAPH_CONTRACT_VERSION } from './tool.js';
 export { GRAPH_ENV_SPECS } from './cli/pressure-monitor.js';
 export { pickAdapter, createAdapterRegistry, currentAdapterRegistry, GraphAdapterRegistry, } from './lang-adapter/registry.js';
 export { GraphAdapterSelector } from './lang-adapter/selector.js';
+// Size-guarded source-file reader for adapter parse steps (per-file 10 MB cap,
+// shared with the fitness engine's FileAccessor policy). Adapters use this
+// instead of a raw `readFileSync` so one pathological source file can't spike
+// graph memory; oversize throws and the adapter's per-file catch records a
+// ParseError + continues.
+export { readSourceFileGuarded, MAX_SOURCE_FILE_BYTES } from './lang-adapter/read-source.js';
 export { truncateForCallEdge, CALL_EDGE_TEXT_MAX, CREATION_EDGE_PREFIX, CREATION_EDGE_TEXT_MAX, appendEdge, createMutableStats, pushCreationEdge, } from './lang-adapter/edge-helpers.js';
 // ── Shared body-digest primitives ─────────────────────────────────
 //
@@ -29,6 +35,11 @@ export { truncateForCallEdge, CALL_EDGE_TEXT_MAX, CREATION_EDGE_PREFIX, CREATION
 // only its language-specific comment stripper / normalizer (round-3
 // audit 2026-05-30, finding D). Mirrors the edge-helpers rationale.
 export { normalizeWhitespace, hashBody } from './lang-adapter/body-digest.js';
+export { NEAR_DUP_SIGNATURE_K, NEAR_DUP_LSH_BANDS, NEAR_DUP_LSH_ROWS, shingle, bodySignature, estimateJaccard, lshBandHashes, digestCanonicalBody, } from './lang-adapter/near-duplicate-signature.js';
+export { languageOfFile } from './lang-adapter/language-of-file.js';
+// Canonical TS/JS test-file predicate, single-sourced in the shared substrate (ADR-0064,
+// D1). Re-exported so graph-typescript reaches it via the engine barrel (no new dep).
+export { isTestFilePath } from '@opensip-cli/clone-detection';
 // PR 3 of plan 2026-05-23-plan-graph-adapter-package-split.md:
 // first-party adapters live in their own graph-* packages, so the
 // engine no longer re-exports any adapter, parsed-project type, or

package/dist/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,qJAAqJ;AACrJ;;;;;GAKG;AAEH,iEAAiE;AACjE,kEAAkE;AAClE,wBAAwB;AACxB,OAAO,yBAAyB,CAAC;AAGjC,mEAAmE;AACnE,qEAAqE;AACrE,6DAA6D;AAC7D,qBAAqB;AACrB,OAAO,EAAE,SAAS,EAAE,SAAS,IAAI,IAAI,EAAE,MAAM,WAAW,CAAC;AACzD,OAAO,EAAE,sBAAsB,EAAE,MAAM,WAAW,CAAC;AACnD,oEAAoE;AACpE,6EAA6E;AAC7E,OAAO,EAAE,eAAe,EAAE,MAAM,2BAA2B,CAAC;AAsF5D,OAAO,EACL,WAAW,EACX,qBAAqB,EACrB,sBAAsB,EACtB,oBAAoB,GACrB,MAAM,4BAA4B,CAAC;AACpC,OAAO,EAAE,oBAAoB,EAAE,MAAM,4BAA4B,CAAC;AAClE,OAAO,EACL,mBAAmB,EACnB,kBAAkB,EAClB,oBAAoB,EACpB,sBAAsB,EACtB,UAAU,EACV,kBAAkB,EAClB,gBAAgB,GACjB,MAAM,gCAAgC,CAAC;AAWxC,qEAAqE;AACrE,EAAE;AACF,oEAAoE;AACpE,qEAAqE;AACrE,wEAAwE;AACxE,oEAAoE;AACpE,oEAAoE;AACpE,OAAO,EAAE,mBAAmB,EAAE,QAAQ,EAAE,MAAM,+BAA+B,CAAC;AAG9E,+DAA+D;AAC/D,kEAAkE;AAClE,mEAAmE;AACnE,qEAAqE;AACrE,gBAAgB;AAEhB,4EAA4E;AAC5E,gFAAgF;AAChF,OAAO,EAAE,YAAY,EAAE,MAAM,gBAAgB,CAAC;AAE9C,sEAAsE;AACtE,EAAE;AACF,uEAAuE;AACvE,yEAAyE;AACzE,sEAAsE;AACtE,uEAAuE;AACvE,4EAA4E;AAC5E,0EAA0E;AAC1E,wEAAwE;AACxE,0EAA0E;AAC1E,OAAO,EACL,gBAAgB,EAChB,yBAAyB,EACzB,kCAAkC,EAClC,yBAAyB,GAC1B,MAAM,iCAAiC,CAAC;AAOzC,OAAO,EAAE,uBAAuB,EAAE,YAAY,EAAE,MAAM,4BAA4B,CAAC;AAEnF,6EAA6E;AAC7E,qEAAqE;AACrE,6EAA6E;AAC7E,OAAO,EAAE,UAAU,EAAE,MAAM,wBAAwB,CAAC;AAEpD,sEAAsE;AACtE,OAAO,EAAE,iBAAiB,EAAE,MAAM,oBAAoB,CAAC;AAOvD,OAAO,EACL,mBAAmB,EACnB,yBAAyB,EACzB,kBAAkB,EAClB,mBAAmB,EACnB,oBAAoB,GACrB,MAAM,+BAA+B,CAAC;AACvC,OAAO,EACL,mBAAmB,EACnB,oBAAoB,EACpB,mBAAmB,GACpB,MAAM,uBAAuB,CAAC;AAC/B,OAAO,EAAE,oBAAoB,EAAE,MAAM,sBAAsB,CAAC"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,iEAAiE;AACjE,kEAAkE;AAClE,wBAAwB;AACxB,OAAO,yBAAyB,CAAC;AAGjC,mEAAmE;AACnE,qEAAqE;AACrE,6DAA6D;AAC7D,qBAAqB;AACrB,OAAO,EAAE,SAAS,EAAE,SAAS,IAAI,IAAI,EAAE,MAAM,WAAW,CAAC;AACzD,OAAO,EAAE,sBAAsB,EAAE,MAAM,WAAW,CAAC;AACnD,oEAAoE;AACpE,6EAA6E;AAC7E,OAAO,EAAE,eAAe,EAAE,MAAM,2BAA2B,CAAC;AAsF5D,OAAO,EACL,WAAW,EACX,qBAAqB,EACrB,sBAAsB,EACtB,oBAAoB,GACrB,MAAM,4BAA4B,CAAC;AACpC,OAAO,EAAE,oBAAoB,EAAE,MAAM,4BAA4B,CAAC;AAClE,+EAA+E;AAC/E,2EAA2E;AAC3E,8EAA8E;AAC9E,2EAA2E;AAC3E,0BAA0B;AAC1B,OAAO,EAAE,qBAAqB,EAAE,qBAAqB,EAAE,MAAM,+BAA+B,CAAC;AAC7F,OAAO,EACL,mBAAmB,EACnB,kBAAkB,EAClB,oBAAoB,EACpB,sBAAsB,EACtB,UAAU,EACV,kBAAkB,EAClB,gBAAgB,GACjB,MAAM,gCAAgC,CAAC;AAWxC,qEAAqE;AACrE,EAAE;AACF,oEAAoE;AACpE,qEAAqE;AACrE,wEAAwE;AACxE,oEAAoE;AACpE,oEAAoE;AACpE,OAAO,EAAE,mBAAmB,EAAE,QAAQ,EAAE,MAAM,+BAA+B,CAAC;AAE9E,OAAO,EACL,oBAAoB,EACpB,kBAAkB,EAClB,iBAAiB,EACjB,OAAO,EACP,aAAa,EACb,eAAe,EACf,aAAa,EACb,mBAAmB,GACpB,MAAM,4CAA4C,CAAC;AAEpD,OAAO,EAAE,cAAc,EAAE,MAAM,oCAAoC,CAAC;AACpE,yFAAyF;AACzF,sFAAsF;AACtF,OAAO,EAAE,cAAc,EAAE,MAAM,8BAA8B,CAAC;AAE9D,+DAA+D;AAC/D,kEAAkE;AAClE,mEAAmE;AACnE,qEAAqE;AACrE,gBAAgB;AAEhB,4EAA4E;AAC5E,gFAAgF;AAChF,OAAO,EAAE,YAAY,EAAE,MAAM,gBAAgB,CAAC;AAE9C,sEAAsE;AACtE,EAAE;AACF,uEAAuE;AACvE,yEAAyE;AACzE,sEAAsE;AACtE,uEAAuE;AACvE,4EAA4E;AAC5E,0EAA0E;AAC1E,wEAAwE;AACxE,0EAA0E;AAC1E,OAAO,EACL,gBAAgB,EAChB,yBAAyB,EACzB,kCAAkC,EAClC,yBAAyB,GAC1B,MAAM,iCAAiC,CAAC;AAOzC,OAAO,EAAE,uBAAuB,EAAE,YAAY,EAAE,MAAM,4BAA4B,CAAC;AAEnF,6EAA6E;AAC7E,qEAAqE;AACrE,6EAA6E;AAC7E,OAAO,EAAE,UAAU,EAAE,MAAM,wBAAwB,CAAC;AAEpD,sEAAsE;AACtE,OAAO,EAAE,iBAAiB,EAAE,MAAM,oBAAoB,CAAC;AAOvD,OAAO,EACL,mBAAmB,EACnB,yBAAyB,EACzB,kBAAkB,EAClB,mBAAmB,EACnB,oBAAoB,GACrB,MAAM,+BAA+B,CAAC;AACvC,OAAO,EACL,mBAAmB,EACnB,oBAAoB,EACpB,mBAAmB,GACpB,MAAM,uBAAuB,CAAC;AAC/B,OAAO,EAAE,oBAAoB,EAAE,MAAM,sBAAsB,CAAC"}

package/dist/internal.d.ts

@@ -18,7 +18,7 @@ export { runGraph, GRAPH_STAGES } from './cli/orchestrate.js';
 export type { GraphStage, GraphProgressEvent, GraphProgressCallback, RunGraphInput, RunGraphResult, } from './cli/orchestrate.js';
 export { executeGraph, buildUnifiedReportLines } from './cli/graph.js';
 export type { UnifiedReportInput } from './cli/graph.js';
-export { graphDoneTableNode } from './cli/graph-envelope-view.js';
+export { envelopeToLiveRunTableRows } from './cli/graph-envelope-view.js';
 export { CatalogRepo } from './persistence/catalog-repo.js';
 export { MemoryPressureError } from './cli/pressure-monitor.js';
 export { HEAP_TARGETS, decideHeapTargetMb, systemHasMemoryFor, runHeapPreflight, totalSystemMemoryMb, } from './cli/heap-preflight.js';

package/dist/internal.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"internal.d.ts","sourceRoot":"","sources":["../src/internal.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAKH,OAAO,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAC;AAKrD,YAAY,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAE9C,OAAO,EAAE,sBAAsB,EAAE,MAAM,iCAAiC,CAAC;AACzE,OAAO,EAAE,oBAAoB,EAAE,MAAM,gCAAgC,CAAC;AACtE,OAAO,EAAE,0BAA0B,EAAE,MAAM,qCAAqC,CAAC;AACjF,OAAO,EAAE,iBAAiB,EAAE,MAAM,2BAA2B,CAAC;AAU9D,OAAO,EAAE,QAAQ,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAC9D,YAAY,EACV,UAAU,EACV,kBAAkB,EAClB,qBAAqB,EACrB,aAAa,EACb,cAAc,GACf,MAAM,sBAAsB,CAAC;AAC9B,OAAO,EAAE,YAAY,EAAE,uBAAuB,EAAE,MAAM,gBAAgB,CAAC;AACvE,YAAY,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAC;AAMzD,OAAO,EAAE,kBAAkB,EAAE,MAAM,8BAA8B,CAAC;AAQlE,OAAO,EAAE,WAAW,EAAE,MAAM,+BAA+B,CAAC;AAC5D,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EACL,YAAY,EACZ,kBAAkB,EAClB,kBAAkB,EAClB,gBAAgB,EAChB,mBAAmB,GACpB,MAAM,yBAAyB,CAAC;AACjC,YAAY,EAAE,KAAK,EAAE,gBAAgB,EAAE,MAAM,kCAAkC,CAAC"}
+{"version":3,"file":"internal.d.ts","sourceRoot":"","sources":["../src/internal.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAKH,OAAO,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAC;AAKrD,YAAY,EAAE,WAAW,EAAE,MAAM,YAAY,CAAC;AAE9C,OAAO,EAAE,sBAAsB,EAAE,MAAM,iCAAiC,CAAC;AACzE,OAAO,EAAE,oBAAoB,EAAE,MAAM,gCAAgC,CAAC;AACtE,OAAO,EAAE,0BAA0B,EAAE,MAAM,qCAAqC,CAAC;AACjF,OAAO,EAAE,iBAAiB,EAAE,MAAM,2BAA2B,CAAC;AAU9D,OAAO,EAAE,QAAQ,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAC9D,YAAY,EACV,UAAU,EACV,kBAAkB,EAClB,qBAAqB,EACrB,aAAa,EACb,cAAc,GACf,MAAM,sBAAsB,CAAC;AAC9B,OAAO,EAAE,YAAY,EAAE,uBAAuB,EAAE,MAAM,gBAAgB,CAAC;AACvE,YAAY,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAC;AAIzD,OAAO,EAAE,0BAA0B,EAAE,MAAM,8BAA8B,CAAC;AAQ1E,OAAO,EAAE,WAAW,EAAE,MAAM,+BAA+B,CAAC;AAC5D,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EACL,YAAY,EACZ,kBAAkB,EAClB,kBAAkB,EAClB,gBAAgB,EAChB,mBAAmB,GACpB,MAAM,yBAAyB,CAAC;AACjC,YAAY,EAAE,KAAK,EAAE,gBAAgB,EAAE,MAAM,kCAAkC,CAAC"}

package/dist/internal.js

@@ -18,7 +18,7 @@ export { duplicatedFunctionBodyRule } from './rules/duplicated-function-body.js'
 export { orphanSubtreeRule } from './rules/orphan-subtree.js';
 // ── Orchestration / CLI-handler surface (ADR-0009, Finding 3) ──────
 //
-// These drive the six-stage pipeline and the `graph` CLI command. The
+// These drive the seven-stage pipeline and the `graph` CLI command. The
 // production path reaches them via `graphTool.commandSpecs` (host-mounted,
 // and the parent repo via the `catalog-export` subcommand), never by importing these
 // symbols — so they are private. Only the cross-package adapter and CLI
@@ -26,11 +26,9 @@ export { orphanSubtreeRule } from './rules/orphan-subtree.js';
 // end without going through Commander.
 export { runGraph, GRAPH_STAGES } from './cli/orchestrate.js';
 export { executeGraph, buildUnifiedReportLines } from './cli/graph.js';
-// The graph live verbose/detail per-unit table node. Internal — exported only
-// so the host-side live/static parity proof
-// (`packages/cli/src/ui/__tests__/graph-live-static-parity.test.tsx`) can assert
-// it renders byte-identically to the static `envelopeToTableView` table.
-export { graphDoneTableNode } from './cli/graph-envelope-view.js';
+// Graph envelope → live-run table rows. Internal — exported for the host-side
+// live/static parity proof (`graph-live-static-parity.test.tsx`).
+export { envelopeToLiveRunTableRows } from './cli/graph-envelope-view.js';
 // `CatalogRepo` — the engine's SQLite/Drizzle catalog persistence repo. Used
 // internally by graph (orchestrator cache, `tool.ts` report contribution,
 // `lookup`/`catalog-export`). It was briefly on the public barrel for fitness's

package/dist/internal.js.map

@@ -1 +1 @@
-{"version":3,"file":"internal.js","sourceRoot":"","sources":["../src/internal.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,yEAAyE;AACzE,+EAA+E;AAC/E,8DAA8D;AAC9D,OAAO,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAC;AAOrD,OAAO,EAAE,sBAAsB,EAAE,MAAM,iCAAiC,CAAC;AACzE,OAAO,EAAE,oBAAoB,EAAE,MAAM,gCAAgC,CAAC;AACtE,OAAO,EAAE,0BAA0B,EAAE,MAAM,qCAAqC,CAAC;AACjF,OAAO,EAAE,iBAAiB,EAAE,MAAM,2BAA2B,CAAC;AAE9D,sEAAsE;AACtE,EAAE;AACF,sEAAsE;AACtE,2EAA2E;AAC3E,qFAAqF;AACrF,wEAAwE;AACxE,wEAAwE;AACxE,uCAAuC;AACvC,OAAO,EAAE,QAAQ,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAQ9D,OAAO,EAAE,YAAY,EAAE,uBAAuB,EAAE,MAAM,gBAAgB,CAAC;AAGvE,8EAA8E;AAC9E,4CAA4C;AAC5C,iFAAiF;AACjF,yEAAyE;AACzE,OAAO,EAAE,kBAAkB,EAAE,MAAM,8BAA8B,CAAC;AAElE,6EAA6E;AAC7E,0EAA0E;AAC1E,gFAAgF;AAChF,0EAA0E;AAC1E,8EAA8E;AAC9E,kFAAkF;AAClF,OAAO,EAAE,WAAW,EAAE,MAAM,+BAA+B,CAAC;AAC5D,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EACL,YAAY,EACZ,kBAAkB,EAClB,kBAAkB,EAClB,gBAAgB,EAChB,mBAAmB,GACpB,MAAM,yBAAyB,CAAC"}
+{"version":3,"file":"internal.js","sourceRoot":"","sources":["../src/internal.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,yEAAyE;AACzE,+EAA+E;AAC/E,8DAA8D;AAC9D,OAAO,EAAE,YAAY,EAAE,MAAM,uBAAuB,CAAC;AAOrD,OAAO,EAAE,sBAAsB,EAAE,MAAM,iCAAiC,CAAC;AACzE,OAAO,EAAE,oBAAoB,EAAE,MAAM,gCAAgC,CAAC;AACtE,OAAO,EAAE,0BAA0B,EAAE,MAAM,qCAAqC,CAAC;AACjF,OAAO,EAAE,iBAAiB,EAAE,MAAM,2BAA2B,CAAC;AAE9D,sEAAsE;AACtE,EAAE;AACF,wEAAwE;AACxE,2EAA2E;AAC3E,qFAAqF;AACrF,wEAAwE;AACxE,wEAAwE;AACxE,uCAAuC;AACvC,OAAO,EAAE,QAAQ,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAQ9D,OAAO,EAAE,YAAY,EAAE,uBAAuB,EAAE,MAAM,gBAAgB,CAAC;AAGvE,8EAA8E;AAC9E,kEAAkE;AAClE,OAAO,EAAE,0BAA0B,EAAE,MAAM,8BAA8B,CAAC;AAE1E,6EAA6E;AAC7E,0EAA0E;AAC1E,gFAAgF;AAChF,0EAA0E;AAC1E,8EAA8E;AAC9E,kFAAkF;AAClF,OAAO,EAAE,WAAW,EAAE,MAAM,+BAA+B,CAAC;AAC5D,OAAO,EAAE,mBAAmB,EAAE,MAAM,2BAA2B,CAAC;AAChE,OAAO,EACL,YAAY,EACZ,kBAAkB,EAClB,kBAAkB,EAClB,gBAAgB,EAChB,mBAAmB,GACpB,MAAM,yBAAyB,CAAC"}

package/dist/lang-adapter/__tests__/language-of-file.test.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=language-of-file.test.d.ts.map

package/dist/lang-adapter/__tests__/language-of-file.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"language-of-file.test.d.ts","sourceRoot":"","sources":["../../../src/lang-adapter/__tests__/language-of-file.test.ts"],"names":[],"mappings":""}

package/dist/lang-adapter/__tests__/language-of-file.test.js

@@ -0,0 +1,36 @@
+import { describe, expect, it } from 'vitest';
+import { languageOfFile } from '../language-of-file.js';
+describe('languageOfFile', () => {
+    it('maps TypeScript extensions', () => {
+        expect(languageOfFile('src/a.ts')).toBe('typescript');
+        expect(languageOfFile('src/a.tsx')).toBe('typescript');
+        expect(languageOfFile('src/a.mts')).toBe('typescript');
+        expect(languageOfFile('src/a.cts')).toBe('typescript');
+    });
+    it('maps JavaScript extensions to typescript (same adapter; .ts↔.js clones are real)', () => {
+        expect(languageOfFile('src/a.js')).toBe('typescript');
+        expect(languageOfFile('src/a.jsx')).toBe('typescript');
+        expect(languageOfFile('src/a.mjs')).toBe('typescript');
+        expect(languageOfFile('src/a.cjs')).toBe('typescript');
+        // A .ts and a .js file are the same language → clones between them are detected.
+        expect(languageOfFile('a.ts')).toBe(languageOfFile('b.js'));
+    });
+    it('maps Python extensions without truncating .pyi', () => {
+        expect(languageOfFile('mod.py')).toBe('python');
+        expect(languageOfFile('stub.pyi')).toBe('python');
+    });
+    it('maps Go, Java, and Rust', () => {
+        expect(languageOfFile('main.go')).toBe('go');
+        expect(languageOfFile('App.java')).toBe('java');
+        expect(languageOfFile('lib.rs')).toBe('rust');
+    });
+    it('returns undefined for unknown extensions', () => {
+        expect(languageOfFile('readme.md')).toBeUndefined();
+        expect(languageOfFile('noext')).toBeUndefined();
+    });
+    it('two files share a language iff languageOfFile is equal', () => {
+        expect(languageOfFile('a.ts')).toBe(languageOfFile('b.tsx'));
+        expect(languageOfFile('a.go')).not.toBe(languageOfFile('b.ts'));
+    });
+});
+//# sourceMappingURL=language-of-file.test.js.map

package/dist/lang-adapter/__tests__/language-of-file.test.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"language-of-file.test.js","sourceRoot":"","sources":["../../../src/lang-adapter/__tests__/language-of-file.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,QAAQ,CAAC;AAE9C,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AAExD,QAAQ,CAAC,gBAAgB,EAAE,GAAG,EAAE;IAC9B,EAAE,CAAC,4BAA4B,EAAE,GAAG,EAAE;QACpC,MAAM,CAAC,cAAc,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QACtD,MAAM,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QACvD,MAAM,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QACvD,MAAM,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;IACzD,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,kFAAkF,EAAE,GAAG,EAAE;QAC1F,MAAM,CAAC,cAAc,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QACtD,MAAM,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QACvD,MAAM,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QACvD,MAAM,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;QACvD,iFAAiF;QACjF,MAAM,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC;IAC9D,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,gDAAgD,EAAE,GAAG,EAAE;QACxD,MAAM,CAAC,cAAc,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAChD,MAAM,CAAC,cAAc,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IACpD,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,yBAAyB,EAAE,GAAG,EAAE;QACjC,MAAM,CAAC,cAAc,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC7C,MAAM,CAAC,cAAc,CAAC,UAAU,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAChD,MAAM,CAAC,cAAc,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAChD,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,0CAA0C,EAAE,GAAG,EAAE;QAClD,MAAM,CAAC,cAAc,CAAC,WAAW,CAAC,CAAC,CAAC,aAAa,EAAE,CAAC;QACpD,MAAM,CAAC,cAAc,CAAC,OAAO,CAAC,CAAC,CAAC,aAAa,EAAE,CAAC;IAClD,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,wDAAwD,EAAE,GAAG,EAAE;QAChE,MAAM,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,cAAc,CAAC,OAAO,CAAC,CAAC,CAAC;QAC7D,MAAM,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC;IAClE,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}

package/dist/lang-adapter/__tests__/near-duplicate-signature.test.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=near-duplicate-signature.test.d.ts.map

package/dist/lang-adapter/__tests__/near-duplicate-signature.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"near-duplicate-signature.test.d.ts","sourceRoot":"","sources":["../../../src/lang-adapter/__tests__/near-duplicate-signature.test.ts"],"names":[],"mappings":""}

package/dist/lang-adapter/__tests__/near-duplicate-signature.test.js

@@ -0,0 +1,78 @@
+import { describe, expect, it } from 'vitest';
+import { NEAR_DUP_LSH_BANDS, NEAR_DUP_LSH_ROWS, NEAR_DUP_SIGNATURE_K, bodySignature, digestCanonicalBody, estimateJaccard, lshBandHashes, shingle, } from '../near-duplicate-signature.js';
+describe('near-duplicate-signature', () => {
+    const base = 'function processItems(items) { const out = []; for (const item of items) { out.push(transform(item)); } return out; }';
+    const nearEdit = 'function processItems(items) { const result = []; for (const item of items) { result.push(transform(item)); } return result; }';
+    const unrelated = 'export function validateConfig(cfg) { if (!cfg.apiKey) throw new Error("missing"); return cfg; }';
+    it('identical canonical text yields Jaccard 1.0', () => {
+        const a = bodySignature(base);
+        const b = bodySignature(base);
+        expect(estimateJaccard(a, b)).toBe(1);
+    });
+    it('one-token edit yields high but sub-1.0 Jaccard', () => {
+        const j = estimateJaccard(bodySignature(base), bodySignature(nearEdit));
+        expect(j).toBeGreaterThan(0.5);
+        expect(j).toBeLessThan(1);
+    });
+    it('unrelated bodies yield low Jaccard', () => {
+        expect(estimateJaccard(bodySignature(base), bodySignature(unrelated))).toBeLessThan(0.5);
+    });
+    it('signatures are stable across runs', () => {
+        expect(bodySignature(base)).toEqual(bodySignature(base));
+    });
+    it('bands × rows equals k', () => {
+        expect(NEAR_DUP_LSH_BANDS * NEAR_DUP_LSH_ROWS).toBe(NEAR_DUP_SIGNATURE_K);
+    });
+    it('LSH knee is near the 0.85 threshold', () => {
+        const knee = (1 / NEAR_DUP_LSH_BANDS) ** (1 / NEAR_DUP_LSH_ROWS);
+        expect(knee).toBeCloseTo(0.878, 2);
+    });
+    it('shingle emits char 5-grams', () => {
+        expect(shingle('abcdef').size).toBe(2);
+        expect(shingle('ab').size).toBe(1);
+        expect(shingle('').size).toBe(0);
+    });
+    it('digestCanonicalBody returns hash, size, and signature', () => {
+        const d = digestCanonicalBody(base);
+        expect(d.hash).toMatch(/^[a-f0-9]{64}$/);
+        expect(d.size).toBeGreaterThan(0);
+        expect(d.signature?.length).toBe(NEAR_DUP_SIGNATURE_K);
+    });
+    it('lshBandHashes returns one hash per band', () => {
+        const sig = bodySignature(base);
+        expect(lshBandHashes(sig, NEAR_DUP_LSH_BANDS, NEAR_DUP_LSH_ROWS).length).toBe(NEAR_DUP_LSH_BANDS);
+    });
+    it('bodySignature returns empty for empty canonical (no shingles)', () => {
+        expect(bodySignature('')).toEqual([]);
+    });
+    it('bodySignature with custom k derives k seeds and returns k values', () => {
+        const sig = bodySignature(base, 32);
+        expect(sig.length).toBe(32);
+        // Custom-k path is deterministic and stable across calls.
+        expect(bodySignature(base, 32)).toEqual(sig);
+        // Identical body still self-matches under a custom k.
+        expect(estimateJaccard(sig, bodySignature(base, 32))).toBe(1);
+    });
+    it('bodySignature produces a full-width signature for sub-gram bodies', () => {
+        // Body shorter than the gram size yields a single whole-string shingle,
+        // exercising the single-base MinHash loop across all k positions.
+        const short = bodySignature('ab');
+        expect(short.length).toBe(NEAR_DUP_SIGNATURE_K);
+        expect(short.every((v) => Number.isInteger(v))).toBe(true);
+    });
+    it('estimateJaccard returns 0 when either signature is empty', () => {
+        const sig = bodySignature(base);
+        expect(estimateJaccard([], sig)).toBe(0);
+        expect(estimateJaccard(sig, [])).toBe(0);
+        expect(estimateJaccard([], [])).toBe(0);
+    });
+    it('estimateJaccard returns 0 for mismatched signature lengths', () => {
+        expect(estimateJaccard(bodySignature(base), bodySignature(base, 32))).toBe(0);
+    });
+    it('digestCanonicalBody omits signature for empty canonical', () => {
+        const d = digestCanonicalBody('');
+        expect(d.hash).toMatch(/^[a-f0-9]{64}$/);
+        expect(d.signature).toBeUndefined();
+    });
+});
+//# sourceMappingURL=near-duplicate-signature.test.js.map

package/dist/lang-adapter/__tests__/near-duplicate-signature.test.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"near-duplicate-signature.test.js","sourceRoot":"","sources":["../../../src/lang-adapter/__tests__/near-duplicate-signature.test.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,QAAQ,CAAC;AAE9C,OAAO,EACL,kBAAkB,EAClB,iBAAiB,EACjB,oBAAoB,EACpB,aAAa,EACb,mBAAmB,EACnB,eAAe,EACf,aAAa,EACb,OAAO,GACR,MAAM,gCAAgC,CAAC;AAExC,QAAQ,CAAC,0BAA0B,EAAE,GAAG,EAAE;IACxC,MAAM,IAAI,GACR,uHAAuH,CAAC;IAC1H,MAAM,QAAQ,GACZ,gIAAgI,CAAC;IACnI,MAAM,SAAS,GACb,kGAAkG,CAAC;IAErG,EAAE,CAAC,6CAA6C,EAAE,GAAG,EAAE;QACrD,MAAM,CAAC,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;QAC9B,MAAM,CAAC,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;QAC9B,MAAM,CAAC,eAAe,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACxC,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,gDAAgD,EAAE,GAAG,EAAE;QACxD,MAAM,CAAC,GAAG,eAAe,CAAC,aAAa,CAAC,IAAI,CAAC,EAAE,aAAa,CAAC,QAAQ,CAAC,CAAC,CAAC;QACxE,MAAM,CAAC,CAAC,CAAC,CAAC,eAAe,CAAC,GAAG,CAAC,CAAC;QAC/B,MAAM,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;IAC5B,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,oCAAoC,EAAE,GAAG,EAAE;QAC5C,MAAM,CAAC,eAAe,CAAC,aAAa,CAAC,IAAI,CAAC,EAAE,aAAa,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC;IAC3F,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,mCAAmC,EAAE,GAAG,EAAE;QAC3C,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,OAAO,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC;IAC3D,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,uBAAuB,EAAE,GAAG,EAAE;QAC/B,MAAM,CAAC,kBAAkB,GAAG,iBAAiB,CAAC,CAAC,IAAI,CAAC,oBAAoB,CAAC,CAAC;IAC5E,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,qCAAqC,EAAE,GAAG,EAAE;QAC7C,MAAM,IAAI,GAAG,CAAC,CAAC,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAC,GAAG,iBAAiB,CAAC,CAAC;QACjE,MAAM,CAAC,IAAI,CAAC,CAAC,WAAW,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC;IACrC,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,4BAA4B,EAAE,GAAG,EAAE;QACpC,MAAM,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACvC,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACnC,MAAM,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IACnC,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,uDAAuD,EAAE,GAAG,EAAE;QAC/D,MAAM,CAAC,GAAG,mBAAmB,CAAC,IAAI,CAAC,CAAC;QACpC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC;QACzC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;QAClC,MAAM,CAAC,CAAC,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC,IAAI,CAAC,oBAAoB,CAAC,CAAC;IACzD,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,yCAAyC,EAAE,GAAG,EAAE;QACjD,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;QAChC,MAAM,CAAC,aAAa,CAAC,GAAG,EAAE,kBAAkB,EAAE,iBAAiB,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,CAC3E,kBAAkB,CACnB,CAAC;IACJ,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,+DAA+D,EAAE,GAAG,EAAE;QACvE,MAAM,CAAC,aAAa,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IACxC,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,kEAAkE,EAAE,GAAG,EAAE;QAC1E,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;QACpC,MAAM,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAC5B,0DAA0D;QAC1D,MAAM,CAAC,aAAa,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QAC7C,sDAAsD;QACtD,MAAM,CAAC,eAAe,CAAC,GAAG,EAAE,aAAa,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAChE,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,mEAAmE,EAAE,GAAG,EAAE;QAC3E,wEAAwE;QACxE,kEAAkE;QAClE,MAAM,KAAK,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;QAClC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,oBAAoB,CAAC,CAAC;QAChD,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC7D,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,0DAA0D,EAAE,GAAG,EAAE;QAClE,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,CAAC,CAAC;QAChC,MAAM,CAAC,eAAe,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACzC,MAAM,CAAC,eAAe,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACzC,MAAM,CAAC,eAAe,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC1C,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,4DAA4D,EAAE,GAAG,EAAE;QACpE,MAAM,CAAC,eAAe,CAAC,aAAa,CAAC,IAAI,CAAC,EAAE,aAAa,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAChF,CAAC,CAAC,CAAC;IAEH,EAAE,CAAC,yDAAyD,EAAE,GAAG,EAAE;QACjE,MAAM,CAAC,GAAG,mBAAmB,CAAC,EAAE,CAAC,CAAC;QAClC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,gBAAgB,CAAC,CAAC;QACzC,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,aAAa,EAAE,CAAC;IACtC,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC"}

package/dist/lang-adapter/body-digest.d.ts

@@ -1,46 +1,14 @@
 /**
- * @fileoverview Shared body-digest primitives for GraphLanguageAdapters.
+ * @fileoverview Body-digest primitives — relocated to `@opensip-cli/clone-detection`.
  *
- * The catalog's `bodyHash` contract is "reduce a function body to a
- * canonical string, then SHA-256 it." The normalize-to-hash TAIL of that
- * pipeline — the `BodyDigest` shape, the whitespace-collapse normalizer,
- * and the hash+size step — was byte-identical across every adapter that
- * hashes bodies (go/java/rust real bodies, plus the synthetic-body path
- * in all four tree-sitter packs). Per the round-3 modular-monolith audit
- * (2026-05-30, finding D) it lives here on the contract layer instead of
- * being copy-pasted into each pack.
+ * The hash/normalize tail (`BodyDigest`, `normalizeWhitespace`, `hashBody`) moved to
+ * the shared layer-2 substrate (ADR-0064) so both graph and yagni single-source it and
+ * cannot diverge. This module re-exports the same symbols verbatim so every existing
+ * graph-engine + adapter importer of `'../lang-adapter/body-digest.js'` is unchanged
+ * and `bodyHash` values stay byte-identical.
  *
- * This mirrors edge-helpers.ts exactly: tiny utilities that used to be
- * duplicated under each adapter now sit in lang-adapter/, which adapters
- * already import types from — so importing one more helper is
- * structurally consistent and violates no layering rule (lang-* adapters
- * may not import each OTHER, but lang-adapter/ is fair game).
- *
- * What deliberately STAYS in each adapter pack: the language-specific
- * comment stripper (Go rune literals vs Java text blocks vs Rust nested
- * block comments) and any language-specific normalization. Python, for
- * example, substitutes its own `normalizePythonBody` for real bodies
- * because indentation is semantically significant there — it still
- * composes `hashBody` + `normalizeWhitespace` for synthetic bodies. The
- * primitives are intentionally low-level (not a single `digestBody(text,
- * stripFn)`) precisely so a pack can swap the normalization step.
- */
-/** Catalog body fingerprint: SHA-256 hex hash + canonical-text length. */
-export interface BodyDigest {
-    readonly hash: string;
-    readonly size: number;
-}
-/**
- * Collapse every run of whitespace to a single space and trim. The
- * default body normalization for whitespace-insensitive languages
- * (Go/Java/Rust) and for synthetic (module-init) bodies in every pack.
- */
-export declare function normalizeWhitespace(s: string): string;
-/**
- * SHA-256 an already-normalized ("canonical") body string into a
- * `BodyDigest`. `size` is the canonical-text length so it tracks the
- * hashed content, not the raw source. Callers normalize first — via
- * `normalizeWhitespace` or a language-specific normalizer.
+ * Language-specific comment strippers / normalizers still live in each adapter pack;
+ * only the canonical-string→hash tail is shared (see the package).
  */
-export declare function hashBody(canonical: string): BodyDigest;
+export { normalizeWhitespace, hashBody, type BodyDigest } from '@opensip-cli/clone-detection';
 //# sourceMappingURL=body-digest.d.ts.map

package/dist/lang-adapter/body-digest.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"body-digest.d.ts","sourceRoot":"","sources":["../../src/lang-adapter/body-digest.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AAIH,0EAA0E;AAC1E,MAAM,WAAW,UAAU;IACzB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;CACvB;AAED;;;;GAIG;AACH,wBAAgB,mBAAmB,CAAC,CAAC,EAAE,MAAM,GAAG,MAAM,CAErD;AAED;;;;;GAKG;AACH,wBAAgB,QAAQ,CAAC,SAAS,EAAE,MAAM,GAAG,UAAU,CAKtD"}
+{"version":3,"file":"body-digest.d.ts","sourceRoot":"","sources":["../../src/lang-adapter/body-digest.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,OAAO,EAAE,mBAAmB,EAAE,QAAQ,EAAE,KAAK,UAAU,EAAE,MAAM,8BAA8B,CAAC"}

package/dist/lang-adapter/body-digest.js

@@ -1,49 +1,14 @@
 /**
- * @fileoverview Shared body-digest primitives for GraphLanguageAdapters.
+ * @fileoverview Body-digest primitives — relocated to `@opensip-cli/clone-detection`.
  *
- * The catalog's `bodyHash` contract is "reduce a function body to a
- * canonical string, then SHA-256 it." The normalize-to-hash TAIL of that
- * pipeline — the `BodyDigest` shape, the whitespace-collapse normalizer,
- * and the hash+size step — was byte-identical across every adapter that
- * hashes bodies (go/java/rust real bodies, plus the synthetic-body path
- * in all four tree-sitter packs). Per the round-3 modular-monolith audit
- * (2026-05-30, finding D) it lives here on the contract layer instead of
- * being copy-pasted into each pack.
+ * The hash/normalize tail (`BodyDigest`, `normalizeWhitespace`, `hashBody`) moved to
+ * the shared layer-2 substrate (ADR-0064) so both graph and yagni single-source it and
+ * cannot diverge. This module re-exports the same symbols verbatim so every existing
+ * graph-engine + adapter importer of `'../lang-adapter/body-digest.js'` is unchanged
+ * and `bodyHash` values stay byte-identical.
  *
- * This mirrors edge-helpers.ts exactly: tiny utilities that used to be
- * duplicated under each adapter now sit in lang-adapter/, which adapters
- * already import types from — so importing one more helper is
- * structurally consistent and violates no layering rule (lang-* adapters
- * may not import each OTHER, but lang-adapter/ is fair game).
- *
- * What deliberately STAYS in each adapter pack: the language-specific
- * comment stripper (Go rune literals vs Java text blocks vs Rust nested
- * block comments) and any language-specific normalization. Python, for
- * example, substitutes its own `normalizePythonBody` for real bodies
- * because indentation is semantically significant there — it still
- * composes `hashBody` + `normalizeWhitespace` for synthetic bodies. The
- * primitives are intentionally low-level (not a single `digestBody(text,
- * stripFn)`) precisely so a pack can swap the normalization step.
- */
-import { createHash } from 'node:crypto';
-/**
- * Collapse every run of whitespace to a single space and trim. The
- * default body normalization for whitespace-insensitive languages
- * (Go/Java/Rust) and for synthetic (module-init) bodies in every pack.
- */
-export function normalizeWhitespace(s) {
-    return s.replaceAll(/\s+/g, ' ').trim();
-}
-/**
- * SHA-256 an already-normalized ("canonical") body string into a
- * `BodyDigest`. `size` is the canonical-text length so it tracks the
- * hashed content, not the raw source. Callers normalize first — via
- * `normalizeWhitespace` or a language-specific normalizer.
+ * Language-specific comment strippers / normalizers still live in each adapter pack;
+ * only the canonical-string→hash tail is shared (see the package).
  */
-export function hashBody(canonical) {
-    return {
-        hash: createHash('sha256').update(canonical, 'utf8').digest('hex'),
-        size: canonical.length,
-    };
-}
+export { normalizeWhitespace, hashBody } from '@opensip-cli/clone-detection';
 //# sourceMappingURL=body-digest.js.map

package/dist/lang-adapter/body-digest.js.map

@@ -1 +1 @@
-{"version":3,"file":"body-digest.js","sourceRoot":"","sources":["../../src/lang-adapter/body-digest.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAQzC;;;;GAIG;AACH,MAAM,UAAU,mBAAmB,CAAC,CAAS;IAC3C,OAAO,CAAC,CAAC,UAAU,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;AAC1C,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,QAAQ,CAAC,SAAiB;IACxC,OAAO;QACL,IAAI,EAAE,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC;QAClE,IAAI,EAAE,SAAS,CAAC,MAAM;KACvB,CAAC;AACJ,CAAC"}
+{"version":3,"file":"body-digest.js","sourceRoot":"","sources":["../../src/lang-adapter/body-digest.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AAEH,OAAO,EAAE,mBAAmB,EAAE,QAAQ,EAAmB,MAAM,8BAA8B,CAAC"}

package/dist/lang-adapter/language-of-file.d.ts

@@ -0,0 +1,15 @@
+/**
+ * Canonical language id by file extension — engine-homed for rule gating.
+ *
+ * The near-clone rule must gate candidate pairs to the same language. The
+ * equivalent filter in graph-adapter-common cannot be imported from the engine
+ * rule layer (adapter-common depends on the engine). This module is the
+ * canonical extension map consumed by rules; adapter-common may delegate here
+ * in a follow-up.
+ */
+/**
+ * Return the canonical language id for `filePath` by extension, or `undefined`
+ * when the extension is not recognized.
+ */
+export declare function languageOfFile(filePath: string): string | undefined;
+//# sourceMappingURL=language-of-file.d.ts.map

package/dist/lang-adapter/language-of-file.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"language-of-file.d.ts","sourceRoot":"","sources":["../../src/lang-adapter/language-of-file.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AA0BH;;;GAGG;AACH,wBAAgB,cAAc,CAAC,QAAQ,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAKnE"}

package/dist/lang-adapter/language-of-file.js

@@ -0,0 +1,43 @@
+/**
+ * Canonical language id by file extension — engine-homed for rule gating.
+ *
+ * The near-clone rule must gate candidate pairs to the same language. The
+ * equivalent filter in graph-adapter-common cannot be imported from the engine
+ * rule layer (adapter-common depends on the engine). This module is the
+ * canonical extension map consumed by rules; adapter-common may delegate here
+ * in a follow-up.
+ */
+/**
+ * Extension suffix → canonical graph language id. Longest-match first per path.
+ * `.js`/`.jsx`/`.mjs`/`.cjs` map to `typescript` because the graph-typescript
+ * adapter analyzes them and a `.ts`↔`.js` (or `.js`↔`.js`) near-clone is a real,
+ * actionable clone — omitting them would silently skip every JS clone pair.
+ */
+const EXTENSION_TO_LANGUAGE = {
+    '.tsx': 'typescript',
+    '.mts': 'typescript',
+    '.cts': 'typescript',
+    '.ts': 'typescript',
+    '.jsx': 'typescript',
+    '.mjs': 'typescript',
+    '.cjs': 'typescript',
+    '.js': 'typescript',
+    '.pyi': 'python',
+    '.py': 'python',
+    '.java': 'java',
+    '.go': 'go',
+    '.rs': 'rust',
+};
+const SORTED_EXTENSIONS = Object.keys(EXTENSION_TO_LANGUAGE).sort((a, b) => b.length - a.length);
+/**
+ * Return the canonical language id for `filePath` by extension, or `undefined`
+ * when the extension is not recognized.
+ */
+export function languageOfFile(filePath) {
+    for (const ext of SORTED_EXTENSIONS) {
+        if (filePath.endsWith(ext))
+            return EXTENSION_TO_LANGUAGE[ext];
+    }
+    return undefined;
+}
+//# sourceMappingURL=language-of-file.js.map

package/dist/lang-adapter/language-of-file.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"language-of-file.js","sourceRoot":"","sources":["../../src/lang-adapter/language-of-file.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH;;;;;GAKG;AACH,MAAM,qBAAqB,GAAqC;IAC9D,MAAM,EAAE,YAAY;IACpB,MAAM,EAAE,YAAY;IACpB,MAAM,EAAE,YAAY;IACpB,KAAK,EAAE,YAAY;IACnB,MAAM,EAAE,YAAY;IACpB,MAAM,EAAE,YAAY;IACpB,MAAM,EAAE,YAAY;IACpB,KAAK,EAAE,YAAY;IACnB,MAAM,EAAE,QAAQ;IAChB,KAAK,EAAE,QAAQ;IACf,OAAO,EAAE,MAAM;IACf,KAAK,EAAE,IAAI;IACX,KAAK,EAAE,MAAM;CACd,CAAC;AAEF,MAAM,iBAAiB,GAAG,MAAM,CAAC,IAAI,CAAC,qBAAqB,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC;AAEjG;;;GAGG;AACH,MAAM,UAAU,cAAc,CAAC,QAAgB;IAC7C,KAAK,MAAM,GAAG,IAAI,iBAAiB,EAAE,CAAC;QACpC,IAAI,QAAQ,CAAC,QAAQ,CAAC,GAAG,CAAC;YAAE,OAAO,qBAAqB,CAAC,GAAG,CAAC,CAAC;IAChE,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC"}

package/dist/lang-adapter/near-duplicate-signature.d.ts

@@ -0,0 +1,12 @@
+/**
+ * @fileoverview Near-duplicate MinHash/LSH primitives — relocated to
+ * `@opensip-cli/clone-detection`.
+ *
+ * The signature constants + digest tail moved to the shared layer-2 substrate
+ * (ADR-0064) so graph and yagni single-source them. This module re-exports the same
+ * symbols verbatim so existing importers of `'../lang-adapter/near-duplicate-signature.js'`
+ * are unchanged and signature VALUES (and the `sig=` cache-key segment) stay
+ * byte-identical. `NEAR_DUP_SIGNATURE_VERSION` is unchanged by the move.
+ */
+export { NEAR_DUP_SIGNATURE_K, NEAR_DUP_LSH_BANDS, NEAR_DUP_LSH_ROWS, NEAR_DUP_SIGNATURE_VERSION, shingle, bodySignature, estimateJaccard, lshBandHashes, digestCanonicalBody, type BodyDigestWithSignature, } from '@opensip-cli/clone-detection';
+//# sourceMappingURL=near-duplicate-signature.d.ts.map

package/dist/lang-adapter/near-duplicate-signature.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"near-duplicate-signature.d.ts","sourceRoot":"","sources":["../../src/lang-adapter/near-duplicate-signature.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EACL,oBAAoB,EACpB,kBAAkB,EAClB,iBAAiB,EACjB,0BAA0B,EAC1B,OAAO,EACP,aAAa,EACb,eAAe,EACf,aAAa,EACb,mBAAmB,EACnB,KAAK,uBAAuB,GAC7B,MAAM,8BAA8B,CAAC"}

package/dist/lang-adapter/near-duplicate-signature.js

@@ -0,0 +1,12 @@
+/**
+ * @fileoverview Near-duplicate MinHash/LSH primitives — relocated to
+ * `@opensip-cli/clone-detection`.
+ *
+ * The signature constants + digest tail moved to the shared layer-2 substrate
+ * (ADR-0064) so graph and yagni single-source them. This module re-exports the same
+ * symbols verbatim so existing importers of `'../lang-adapter/near-duplicate-signature.js'`
+ * are unchanged and signature VALUES (and the `sig=` cache-key segment) stay
+ * byte-identical. `NEAR_DUP_SIGNATURE_VERSION` is unchanged by the move.
+ */
+export { NEAR_DUP_SIGNATURE_K, NEAR_DUP_LSH_BANDS, NEAR_DUP_LSH_ROWS, NEAR_DUP_SIGNATURE_VERSION, shingle, bodySignature, estimateJaccard, lshBandHashes, digestCanonicalBody, } from '@opensip-cli/clone-detection';
+//# sourceMappingURL=near-duplicate-signature.js.map

package/dist/lang-adapter/near-duplicate-signature.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"near-duplicate-signature.js","sourceRoot":"","sources":["../../src/lang-adapter/near-duplicate-signature.ts"],"names":[],"mappings":"AAAA;;;;;;;;;GASG;AAEH,OAAO,EACL,oBAAoB,EACpB,kBAAkB,EAClB,iBAAiB,EACjB,0BAA0B,EAC1B,OAAO,EACP,aAAa,EACb,eAAe,EACf,aAAa,EACb,mBAAmB,GAEpB,MAAM,8BAA8B,CAAC"}

package/dist/lang-adapter/read-source.d.ts

@@ -0,0 +1,20 @@
+/**
+ * Maximum bytes a single source file may occupy when read for parsing (10 MB).
+ * Mirrors the fitness engine's `FileAccessor` `FILE_TOO_LARGE` ceiling.
+ */
+export declare const MAX_SOURCE_FILE_BYTES = 10000000;
+/**
+ * Read a source file as UTF-8 with a per-file size guard. Throws if the file
+ * exceeds `maxBytes` (default {@link MAX_SOURCE_FILE_BYTES}) — by `statSync`
+ * size (checked before the read, so an oversized file is never loaded) or by
+ * post-read length (closing the stat→read TOCTOU window). A missing/unreadable
+ * file surfaces the underlying `statSync` error, exactly as the prior direct
+ * `readFileSync` surfaced its own.
+ *
+ * @throws {Error} When the file exceeds `maxBytes` (by `statSync` size or
+ *   post-read length), or when the underlying `statSync`/`readFileSync` fails
+ *   (e.g. a missing or unreadable file). Adapter parse loops catch this per file
+ *   and record a `ParseError`, so the throw is non-fatal.
+ */
+export declare function readSourceFileGuarded(path: string, maxBytes?: number): string;
+//# sourceMappingURL=read-source.d.ts.map

package/dist/lang-adapter/read-source.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"read-source.d.ts","sourceRoot":"","sources":["../../src/lang-adapter/read-source.ts"],"names":[],"mappings":"AAmBA;;;GAGG;AACH,eAAO,MAAM,qBAAqB,WAAa,CAAC;AAEhD;;;;;;;;;;;;GAYG;AACH,wBAAgB,qBAAqB,CACnC,IAAI,EAAE,MAAM,EACZ,QAAQ,GAAE,MAA8B,GACvC,MAAM,CAYR"}

package/dist/lang-adapter/read-source.js

@@ -0,0 +1,48 @@
+/**
+ * read-source — the single size-guarded source-file reader for graph language
+ * adapters.
+ *
+ * Adapter parse steps read project SOURCE files into memory before handing them
+ * to a parser. Reading a whole file is unavoidable, but an *unbounded* read lets
+ * one pathological file (a multi-hundred-MB generated/minified bundle) spike
+ * graph memory. This wrapper caps a single source file at
+ * {@link MAX_SOURCE_FILE_BYTES} (10 MB — the same ceiling the fitness engine's
+ * `FileAccessor` enforces, so the two tools share one source-size policy), using
+ * a `statSync` pre-check so an oversized file is rejected WITHOUT being loaded.
+ *
+ * On oversize it THROWS. Adapter parse loops already wrap each per-file read in a
+ * try/catch that records a `ParseError` and continues, so an oversized file is
+ * skipped with a recorded diagnostic — never fatal, consistent with how a
+ * read/parse failure on any single file is already handled.
+ */
+import { readFileSync, statSync } from 'node:fs';
+/**
+ * Maximum bytes a single source file may occupy when read for parsing (10 MB).
+ * Mirrors the fitness engine's `FileAccessor` `FILE_TOO_LARGE` ceiling.
+ */
+export const MAX_SOURCE_FILE_BYTES = 10_000_000;
+/**
+ * Read a source file as UTF-8 with a per-file size guard. Throws if the file
+ * exceeds `maxBytes` (default {@link MAX_SOURCE_FILE_BYTES}) — by `statSync`
+ * size (checked before the read, so an oversized file is never loaded) or by
+ * post-read length (closing the stat→read TOCTOU window). A missing/unreadable
+ * file surfaces the underlying `statSync` error, exactly as the prior direct
+ * `readFileSync` surfaced its own.
+ *
+ * @throws {Error} When the file exceeds `maxBytes` (by `statSync` size or
+ *   post-read length), or when the underlying `statSync`/`readFileSync` fails
+ *   (e.g. a missing or unreadable file). Adapter parse loops catch this per file
+ *   and record a `ParseError`, so the throw is non-fatal.
+ */
+export function readSourceFileGuarded(path, maxBytes = MAX_SOURCE_FILE_BYTES) {
+    const { size } = statSync(path);
+    if (size > maxBytes) {
+        throw new Error(`source file exceeds ${maxBytes}-byte size guard (${size} bytes): ${path}`);
+    }
+    const text = readFileSync(path, 'utf8');
+    if (text.length > maxBytes) {
+        throw new Error(`source file exceeds ${maxBytes}-byte size guard after read (${text.length} bytes): ${path}`);
+    }
+    return text;
+}
+//# sourceMappingURL=read-source.js.map

package/dist/lang-adapter/read-source.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"read-source.js","sourceRoot":"","sources":["../../src/lang-adapter/read-source.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AACH,OAAO,EAAE,YAAY,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AAEjD;;;GAGG;AACH,MAAM,CAAC,MAAM,qBAAqB,GAAG,UAAU,CAAC;AAEhD;;;;;;;;;;;;GAYG;AACH,MAAM,UAAU,qBAAqB,CACnC,IAAY,EACZ,WAAmB,qBAAqB;IAExC,MAAM,EAAE,IAAI,EAAE,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC;IAChC,IAAI,IAAI,GAAG,QAAQ,EAAE,CAAC;QACpB,MAAM,IAAI,KAAK,CAAC,uBAAuB,QAAQ,qBAAqB,IAAI,YAAY,IAAI,EAAE,CAAC,CAAC;IAC9F,CAAC;IACD,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;IACxC,IAAI,IAAI,CAAC,MAAM,GAAG,QAAQ,EAAE,CAAC;QAC3B,MAAM,IAAI,KAAK,CACb,uBAAuB,QAAQ,gCAAgC,IAAI,CAAC,MAAM,YAAY,IAAI,EAAE,CAC7F,CAAC;IACJ,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC"}

package/dist/lang-adapter/read-source.test.d.ts

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=read-source.test.d.ts.map

package/dist/lang-adapter/read-source.test.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"read-source.test.d.ts","sourceRoot":"","sources":["../../src/lang-adapter/read-source.test.ts"],"names":[],"mappings":""}

package/dist/lang-adapter/read-source.test.js

@@ -0,0 +1,41 @@
+/**
+ * read-source — per-file source size guard for adapter parse steps.
+ *
+ * Proves the guard rejects an oversized file (the unbounded-memory hardening)
+ * and otherwise behaves like a plain UTF-8 read, so the adapters' existing
+ * per-file try/catch records a ParseError + continues on the throw.
+ */
+import { mkdtempSync, rmSync, writeFileSync } from 'node:fs';
+import { tmpdir } from 'node:os';
+import { join } from 'node:path';
+import { afterEach, describe, expect, it } from 'vitest';
+import { MAX_SOURCE_FILE_BYTES, readSourceFileGuarded } from './read-source.js';
+describe('readSourceFileGuarded', () => {
+    let dir;
+    afterEach(() => {
+        if (dir !== undefined)
+            rmSync(dir, { recursive: true, force: true });
+        dir = undefined;
+    });
+    function tmpFile(name, content) {
+        dir = mkdtempSync(join(tmpdir(), 'read-source-'));
+        const filePath = join(dir, name);
+        writeFileSync(filePath, content, 'utf8');
+        return filePath;
+    }
+    it('reads a normal source file as UTF-8', () => {
+        const filePath = tmpFile('a.ts', 'export const x = 1;\n');
+        expect(readSourceFileGuarded(filePath)).toBe('export const x = 1;\n');
+    });
+    it('throws (does not load) when the file exceeds the byte guard', () => {
+        const filePath = tmpFile('big.ts', 'x'.repeat(64));
+        expect(() => readSourceFileGuarded(filePath, 16)).toThrow(/size guard/);
+    });
+    it('defaults to the shared 10MB ceiling', () => {
+        expect(MAX_SOURCE_FILE_BYTES).toBe(10_000_000);
+    });
+    it('surfaces a stat error for a missing file (so the adapter catch records a ParseError)', () => {
+        expect(() => readSourceFileGuarded(join(tmpdir(), 'opensip-missing-source-xyz.ts'))).toThrow();
+    });
+});
+//# sourceMappingURL=read-source.test.js.map