@opensip-cli/clone-detection 0.1.12

package/dist/find-near-duplicates.js

@@ -0,0 +1,247 @@
+/**
+ * Near-duplicate body detection — LSH-banded MinHash, the single implementation graph's
+ * `near-duplicate-function-body` rule consumes (ADR-0064). Pure over `CloneCandidate[]`;
+ * emits tool-agnostic clusters (no `Signal`). Candidate pairs exclude identical
+ * `bodyHash` edges and cross-language false positives (via the pre-resolved
+ * `candidate.language`).
+ */
+import { isEligibleKind } from './find-duplicate-bodies.js';
+import { NEAR_DUP_LSH_BANDS, NEAR_DUP_SIGNATURE_K, estimateJaccard, lshBandHashes, } from './near-duplicate-signature.js';
+const DEFAULT_MIN_SIMILARITY = 0.85;
+const DEFAULT_MIN_BODY_SIZE = 200;
+const MAX_CLUSTER_SIZE = 50;
+/**
+ * Detect near-duplicate body clusters (Jaccard ≥ `minSimilarity`, same language, distinct
+ * exact hash). Returns one cluster per connected component with ≥2 near members.
+ */
+export function findNearDuplicates(candidates, opts = {}) {
+    const minSimilarity = opts.minSimilarity ?? DEFAULT_MIN_SIMILARITY;
+    const minBodySize = opts.minBodySize ?? DEFAULT_MIN_BODY_SIZE;
+    const bands = opts.lshBands ?? NEAR_DUP_LSH_BANDS;
+    const rows = NEAR_DUP_SIGNATURE_K / bands;
+    // bands MUST divide k evenly — otherwise `rows` is fractional and band slicing is
+    // misaligned. `rows * bands === k` alone does not catch this (128/7*7 round-trips to
+    // 128 in IEEE-754), so test integrality.
+    if (!Number.isInteger(rows) || rows < 1)
+        return [];
+    const eligible = collectEligible(candidates, minBodySize);
+    if (eligible.length < 2)
+        return [];
+    const edges = buildNearEdges(eligible, minSimilarity, bands, rows);
+    const components = clusterComponents(eligible.length, edges);
+    return emitComponentClusters(eligible, components, edges);
+}
+function collectEligible(candidates, minBodySize) {
+    const out = [];
+    for (const occ of candidates) {
+        if (!isEligibleKind(occ))
+            continue;
+        if (occ.bodySignature?.length !== NEAR_DUP_SIGNATURE_K)
+            continue;
+        if (occ.bodySize !== undefined && occ.bodySize < minBodySize)
+            continue;
+        out.push(occ);
+    }
+    return out;
+}
+function buildNearEdges(eligible, minSimilarity, bands, rows) {
+    const buckets = indexLshBuckets(eligible, bands, rows);
+    const edges = [];
+    const seenPairs = new Set();
+    for (const indices of buckets.values()) {
+        collectBucketEdges(indices, eligible, minSimilarity, seenPairs, edges);
+    }
+    return edges;
+}
+function indexLshBuckets(eligible, bands, rows) {
+    const buckets = new Map();
+    for (const [i, occ] of eligible.entries()) {
+        if (!occ.bodySignature)
+            continue;
+        const bandHashes = lshBandHashes(occ.bodySignature, bands, rows);
+        for (const [band, bandHash] of bandHashes.entries()) {
+            const key = `${String(band)}:${bandHash ?? ''}`;
+            const list = buckets.get(key) ?? [];
+            list.push(i);
+            buckets.set(key, list);
+        }
+    }
+    return buckets;
+}
+function collectBucketEdges(indices, eligible, minSimilarity, seenPairs, edges) {
+    if (indices.length < 2)
+        return;
+    for (let i = 0; i < indices.length; i++) {
+        for (let j = i + 1; j < indices.length; j++) {
+            const ai = indices[i];
+            const bi = indices[j];
+            if (ai === undefined || bi === undefined)
+                continue;
+            const key = pairKey(ai, bi);
+            if (seenPairs.has(key))
+                continue;
+            seenPairs.add(key);
+            const edge = tryNearEdge(ai, bi, eligible, minSimilarity);
+            if (edge)
+                edges.push(edge);
+        }
+    }
+}
+function tryNearEdge(ai, bi, eligible, minSimilarity) {
+    const a = eligible[ai];
+    const b = eligible[bi];
+    if (!a?.bodySignature || !b?.bodySignature)
+        return undefined;
+    if (a.bodyHash === b.bodyHash)
+        return undefined;
+    // Same-language gate. `language` is the caller-resolved equivalent of graph's
+    // `languageOfFile(filePath)`; undefined on either side skips the pair (identical to
+    // `languageOfFile(...) === undefined`).
+    const langA = a.language;
+    const langB = b.language;
+    if (langA === undefined || langB === undefined || langA !== langB)
+        return undefined;
+    const similarity = estimateJaccard(a.bodySignature, b.bodySignature);
+    if (similarity < minSimilarity)
+        return undefined;
+    return { a: ai, b: bi, similarity };
+}
+function clusterComponents(size, edges) {
+    const uf = new UnionFind(size);
+    for (const e of edges)
+        uf.union(e.a, e.b);
+    const byRoot = new Map();
+    for (let i = 0; i < size; i++) {
+        const root = uf.find(i);
+        const list = byRoot.get(root) ?? [];
+        list.push(i);
+        byRoot.set(root, list);
+    }
+    return [...byRoot.values()].filter((c) => c.length >= 2);
+}
+function emitComponentClusters(eligible, components, edges) {
+    const edgeByPair = buildEdgeSimilarityIndex(edges);
+    const clusters = [];
+    for (const component of components) {
+        const cluster = buildComponentCluster(eligible, component, edges, edgeByPair);
+        if (cluster)
+            clusters.push(cluster);
+    }
+    return clusters;
+}
+function buildEdgeSimilarityIndex(edges) {
+    const edgeByPair = new Map();
+    for (const e of edges) {
+        const key = pairKey(e.a, e.b);
+        const prev = edgeByPair.get(key);
+        if (prev === undefined || e.similarity > prev)
+            edgeByPair.set(key, e.similarity);
+    }
+    return edgeByPair;
+}
+function buildComponentCluster(eligible, component, edges, edgeByPair) {
+    const nearIndices = nearIndicesInComponent(component, edges);
+    if (nearIndices.size < 2)
+        return undefined;
+    if (component.length > MAX_CLUSTER_SIZE)
+        return undefined;
+    const members = component.map((i) => eligible[i]).filter((o) => !!o);
+    const anchor = lowestByLocation(members);
+    const nearMembers = [...nearIndices]
+        .map((i) => eligible[i]?.qualifiedName)
+        .filter((n) => n !== undefined)
+        .sort();
+    const exactMembers = exactMembersInComponent(members);
+    const maxSim = maxSimilarityAmong(nearIndices, edgeByPair);
+    return {
+        anchor,
+        nearMembers,
+        exactMembers,
+        estimatedSimilarity: maxSim,
+        clusterSize: component.length,
+    };
+}
+function nearIndicesInComponent(component, edges) {
+    const nearIndices = new Set();
+    const componentSet = new Set(component);
+    for (const e of edges) {
+        if (!componentSet.has(e.a) && !componentSet.has(e.b))
+            continue;
+        nearIndices.add(e.a);
+        nearIndices.add(e.b);
+    }
+    return nearIndices;
+}
+function exactMembersInComponent(members) {
+    const hashCounts = new Map();
+    for (const m of members)
+        hashCounts.set(m.bodyHash, (hashCounts.get(m.bodyHash) ?? 0) + 1);
+    return members
+        .filter((m) => (hashCounts.get(m.bodyHash) ?? 0) > 1)
+        .map((m) => m.qualifiedName)
+        .sort();
+}
+function maxSimilarityAmong(nearIndices, edgeByPair) {
+    let maxSim = 0;
+    for (const i of nearIndices) {
+        for (const j of nearIndices) {
+            if (i >= j)
+                continue;
+            const sim = edgeByPair.get(pairKey(i, j));
+            if (sim !== undefined && sim > maxSim)
+                maxSim = sim;
+        }
+    }
+    return maxSim;
+}
+function lowestByLocation(occs) {
+    return occs.reduce((lo, c) => {
+        if (c.filePath < lo.filePath)
+            return c;
+        if (c.filePath > lo.filePath)
+            return lo;
+        if (c.line < lo.line)
+            return c;
+        if (c.line > lo.line)
+            return lo;
+        return c.column < lo.column ? c : lo;
+    });
+}
+function pairKey(a, b) {
+    return a < b ? `${String(a)}:${String(b)}` : `${String(b)}:${String(a)}`;
+}
+class UnionFind {
+    parent;
+    rank;
+    constructor(size) {
+        this.parent = Array.from({ length: size }, (_, i) => i);
+        this.rank = Array.from({ length: size }, () => 0);
+    }
+    find(x) {
+        const p = this.parent[x];
+        if (p === undefined || p === x)
+            return x;
+        const root = this.find(p);
+        this.parent[x] = root;
+        return root;
+    }
+    union(a, b) {
+        const ra = this.find(a);
+        const rb = this.find(b);
+        if (ra === rb)
+            return;
+        const rankA = this.rank[ra] ?? 0;
+        const rankB = this.rank[rb] ?? 0;
+        if (rankA < rankB) {
+            this.parent[ra] = rb;
+        }
+        else if (rankA > rankB) {
+            this.parent[rb] = ra;
+        }
+        else {
+            this.parent[rb] = ra;
+            this.rank[ra] = rankA + 1;
+        }
+    }
+}
+//# sourceMappingURL=find-near-duplicates.js.map

package/dist/find-near-duplicates.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"find-near-duplicates.js","sourceRoot":"","sources":["../src/find-near-duplicates.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAAE,cAAc,EAAE,MAAM,4BAA4B,CAAC;AAC5D,OAAO,EACL,kBAAkB,EAClB,oBAAoB,EACpB,eAAe,EACf,aAAa,GACd,MAAM,+BAA+B,CAAC;AAIvC,MAAM,sBAAsB,GAAG,IAAI,CAAC;AACpC,MAAM,qBAAqB,GAAG,GAAG,CAAC;AAClC,MAAM,gBAAgB,GAAG,EAAE,CAAC;AAE5B;;;GAGG;AACH,MAAM,UAAU,kBAAkB,CAChC,UAAqC,EACrC,OAAoB,EAAE;IAEtB,MAAM,aAAa,GAAG,IAAI,CAAC,aAAa,IAAI,sBAAsB,CAAC;IACnE,MAAM,WAAW,GAAG,IAAI,CAAC,WAAW,IAAI,qBAAqB,CAAC;IAC9D,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,IAAI,kBAAkB,CAAC;IAClD,MAAM,IAAI,GAAG,oBAAoB,GAAG,KAAK,CAAC;IAC1C,kFAAkF;IAClF,qFAAqF;IACrF,yCAAyC;IACzC,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,IAAI,GAAG,CAAC;QAAE,OAAO,EAAE,CAAC;IAEnD,MAAM,QAAQ,GAAG,eAAe,CAAC,UAAU,EAAE,WAAW,CAAC,CAAC;IAC1D,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,EAAE,CAAC;IAEnC,MAAM,KAAK,GAAG,cAAc,CAAC,QAAQ,EAAE,aAAa,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;IACnE,MAAM,UAAU,GAAG,iBAAiB,CAAC,QAAQ,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC;IAC7D,OAAO,qBAAqB,CAAC,QAAQ,EAAE,UAAU,EAAE,KAAK,CAAC,CAAC;AAC5D,CAAC;AAED,SAAS,eAAe,CACtB,UAAqC,EACrC,WAAmB;IAEnB,MAAM,GAAG,GAAqB,EAAE,CAAC;IACjC,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAC7B,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC;YAAE,SAAS;QACnC,IAAI,GAAG,CAAC,aAAa,EAAE,MAAM,KAAK,oBAAoB;YAAE,SAAS;QACjE,IAAI,GAAG,CAAC,QAAQ,KAAK,SAAS,IAAI,GAAG,CAAC,QAAQ,GAAG,WAAW;YAAE,SAAS;QACvE,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAChB,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAQD,SAAS,cAAc,CACrB,QAAmC,EACnC,aAAqB,EACrB,KAAa,EACb,IAAY;IAEZ,MAAM,OAAO,GAAG,eAAe,CAAC,QAAQ,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;IACvD,MAAM,KAAK,GAAe,EAAE,CAAC;IAC7B,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;IACpC,KAAK,MAAM,OAAO,IAAI,OAAO,CAAC,MAAM,EAAE,EAAE,CAAC;QACvC,kBAAkB,CAAC,OAAO,EAAE,QAAQ,EAAE,aAAa,EAAE,SAAS,EAAE,KAAK,CAAC,CAAC;IACzE,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,eAAe,CACtB,QAAmC,EACnC,KAAa,EACb,IAAY;IAEZ,MAAM,OAAO,GAAG,IAAI,GAAG,EAAoB,CAAC;IAC5C,KAAK,MAAM,CAAC,CAAC,EAAE,GAAG,CAAC,IAAI,QAAQ,CAAC,OAAO,EAAE,EAAE,CAAC;QAC1C,IAAI,CAAC,GAAG,CAAC,aAAa;YAAE,SAAS;QACjC,MAAM,UAAU,GAAG,aAAa,CAAC,GAAG,CAAC,aAAa,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;QACjE,KAAK,MAAM,CAAC,IAAI,EAAE,QAAQ,CAAC,IAAI,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC;YACpD,MAAM,GAAG,GAAG,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,QAAQ,IAAI,EAAE,EAAE,CAAC;YAChD,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC;YACpC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;YACb,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;QACzB,CAAC;IACH,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,SAAS,kBAAkB,CACzB,OAA0B,EAC1B,QAAmC,EACnC,aAAqB,EACrB,SAAsB,EACtB,KAAiB;IAEjB,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO;IAC/B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACxC,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC5C,MAAM,EAAE,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;YACtB,MAAM,EAAE,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;YACtB,IAAI,EAAE,KAAK,SAAS,IAAI,EAAE,KAAK,SAAS;gBAAE,SAAS;YACnD,MAAM,GAAG,GAAG,OAAO,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC;YAC5B,IAAI,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC;gBAAE,SAAS;YACjC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YACnB,MAAM,IAAI,GAAG,WAAW,CAAC,EAAE,EAAE,EAAE,EAAE,QAAQ,EAAE,aAAa,CAAC,CAAC;YAC1D,IAAI,IAAI;gBAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAC7B,CAAC;IACH,CAAC;AACH,CAAC;AAED,SAAS,WAAW,CAClB,EAAU,EACV,EAAU,EACV,QAAmC,EACnC,aAAqB;IAErB,MAAM,CAAC,GAAG,QAAQ,CAAC,EAAE,CAAC,CAAC;IACvB,MAAM,CAAC,GAAG,QAAQ,CAAC,EAAE,CAAC,CAAC;IACvB,IAAI,CAAC,CAAC,EAAE,aAAa,IAAI,CAAC,CAAC,EAAE,aAAa;QAAE,OAAO,SAAS,CAAC;IAC7D,IAAI,CAAC,CAAC,QAAQ,KAAK,CAAC,CAAC,QAAQ;QAAE,OAAO,SAAS,CAAC;IAEhD,8EAA8E;IAC9E,oFAAoF;IACpF,wCAAwC;IACxC,MAAM,KAAK,GAAG,CAAC,CAAC,QAAQ,CAAC;IACzB,MAAM,KAAK,GAAG,CAAC,CAAC,QAAQ,CAAC;IACzB,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,KAAK;QAAE,OAAO,SAAS,CAAC;IAEpF,MAAM,UAAU,GAAG,eAAe,CAAC,CAAC,CAAC,aAAa,EAAE,CAAC,CAAC,aAAa,CAAC,CAAC;IACrE,IAAI,UAAU,GAAG,aAAa;QAAE,OAAO,SAAS,CAAC;IACjD,OAAO,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,UAAU,EAAE,CAAC;AACtC,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAY,EAAE,KAA0B;IACjE,MAAM,EAAE,GAAG,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC;IAC/B,KAAK,MAAM,CAAC,IAAI,KAAK;QAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAE1C,MAAM,MAAM,GAAG,IAAI,GAAG,EAAoB,CAAC;IAC3C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9B,MAAM,IAAI,GAAG,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACxB,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;QACpC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACb,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;IACzB,CAAC;IACD,OAAO,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,CAAC;AAC3D,CAAC;AAED,SAAS,qBAAqB,CAC5B,QAAmC,EACnC,UAA+B,EAC/B,KAA0B;IAE1B,MAAM,UAAU,GAAG,wBAAwB,CAAC,KAAK,CAAC,CAAC;IACnD,MAAM,QAAQ,GAA2B,EAAE,CAAC;IAC5C,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;QACnC,MAAM,OAAO,GAAG,qBAAqB,CAAC,QAAQ,EAAE,SAAS,EAAE,KAAK,EAAE,UAAU,CAAC,CAAC;QAC9E,IAAI,OAAO;YAAE,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;IACtC,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,SAAS,wBAAwB,CAAC,KAA0B;IAC1D,MAAM,UAAU,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC7C,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;QACtB,MAAM,GAAG,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;QAC9B,MAAM,IAAI,GAAG,UAAU,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACjC,IAAI,IAAI,KAAK,SAAS,IAAI,CAAC,CAAC,UAAU,GAAG,IAAI;YAAE,UAAU,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,UAAU,CAAC,CAAC;IACnF,CAAC;IACD,OAAO,UAAU,CAAC;AACpB,CAAC;AAED,SAAS,qBAAqB,CAC5B,QAAmC,EACnC,SAA4B,EAC5B,KAA0B,EAC1B,UAAuC;IAEvC,MAAM,WAAW,GAAG,sBAAsB,CAAC,SAAS,EAAE,KAAK,CAAC,CAAC;IAC7D,IAAI,WAAW,CAAC,IAAI,GAAG,CAAC;QAAE,OAAO,SAAS,CAAC;IAC3C,IAAI,SAAS,CAAC,MAAM,GAAG,gBAAgB;QAAE,OAAO,SAAS,CAAC;IAE1D,MAAM,OAAO,GAAG,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAuB,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1F,MAAM,MAAM,GAAG,gBAAgB,CAAC,OAAO,CAAC,CAAC;IACzC,MAAM,WAAW,GAAG,CAAC,GAAG,WAAW,CAAC;SACjC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,aAAa,CAAC;SACtC,MAAM,CAAC,CAAC,CAAC,EAAe,EAAE,CAAC,CAAC,KAAK,SAAS,CAAC;SAC3C,IAAI,EAAE,CAAC;IACV,MAAM,YAAY,GAAG,uBAAuB,CAAC,OAAO,CAAC,CAAC;IACtD,MAAM,MAAM,GAAG,kBAAkB,CAAC,WAAW,EAAE,UAAU,CAAC,CAAC;IAE3D,OAAO;QACL,MAAM;QACN,WAAW;QACX,YAAY;QACZ,mBAAmB,EAAE,MAAM;QAC3B,WAAW,EAAE,SAAS,CAAC,MAAM;KAC9B,CAAC;AACJ,CAAC;AAED,SAAS,sBAAsB,CAC7B,SAA4B,EAC5B,KAA0B;IAE1B,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;IACtC,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,SAAS,CAAC,CAAC;IACxC,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;QACtB,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;YAAE,SAAS;QAC/D,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACrB,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACvB,CAAC;IACD,OAAO,WAAW,CAAC;AACrB,CAAC;AAED,SAAS,uBAAuB,CAAC,OAAkC;IACjE,MAAM,UAAU,GAAG,IAAI,GAAG,EAAkB,CAAC;IAC7C,KAAK,MAAM,CAAC,IAAI,OAAO;QAAE,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,EAAE,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAC3F,OAAO,OAAO;SACX,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;SACpD,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,aAAa,CAAC;SAC3B,IAAI,EAAE,CAAC;AACZ,CAAC;AAED,SAAS,kBAAkB,CACzB,WAAgC,EAChC,UAAuC;IAEvC,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,KAAK,MAAM,CAAC,IAAI,WAAW,EAAE,CAAC;QAC5B,KAAK,MAAM,CAAC,IAAI,WAAW,EAAE,CAAC;YAC5B,IAAI,CAAC,IAAI,CAAC;gBAAE,SAAS;YACrB,MAAM,GAAG,GAAG,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;YAC1C,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,GAAG,MAAM;gBAAE,MAAM,GAAG,GAAG,CAAC;QACtD,CAAC;IACH,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,gBAAgB,CAAC,IAA+B;IACvD,OAAO,IAAI,CAAC,MAAM,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,EAAE;QAC3B,IAAI,CAAC,CAAC,QAAQ,GAAG,EAAE,CAAC,QAAQ;YAAE,OAAO,CAAC,CAAC;QACvC,IAAI,CAAC,CAAC,QAAQ,GAAG,EAAE,CAAC,QAAQ;YAAE,OAAO,EAAE,CAAC;QACxC,IAAI,CAAC,CAAC,IAAI,GAAG,EAAE,CAAC,IAAI;YAAE,OAAO,CAAC,CAAC;QAC/B,IAAI,CAAC,CAAC,IAAI,GAAG,EAAE,CAAC,IAAI;YAAE,OAAO,EAAE,CAAC;QAChC,OAAO,CAAC,CAAC,MAAM,GAAG,EAAE,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACvC,CAAC,CAAC,CAAC;AACL,CAAC;AAED,SAAS,OAAO,CAAC,CAAS,EAAE,CAAS;IACnC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,MAAM,CAAC,CAAC,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC;AAC3E,CAAC;AAED,MAAM,SAAS;IACI,MAAM,CAAW;IACjB,IAAI,CAAW;IAEhC,YAAY,IAAY;QACtB,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;QACxD,IAAI,CAAC,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;IACpD,CAAC;IAED,IAAI,CAAC,CAAS;QACZ,MAAM,CAAC,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;QACzB,IAAI,CAAC,KAAK,SAAS,IAAI,CAAC,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC;QACzC,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QAC1B,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,IAAI,CAAC;QACtB,OAAO,IAAI,CAAC;IACd,CAAC;IAED,KAAK,CAAC,CAAS,EAAE,CAAS;QACxB,MAAM,EAAE,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACxB,MAAM,EAAE,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;QACxB,IAAI,EAAE,KAAK,EAAE;YAAE,OAAO;QACtB,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC;QACjC,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,CAAC,CAAC;QACjC,IAAI,KAAK,GAAG,KAAK,EAAE,CAAC;YAClB,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,GAAG,EAAE,CAAC;QACvB,CAAC;aAAM,IAAI,KAAK,GAAG,KAAK,EAAE,CAAC;YACzB,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,GAAG,EAAE,CAAC;QACvB,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC,GAAG,EAAE,CAAC;YACrB,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC;QAC5B,CAAC;IACH,CAAC;CACF"}

package/dist/index.d.ts

@@ -0,0 +1,16 @@
+/**
+ * @opensip-cli/clone-detection — shared function-body clone-detection substrate.
+ *
+ * A pure, `node:crypto`-only leaf package (layer 2) that single-sources the body-hash
+ * + MinHash primitives, the tool-neutral `CloneCandidate` shape, and the duplicate /
+ * near-duplicate detection algorithms + curation policy. Both the graph tool and the
+ * yagni tool depend on it (neither on the other), so there is exactly one
+ * implementation and they cannot diverge (ADR-0064).
+ */
+export { normalizeWhitespace, hashBody, type BodyDigest } from './body-digest.js';
+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 './near-duplicate-signature.js';
+export type { FunctionKind, CloneCandidate, DupOpts, NearDupOpts, DuplicateGroup, CrossPackageAggregate, DuplicateFindings, NearDuplicateCluster, } from './types.js';
+export { findDuplicateBodies, isEligibleKind } from './find-duplicate-bodies.js';
+export { findNearDuplicates } from './find-near-duplicates.js';
+export { isTestFilePath } from './test-file.js';
+//# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAIH,OAAO,EAAE,mBAAmB,EAAE,QAAQ,EAAE,KAAK,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAGlF,OAAO,EACL,oBAAoB,EACpB,kBAAkB,EAClB,iBAAiB,EACjB,0BAA0B,EAC1B,OAAO,EACP,aAAa,EACb,eAAe,EACf,aAAa,EACb,mBAAmB,EACnB,KAAK,uBAAuB,GAC7B,MAAM,+BAA+B,CAAC;AAIvC,YAAY,EACV,YAAY,EACZ,cAAc,EACd,OAAO,EACP,WAAW,EACX,cAAc,EACd,qBAAqB,EACrB,iBAAiB,EACjB,oBAAoB,GACrB,MAAM,YAAY,CAAC;AACpB,OAAO,EAAE,mBAAmB,EAAE,cAAc,EAAE,MAAM,4BAA4B,CAAC;AACjF,OAAO,EAAE,kBAAkB,EAAE,MAAM,2BAA2B,CAAC;AAI/D,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC"}

package/dist/index.js

@@ -0,0 +1,20 @@
+/**
+ * @opensip-cli/clone-detection — shared function-body clone-detection substrate.
+ *
+ * A pure, `node:crypto`-only leaf package (layer 2) that single-sources the body-hash
+ * + MinHash primitives, the tool-neutral `CloneCandidate` shape, and the duplicate /
+ * near-duplicate detection algorithms + curation policy. Both the graph tool and the
+ * yagni tool depend on it (neither on the other), so there is exactly one
+ * implementation and they cannot diverge (ADR-0064).
+ */
+// Body-hash primitives (relocated verbatim from graph — bodyHash is the catalog/cache
+// /equivalence-guardrail identity; the values must never change).
+export { normalizeWhitespace, hashBody } from './body-digest.js';
+// MinHash / LSH near-duplicate primitives + algorithm constants.
+export { NEAR_DUP_SIGNATURE_K, NEAR_DUP_LSH_BANDS, NEAR_DUP_LSH_ROWS, NEAR_DUP_SIGNATURE_VERSION, shingle, bodySignature, estimateJaccard, lshBandHashes, digestCanonicalBody, } from './near-duplicate-signature.js';
+export { findDuplicateBodies, isEligibleKind } from './find-duplicate-bodies.js';
+export { findNearDuplicates } from './find-near-duplicates.js';
+// Canonical TS/JS test-file predicate (single-sourced so every CloneCandidate producer
+// stamps `inTestFile` identically — D1).
+export { isTestFilePath } from './test-file.js';
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,sFAAsF;AACtF,kEAAkE;AAClE,OAAO,EAAE,mBAAmB,EAAE,QAAQ,EAAmB,MAAM,kBAAkB,CAAC;AAElF,iEAAiE;AACjE,OAAO,EACL,oBAAoB,EACpB,kBAAkB,EAClB,iBAAiB,EACjB,0BAA0B,EAC1B,OAAO,EACP,aAAa,EACb,eAAe,EACf,aAAa,EACb,mBAAmB,GAEpB,MAAM,+BAA+B,CAAC;AAcvC,OAAO,EAAE,mBAAmB,EAAE,cAAc,EAAE,MAAM,4BAA4B,CAAC;AACjF,OAAO,EAAE,kBAAkB,EAAE,MAAM,2BAA2B,CAAC;AAE/D,uFAAuF;AACvF,yCAAyC;AACzC,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC"}

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

@@ -0,0 +1,55 @@
+/**
+ * Near-duplicate body signatures — MinHash + LSH primitives.
+ *
+ * Single source of truth for signature constants and the digest tail used by
+ * both tree-sitter adapters and the TypeScript inventory path. Signatures are
+ * derived from the same canonical body string that feeds `bodyHash`.
+ */
+import { type BodyDigest } from './body-digest.js';
+/** Number of MinHash values per body signature. */
+export declare const NEAR_DUP_SIGNATURE_K = 128;
+/** Default LSH band count — co-tuned with {@link NEAR_DUP_LSH_ROWS}. */
+export declare const NEAR_DUP_LSH_BANDS = 8;
+/** Rows per band; `bands × rows === k`. Knee ≈ 0.878 at threshold 0.85. */
+export declare const NEAR_DUP_LSH_ROWS = 16;
+/**
+ * Signature ALGORITHM version. Bump on ANY change to shingling, hashing, or the
+ * permutation scheme that alters signature VALUES. Feeds the `sig=` cache-key
+ * segment so catalogs built with an older algorithm invalidate — mixing old- and
+ * new-algorithm signatures across an incremental build (some occurrences cached,
+ * some re-walked) would corrupt every cross-occurrence Jaccard estimate.
+ *
+ * v1 = k independent SHA-256 hashes per (shingle, seed). v2 = one SHA-256 base
+ * hash per shingle + k cheap 32-bit mixers (~k× fewer hashes, identical MinHash
+ * semantics).
+ */
+export declare const NEAR_DUP_SIGNATURE_VERSION = 2;
+/** Digest including hash, size, and optional near-duplicate MinHash signature. */
+export type BodyDigestWithSignature = BodyDigest;
+/**
+ * Character k-grams from canonical body text. Bodies shorter than `gramSize`
+ * yield a single shingle of the whole string when non-empty.
+ */
+export declare function shingle(canonical: string, gramSize?: number): ReadonlySet<string>;
+/**
+ * Deterministic MinHash signature over char shingles. Stable across runs and
+ * machines (SHA-256 base hash + fixed permutation seeds).
+ *
+ * Each shingle is hashed ONCE with SHA-256 (the only expensive step); the k
+ * MinHash values are then derived by mixing that base hash with k fixed 32-bit
+ * seeds. This is ~k× fewer SHA-256 computations than hashing every
+ * (shingle, seed) pair while preserving MinHash semantics (each of the k mixers
+ * is an independent hash of the shingle universe, so the per-position min
+ * estimates Jaccard exactly as before). Algorithm {@link NEAR_DUP_SIGNATURE_VERSION}.
+ */
+export declare function bodySignature(canonical: string, k?: number): readonly number[];
+/** MinHash Jaccard estimate: equal-position fraction. */
+export declare function estimateJaccard(a: readonly number[], b: readonly number[]): number;
+/** LSH band hashes for candidate-pair generation. */
+export declare function lshBandHashes(signature: readonly number[], bands: number, rows: number): readonly string[];
+/**
+ * Hash + size + signature for an already-normalized canonical body string.
+ * Skips `signature` when the canonical text yields no shingles.
+ */
+export declare function digestCanonicalBody(canonical: string): BodyDigestWithSignature;
+//# sourceMappingURL=near-duplicate-signature.d.ts.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"near-duplicate-signature.d.ts","sourceRoot":"","sources":["../src/near-duplicate-signature.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,OAAO,EAAY,KAAK,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAE7D,mDAAmD;AACnD,eAAO,MAAM,oBAAoB,MAAM,CAAC;AAExC,wEAAwE;AACxE,eAAO,MAAM,kBAAkB,IAAI,CAAC;AAEpC,2EAA2E;AAC3E,eAAO,MAAM,iBAAiB,KAAK,CAAC;AAEpC;;;;;;;;;;GAUG;AACH,eAAO,MAAM,0BAA0B,IAAI,CAAC;AAE5C,kFAAkF;AAClF,MAAM,MAAM,uBAAuB,GAAG,UAAU,CAAC;AAEjD;;;GAGG;AACH,wBAAgB,OAAO,CAAC,SAAS,EAAE,MAAM,EAAE,QAAQ,SAAI,GAAG,WAAW,CAAC,MAAM,CAAC,CAW5E;AAED;;;;;;;;;;GAUG;AACH,wBAAgB,aAAa,CAAC,SAAS,EAAE,MAAM,EAAE,CAAC,SAAuB,GAAG,SAAS,MAAM,EAAE,CAmB5F;AAED,yDAAyD;AACzD,wBAAgB,eAAe,CAAC,CAAC,EAAE,SAAS,MAAM,EAAE,EAAE,CAAC,EAAE,SAAS,MAAM,EAAE,GAAG,MAAM,CAOlF;AAED,qDAAqD;AACrD,wBAAgB,aAAa,CAC3B,SAAS,EAAE,SAAS,MAAM,EAAE,EAC5B,KAAK,EAAE,MAAM,EACb,IAAI,EAAE,MAAM,GACX,SAAS,MAAM,EAAE,CAOnB;AAED;;;GAGG;AACH,wBAAgB,mBAAmB,CAAC,SAAS,EAAE,MAAM,GAAG,uBAAuB,CAO9E"}

package/dist/near-duplicate-signature.js

@@ -0,0 +1,135 @@
+/**
+ * Near-duplicate body signatures — MinHash + LSH primitives.
+ *
+ * Single source of truth for signature constants and the digest tail used by
+ * both tree-sitter adapters and the TypeScript inventory path. Signatures are
+ * derived from the same canonical body string that feeds `bodyHash`.
+ */
+import { createHash } from 'node:crypto';
+import { hashBody } from './body-digest.js';
+/** Number of MinHash values per body signature. */
+export const NEAR_DUP_SIGNATURE_K = 128;
+/** Default LSH band count — co-tuned with {@link NEAR_DUP_LSH_ROWS}. */
+export const NEAR_DUP_LSH_BANDS = 8;
+/** Rows per band; `bands × rows === k`. Knee ≈ 0.878 at threshold 0.85. */
+export const NEAR_DUP_LSH_ROWS = 16;
+/**
+ * Signature ALGORITHM version. Bump on ANY change to shingling, hashing, or the
+ * permutation scheme that alters signature VALUES. Feeds the `sig=` cache-key
+ * segment so catalogs built with an older algorithm invalidate — mixing old- and
+ * new-algorithm signatures across an incremental build (some occurrences cached,
+ * some re-walked) would corrupt every cross-occurrence Jaccard estimate.
+ *
+ * v1 = k independent SHA-256 hashes per (shingle, seed). v2 = one SHA-256 base
+ * hash per shingle + k cheap 32-bit mixers (~k× fewer hashes, identical MinHash
+ * semantics).
+ */
+export const NEAR_DUP_SIGNATURE_VERSION = 2;
+/**
+ * Character k-grams from canonical body text. Bodies shorter than `gramSize`
+ * yield a single shingle of the whole string when non-empty.
+ */
+export function shingle(canonical, gramSize = 5) {
+    const set = new Set();
+    if (canonical.length === 0)
+        return set;
+    if (canonical.length < gramSize) {
+        set.add(canonical);
+        return set;
+    }
+    for (let i = 0; i <= canonical.length - gramSize; i++) {
+        set.add(canonical.slice(i, i + gramSize));
+    }
+    return set;
+}
+/**
+ * Deterministic MinHash signature over char shingles. Stable across runs and
+ * machines (SHA-256 base hash + fixed permutation seeds).
+ *
+ * Each shingle is hashed ONCE with SHA-256 (the only expensive step); the k
+ * MinHash values are then derived by mixing that base hash with k fixed 32-bit
+ * seeds. This is ~k× fewer SHA-256 computations than hashing every
+ * (shingle, seed) pair while preserving MinHash semantics (each of the k mixers
+ * is an independent hash of the shingle universe, so the per-position min
+ * estimates Jaccard exactly as before). Algorithm {@link NEAR_DUP_SIGNATURE_VERSION}.
+ */
+export function bodySignature(canonical, k = NEAR_DUP_SIGNATURE_K) {
+    const shingles = shingle(canonical);
+    if (shingles.size === 0)
+        return [];
+    const seeds = k === NEAR_DUP_SIGNATURE_K ? PERM_SEEDS : derivePermSeeds(k);
+    // Hash each shingle exactly once — the per-(shingle, seed) k-fold SHA-256 of
+    // the v1 algorithm is the cost this removes.
+    const bases = [];
+    for (const gram of shingles)
+        bases.push(baseHash(gram));
+    const sig = [];
+    for (let i = 0; i < k; i++) {
+        const seed = seeds[i] ?? 0;
+        let min = 0xff_ff_ff_ff;
+        for (const base of bases) {
+            const v = mix32(base, seed);
+            if (v < min)
+                min = v;
+        }
+        sig.push(min);
+    }
+    return sig;
+}
+/** MinHash Jaccard estimate: equal-position fraction. */
+export function estimateJaccard(a, b) {
+    if (a.length === 0 || b.length === 0 || a.length !== b.length)
+        return 0;
+    let matches = 0;
+    for (const [i, element] of a.entries()) {
+        if (element === b[i])
+            matches++;
+    }
+    return matches / a.length;
+}
+/** LSH band hashes for candidate-pair generation. */
+export function lshBandHashes(signature, bands, rows) {
+    const out = [];
+    for (let band = 0; band < bands; band++) {
+        const start = band * rows;
+        out.push(signature.slice(start, start + rows).join(','));
+    }
+    return out;
+}
+/**
+ * Hash + size + signature for an already-normalized canonical body string.
+ * Skips `signature` when the canonical text yields no shingles.
+ */
+export function digestCanonicalBody(canonical) {
+    const digest = hashBody(canonical);
+    const signature = bodySignature(canonical);
+    return {
+        ...digest,
+        signature: signature.length > 0 ? signature : undefined,
+    };
+}
+/** SHA-256 → 32-bit base hash of one shingle (computed once per shingle). */
+function baseHash(gram) {
+    return createHash('sha256').update(gram, 'utf8').digest().readUInt32LE(0);
+}
+/**
+ * Cheap 32-bit avalanche of a base hash under one permutation seed (Mueller's
+ * integer finalizer). Division-free, no precision loss — `Math.imul` keeps the
+ * multiply in 32-bit. Distinct seeds yield independent hashes of the shingle.
+ */
+function mix32(h, seed) {
+    const MIX_CONST = 0x4_5d_9f_3b;
+    let x = (h ^ seed) >>> 0;
+    x = Math.imul((x >>> 16) ^ x, MIX_CONST) >>> 0;
+    x = Math.imul((x >>> 16) ^ x, MIX_CONST) >>> 0;
+    return ((x >>> 16) ^ x) >>> 0;
+}
+/** k fixed permutation seeds, derived deterministically (SHA-256 of the index). */
+function derivePermSeeds(k) {
+    return Array.from({ length: k }, (_, i) => createHash('sha256')
+        .update(`opensip-minhash-perm:${String(i)}`, 'utf8')
+        .digest()
+        .readUInt32LE(0));
+}
+const PERM_SEEDS = derivePermSeeds(NEAR_DUP_SIGNATURE_K);
+//# sourceMappingURL=near-duplicate-signature.js.map

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

@@ -0,0 +1 @@
+{"version":3,"file":"near-duplicate-signature.js","sourceRoot":"","sources":["../src/near-duplicate-signature.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAEzC,OAAO,EAAE,QAAQ,EAAmB,MAAM,kBAAkB,CAAC;AAE7D,mDAAmD;AACnD,MAAM,CAAC,MAAM,oBAAoB,GAAG,GAAG,CAAC;AAExC,wEAAwE;AACxE,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,CAAC;AAEpC,2EAA2E;AAC3E,MAAM,CAAC,MAAM,iBAAiB,GAAG,EAAE,CAAC;AAEpC;;;;;;;;;;GAUG;AACH,MAAM,CAAC,MAAM,0BAA0B,GAAG,CAAC,CAAC;AAK5C;;;GAGG;AACH,MAAM,UAAU,OAAO,CAAC,SAAiB,EAAE,QAAQ,GAAG,CAAC;IACrD,MAAM,GAAG,GAAG,IAAI,GAAG,EAAU,CAAC;IAC9B,IAAI,SAAS,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,GAAG,CAAC;IACvC,IAAI,SAAS,CAAC,MAAM,GAAG,QAAQ,EAAE,CAAC;QAChC,GAAG,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QACnB,OAAO,GAAG,CAAC;IACb,CAAC;IACD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,SAAS,CAAC,MAAM,GAAG,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;QACtD,GAAG,CAAC,GAAG,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC;IAC5C,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;;;;;;;;;GAUG;AACH,MAAM,UAAU,aAAa,CAAC,SAAiB,EAAE,CAAC,GAAG,oBAAoB;IACvE,MAAM,QAAQ,GAAG,OAAO,CAAC,SAAS,CAAC,CAAC;IACpC,IAAI,QAAQ,CAAC,IAAI,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IACnC,MAAM,KAAK,GAAG,CAAC,KAAK,oBAAoB,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,CAAC;IAC3E,6EAA6E;IAC7E,6CAA6C;IAC7C,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,IAAI,IAAI,QAAQ;QAAE,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC;IACxD,MAAM,GAAG,GAAa,EAAE,CAAC;IACzB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3B,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAC3B,IAAI,GAAG,GAAG,aAAa,CAAC;QACxB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,MAAM,CAAC,GAAG,KAAK,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;YAC5B,IAAI,CAAC,GAAG,GAAG;gBAAE,GAAG,GAAG,CAAC,CAAC;QACvB,CAAC;QACD,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAChB,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,yDAAyD;AACzD,MAAM,UAAU,eAAe,CAAC,CAAoB,EAAE,CAAoB;IACxE,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,MAAM;QAAE,OAAO,CAAC,CAAC;IACxE,IAAI,OAAO,GAAG,CAAC,CAAC;IAChB,KAAK,MAAM,CAAC,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC;QACvC,IAAI,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC;YAAE,OAAO,EAAE,CAAC;IAClC,CAAC;IACD,OAAO,OAAO,GAAG,CAAC,CAAC,MAAM,CAAC;AAC5B,CAAC;AAED,qDAAqD;AACrD,MAAM,UAAU,aAAa,CAC3B,SAA4B,EAC5B,KAAa,EACb,IAAY;IAEZ,MAAM,GAAG,GAAa,EAAE,CAAC;IACzB,KAAK,IAAI,IAAI,GAAG,CAAC,EAAE,IAAI,GAAG,KAAK,EAAE,IAAI,EAAE,EAAE,CAAC;QACxC,MAAM,KAAK,GAAG,IAAI,GAAG,IAAI,CAAC;QAC1B,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,GAAG,IAAI,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IAC3D,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,mBAAmB,CAAC,SAAiB;IACnD,MAAM,MAAM,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAC;IACnC,MAAM,SAAS,GAAG,aAAa,CAAC,SAAS,CAAC,CAAC;IAC3C,OAAO;QACL,GAAG,MAAM;QACT,SAAS,EAAE,SAAS,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS;KACxD,CAAC;AACJ,CAAC;AAED,6EAA6E;AAC7E,SAAS,QAAQ,CAAC,IAAY;IAC5B,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC,MAAM,EAAE,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;AAC5E,CAAC;AAED;;;;GAIG;AACH,SAAS,KAAK,CAAC,CAAS,EAAE,IAAY;IACpC,MAAM,SAAS,GAAG,YAAY,CAAC;IAC/B,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC;IACzB,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC,GAAG,CAAC,EAAE,SAAS,CAAC,KAAK,CAAC,CAAC;IAC/C,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC,GAAG,CAAC,EAAE,SAAS,CAAC,KAAK,CAAC,CAAC;IAC/C,OAAO,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;AAChC,CAAC;AAED,mFAAmF;AACnF,SAAS,eAAe,CAAC,CAAS;IAChC,OAAO,KAAK,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CACxC,UAAU,CAAC,QAAQ,CAAC;SACjB,MAAM,CAAC,wBAAwB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,MAAM,CAAC;SACnD,MAAM,EAAE;SACR,YAAY,CAAC,CAAC,CAAC,CACnB,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,GAAsB,eAAe,CAAC,oBAAoB,CAAC,CAAC"}

package/dist/test-file.d.ts

@@ -0,0 +1,7 @@
+/**
+ * Returns true if `filePathProjectRel` is a TypeScript / JS test file (or test fixture)
+ * by path convention. Path is project-relative with `/` separators (the same shape
+ * `CloneCandidate.filePath` carries).
+ */
+export declare function isTestFilePath(filePathProjectRel: string): boolean;
+//# sourceMappingURL=test-file.d.ts.map

package/dist/test-file.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"test-file.d.ts","sourceRoot":"","sources":["../src/test-file.ts"],"names":[],"mappings":"AAgBA;;;;GAIG;AACH,wBAAgB,cAAc,CAAC,kBAAkB,EAAE,MAAM,GAAG,OAAO,CAMlE"}

package/dist/test-file.js

@@ -0,0 +1,26 @@
+/**
+ * Canonical TS/JS test-file predicate — single-sourced so every producer of a
+ * `CloneCandidate.inTestFile` flag (graph's TS walk, yagni's TS inventory) classifies
+ * test files IDENTICALLY. A divergence here is exactly the filter-divergence class
+ * ADR-0064 prevents (it would split the cross-tool parity test on `inTestFile`).
+ *
+ * Anchored patterns instead of one alternation; avoids catastrophic backtracking on
+ * pathological inputs. `__fixtures__/` counts as test scaffolding by convention.
+ *
+ * Relocated verbatim from `graph-typescript/src/test-file.ts` (the regexes are unchanged
+ * — classification is byte-stable).
+ */
+const TEST_TESTS_DIR_RE = /(?:^|\/)__tests__\//;
+const TEST_FIXTURES_DIR_RE = /(?:^|\/)__fixtures__\//;
+const TEST_FILE_SUFFIX_RE = /\.test\.(?:ts|tsx|js|jsx)$|_test\.(?:ts|tsx|js|jsx)$/;
+/**
+ * Returns true if `filePathProjectRel` is a TypeScript / JS test file (or test fixture)
+ * by path convention. Path is project-relative with `/` separators (the same shape
+ * `CloneCandidate.filePath` carries).
+ */
+export function isTestFilePath(filePathProjectRel) {
+    return (TEST_TESTS_DIR_RE.test(filePathProjectRel) ||
+        TEST_FIXTURES_DIR_RE.test(filePathProjectRel) ||
+        TEST_FILE_SUFFIX_RE.test(filePathProjectRel));
+}
+//# sourceMappingURL=test-file.js.map

package/dist/test-file.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"test-file.js","sourceRoot":"","sources":["../src/test-file.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AACH,MAAM,iBAAiB,GAAG,qBAAqB,CAAC;AAChD,MAAM,oBAAoB,GAAG,wBAAwB,CAAC;AACtD,MAAM,mBAAmB,GAAG,sDAAsD,CAAC;AAEnF;;;;GAIG;AACH,MAAM,UAAU,cAAc,CAAC,kBAA0B;IACvD,OAAO,CACL,iBAAiB,CAAC,IAAI,CAAC,kBAAkB,CAAC;QAC1C,oBAAoB,CAAC,IAAI,CAAC,kBAAkB,CAAC;QAC7C,mBAAmB,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAC7C,CAAC;AACJ,CAAC"}

package/dist/types.d.ts

@@ -0,0 +1,81 @@
+/**
+ * Tool-neutral input + finding types for the clone-detection substrate.
+ *
+ * `CloneCandidate` is a deliberate STRUCTURAL SUBSET of graph's runtime
+ * `FunctionOccurrence` (ISP — the algorithms accept only the fields the math needs;
+ * graph passes its occurrences directly by structural assignability, yagni builds the
+ * same shape from `lang-typescript`). This is the DIP/ISP boundary that keeps the
+ * substrate free of either tool's types — NOT a future-proofing hook. The cross-tool
+ * parity test (Phase 2) is the seam that proves both producers agree on these fields;
+ * the dep-cruiser leaf rule is the compile-time invariant.
+ */
+/** Function-shape classification — mirrors graph's `FunctionKind` for structural assignability. */
+export type FunctionKind = 'function-declaration' | 'function-expression' | 'arrow' | 'method' | 'constructor' | 'getter' | 'setter' | 'module-init';
+/** The minimal per-function input the detection algorithms consume. */
+export interface CloneCandidate {
+    /** sha256(normalized body) — the grouping key. */
+    readonly bodyHash: string;
+    /** MinHash signature (k=128) — near-duplicate only; absent ⇒ skipped by `findNearDuplicates`. */
+    readonly bodySignature?: readonly number[];
+    /** Normalized body length in chars; absent ⇒ "passes the size floor". */
+    readonly bodySize?: number;
+    /** Canonical body span in lines; absent ⇒ fall back to `endLine − line + 1`. */
+    readonly bodyLines?: number;
+    readonly kind: FunctionKind;
+    readonly inTestFile: boolean;
+    readonly filePath: string;
+    readonly line: number;
+    readonly column: number;
+    readonly endLine: number;
+    readonly simpleName: string;
+    readonly qualifiedName: string;
+    /** Resolved package the caller assigns (graph: `pkgOf`; yagni: nearest `package.json`). */
+    readonly package?: string;
+    /**
+     * Resolved language the caller assigns for the near-duplicate same-language gate
+     * (graph: `languageOfFile(filePath)`). `languageOfFile` stays in graph (single
+     * caller — rule of three not met), so the substrate compares this pre-resolved
+     * field rather than importing a language map. Absent ⇒ the pair is skipped, exactly
+     * as `languageOfFile(...) === undefined` did.
+     */
+    readonly language?: string;
+}
+/** Thresholds for `findDuplicateBodies` (exact). Absent fields take the policy defaults. */
+export interface DupOpts {
+    readonly minLines?: number;
+    readonly minBodySize?: number;
+    readonly minCrossPackagePackages?: number;
+    readonly minCrossPackageBodySize?: number;
+}
+/** Thresholds for `findNearDuplicates` (MinHash/LSH). Absent fields take the policy defaults. */
+export interface NearDupOpts {
+    readonly minSimilarity?: number;
+    readonly minBodySize?: number;
+    readonly lshBands?: number;
+}
+/** A per-instance exact-duplicate group. `members[0]` is the primary (lowest qualifiedName). */
+export interface DuplicateGroup {
+    readonly bodyHash: string;
+    readonly members: readonly CloneCandidate[];
+}
+/** A body duplicated across ≥ `minCrossPackagePackages` distinct packages. */
+export interface CrossPackageAggregate {
+    readonly bodyHash: string;
+    readonly anchor: CloneCandidate;
+    readonly packages: readonly string[];
+    readonly occurrenceCount: number;
+}
+/** Result of {@link findDuplicateBodies}: cross-package aggregates + the surviving per-instance groups. */
+export interface DuplicateFindings {
+    readonly aggregates: readonly CrossPackageAggregate[];
+    readonly groups: readonly DuplicateGroup[];
+}
+/** A near-duplicate cluster (connected component over LSH/Jaccard edges). */
+export interface NearDuplicateCluster {
+    readonly anchor: CloneCandidate;
+    readonly nearMembers: readonly string[];
+    readonly exactMembers: readonly string[];
+    readonly estimatedSimilarity: number;
+    readonly clusterSize: number;
+}
+//# sourceMappingURL=types.d.ts.map

package/dist/types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../src/types.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,mGAAmG;AACnG,MAAM,MAAM,YAAY,GACpB,sBAAsB,GACtB,qBAAqB,GACrB,OAAO,GACP,QAAQ,GACR,aAAa,GACb,QAAQ,GACR,QAAQ,GACR,aAAa,CAAC;AAElB,uEAAuE;AACvE,MAAM,WAAW,cAAc;IAC7B,kDAAkD;IAClD,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,iGAAiG;IACjG,QAAQ,CAAC,aAAa,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;IAC3C,yEAAyE;IACzE,QAAQ,CAAC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAC3B,gFAAgF;IAChF,QAAQ,CAAC,SAAS,CAAC,EAAE,MAAM,CAAC;IAC5B,QAAQ,CAAC,IAAI,EAAE,YAAY,CAAC;IAC5B,QAAQ,CAAC,UAAU,EAAE,OAAO,CAAC;IAC7B,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,MAAM,EAAE,MAAM,CAAC;IACxB,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC;IACzB,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;IAC5B,QAAQ,CAAC,aAAa,EAAE,MAAM,CAAC;IAC/B,2FAA2F;IAC3F,QAAQ,CAAC,OAAO,CAAC,EAAE,MAAM,CAAC;IAC1B;;;;;;OAMG;IACH,QAAQ,CAAC,QAAQ,CAAC,EAAE,MAAM,CAAC;CAC5B;AAED,4FAA4F;AAC5F,MAAM,WAAW,OAAO;IACtB,QAAQ,CAAC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAC3B,QAAQ,CAAC,WAAW,CAAC,EAAE,MAAM,CAAC;IAC9B,QAAQ,CAAC,uBAAuB,CAAC,EAAE,MAAM,CAAC;IAC1C,QAAQ,CAAC,uBAAuB,CAAC,EAAE,MAAM,CAAC;CAC3C;AAED,iGAAiG;AACjG,MAAM,WAAW,WAAW;IAC1B,QAAQ,CAAC,aAAa,CAAC,EAAE,MAAM,CAAC;IAChC,QAAQ,CAAC,WAAW,CAAC,EAAE,MAAM,CAAC;IAC9B,QAAQ,CAAC,QAAQ,CAAC,EAAE,MAAM,CAAC;CAC5B;AAED,gGAAgG;AAChG,MAAM,WAAW,cAAc;IAC7B,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,OAAO,EAAE,SAAS,cAAc,EAAE,CAAC;CAC7C;AAED,8EAA8E;AAC9E,MAAM,WAAW,qBAAqB;IACpC,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,QAAQ,CAAC,MAAM,EAAE,cAAc,CAAC;IAChC,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;IACrC,QAAQ,CAAC,eAAe,EAAE,MAAM,CAAC;CAClC;AAED,2GAA2G;AAC3G,MAAM,WAAW,iBAAiB;IAChC,QAAQ,CAAC,UAAU,EAAE,SAAS,qBAAqB,EAAE,CAAC;IACtD,QAAQ,CAAC,MAAM,EAAE,SAAS,cAAc,EAAE,CAAC;CAC5C;AAED,6EAA6E;AAC7E,MAAM,WAAW,oBAAoB;IACnC,QAAQ,CAAC,MAAM,EAAE,cAAc,CAAC;IAChC,QAAQ,CAAC,WAAW,EAAE,SAAS,MAAM,EAAE,CAAC;IACxC,QAAQ,CAAC,YAAY,EAAE,SAAS,MAAM,EAAE,CAAC;IACzC,QAAQ,CAAC,mBAAmB,EAAE,MAAM,CAAC;IACrC,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAC;CAC9B"}