@monoes/monograph 1.1.0 → 1.2.1

package/dist/src/graph/baseline.js

@@ -0,0 +1,319 @@
+import { readFileSync, writeFileSync, existsSync } from 'fs';
+import { join } from 'path';
+import { createHash } from 'crypto';
+function fingerprintFinding(type, ...parts) {
+    return createHash('sha256').update([type, ...parts].join('::')).digest('hex').slice(0, 16);
+}
+/**
+ * Save the current set of findings as a baseline JSON file.
+ * @param baselinePath - path to write (e.g. .monomind/baseline.json)
+ * @param findings - current findings to persist
+ * @param projectPath - repo path for identification
+ */
+export function saveBaseline(baselinePath, findings, projectPath) {
+    const data = {
+        version: 1,
+        savedAt: new Date().toISOString(),
+        projectPath,
+        findings,
+    };
+    writeFileSync(baselinePath, JSON.stringify(data, null, 2), 'utf-8');
+}
+/**
+ * Load an existing baseline file.
+ */
+export function loadBaseline(baselinePath) {
+    if (!existsSync(baselinePath))
+        return null;
+    try {
+        return JSON.parse(readFileSync(baselinePath, 'utf-8'));
+    }
+    catch {
+        return null;
+    }
+}
+/**
+ * Compare a list of current findings against a baseline.
+ * Returns each finding annotated with introduced:true/false.
+ */
+export function compareWithBaseline(currentFindings, baseline) {
+    if (!baseline) {
+        // No baseline — all findings are "introduced"
+        return currentFindings.map(f => ({ ...f, introduced: true }));
+    }
+    // Primary: match by fingerprint (for baselines that have it)
+    const baselineFingerprints = new Set(baseline.findings
+        .filter(f => f.fingerprint)
+        .map(f => f.fingerprint));
+    // Fallback: match by key (for old baselines without fingerprint)
+    const baselineKeys = new Set(baseline.findings.map(f => f.key));
+    return currentFindings.map(f => {
+        if (f.fingerprint && baselineFingerprints.size > 0) {
+            return { ...f, introduced: !baselineFingerprints.has(f.fingerprint) };
+        }
+        return { ...f, introduced: !baselineKeys.has(f.key) };
+    });
+}
+/**
+ * Extract findings from the database to build a baseline.
+ * Collects: isolated nodes (no edges), nodes with only INFERRED edges,
+ * god nodes (degree > 50).
+ */
+export function extractFindingsFromDb(db, projectPath) {
+    const findings = [];
+    // Isolated nodes (no incoming or outgoing edges)
+    const isolated = db.prepare(`
+    SELECT n.id, n.name, n.file_path, n.start_line, n.label
+    FROM nodes n
+    WHERE NOT EXISTS (SELECT 1 FROM edges e WHERE e.source_id = n.id OR e.target_id = n.id)
+    AND n.label IN ('Function', 'Class', 'Method', 'Interface', 'Variable', 'Module', 'File')
+    LIMIT 500
+  `).all();
+    for (const n of isolated) {
+        findings.push({
+            key: `isolated:${n.file_path ?? n.id}:${n.name}`,
+            type: 'isolated_node',
+            nodeId: n.id,
+            nodeName: n.name,
+            filePath: n.file_path,
+            startLine: n.start_line ?? null,
+            savedAt: new Date().toISOString(),
+            fingerprint: fingerprintFinding('isolated_node', n.file_path ?? '', n.id),
+        });
+    }
+    // God nodes (degree > 50)
+    const gods = db.prepare(`
+    SELECT n.id, n.name, n.file_path, n.start_line,
+           COUNT(DISTINCT e1.id) + COUNT(DISTINCT e2.id) as degree
+    FROM nodes n
+    LEFT JOIN edges e1 ON e1.source_id = n.id
+    LEFT JOIN edges e2 ON e2.target_id = n.id
+    GROUP BY n.id
+    HAVING degree > 50
+    ORDER BY degree DESC
+    LIMIT 100
+  `).all();
+    for (const n of gods) {
+        findings.push({
+            key: `god_node:${n.file_path ?? n.id}:${n.name}`,
+            type: 'god_node',
+            nodeId: n.id,
+            nodeName: n.name,
+            filePath: n.file_path,
+            startLine: n.start_line ?? null,
+            savedAt: new Date().toISOString(),
+            fingerprint: fingerprintFinding('god_node', n.file_path ?? '', n.id),
+        });
+    }
+    // Surprise edges: non-EXTRACTED confidence (cross-community low-confidence edges)
+    const surprises = db.prepare(`
+    SELECT DISTINCT e.id as edge_id, e.source_id, e.target_id, e.confidence,
+           n.name as src_name, n.file_path as src_file, n.start_line as src_line
+    FROM edges e
+    JOIN nodes n ON n.id = e.source_id
+    WHERE e.confidence != 'EXTRACTED'
+    LIMIT 200
+  `).all();
+    for (const e of surprises) {
+        findings.push({
+            key: `surprise:${e.source_id}:${e.target_id}`,
+            type: 'surprise',
+            nodeId: e.source_id,
+            nodeName: e.src_name,
+            filePath: e.src_file,
+            startLine: e.src_line ?? null,
+            savedAt: new Date().toISOString(),
+            fingerprint: fingerprintFinding('surprise', e.source_id, e.target_id),
+        });
+    }
+    // Bridge nodes: nodes whose edges span more than one community_id
+    // Use JOIN+GROUP BY instead of correlated subqueries to avoid N*2 subquery executions
+    const bridges = db.prepare(`
+    SELECT n.id, n.name, n.file_path, n.start_line,
+           COUNT(DISTINCT CASE WHEN e_out.source_id = n.id AND n_tgt.community_id != n.community_id THEN n_tgt.community_id END) +
+           COUNT(DISTINCT CASE WHEN e_in.target_id = n.id AND n_src.community_id != n.community_id THEN n_src.community_id END) AS cross_communities
+    FROM nodes n
+    LEFT JOIN edges e_out ON e_out.source_id = n.id
+    LEFT JOIN nodes n_tgt ON n_tgt.id = e_out.target_id AND n_tgt.community_id IS NOT NULL
+    LEFT JOIN edges e_in ON e_in.target_id = n.id
+    LEFT JOIN nodes n_src ON n_src.id = e_in.source_id AND n_src.community_id IS NOT NULL
+    WHERE n.community_id IS NOT NULL
+    GROUP BY n.id
+    HAVING cross_communities >= 2
+    LIMIT 100
+  `).all();
+    for (const n of bridges) {
+        findings.push({
+            key: `bridge_node:${n.file_path ?? n.id}:${n.name}`,
+            type: 'bridge_node',
+            nodeId: n.id,
+            nodeName: n.name,
+            filePath: n.file_path,
+            startLine: n.start_line ?? null,
+            savedAt: new Date().toISOString(),
+            fingerprint: fingerprintFinding('bridge_node', n.file_path ?? '', n.id),
+        });
+    }
+    // Unreachable exports: exported nodes with no incoming edges
+    const unreachableExports = db.prepare(`
+    SELECT n.id, n.name, n.file_path, n.start_line
+    FROM nodes n
+    WHERE n.is_exported = 1
+    AND n.label IN ('Function', 'Class', 'Method', 'Interface', 'Variable')
+    AND NOT EXISTS (SELECT 1 FROM edges e WHERE e.target_id = n.id)
+    LIMIT 300
+  `).all();
+    for (const n of unreachableExports) {
+        findings.push({
+            key: `unreachable_export:${n.file_path ?? n.id}:${n.name}`,
+            type: 'unreachable_export',
+            nodeId: n.id,
+            nodeName: n.name,
+            filePath: n.file_path,
+            startLine: n.start_line ?? null,
+            savedAt: new Date().toISOString(),
+            fingerprint: fingerprintFinding('unreachable_export', n.file_path ?? '', n.id),
+        });
+    }
+    // Ambiguous edges: edges with AMBIGUOUS confidence
+    const ambiguousEdges = db.prepare(`
+    SELECT DISTINCT e.id as edge_id, e.source_id, e.target_id,
+           n.name as src_name, n.file_path as src_file, n.start_line as src_line
+    FROM edges e
+    JOIN nodes n ON n.id = e.source_id
+    WHERE e.confidence = 'AMBIGUOUS'
+    LIMIT 200
+  `).all();
+    for (const e of ambiguousEdges) {
+        findings.push({
+            key: `ambiguous_edge:${e.source_id}:${e.target_id}`,
+            type: 'ambiguous_edge',
+            nodeId: e.source_id,
+            nodeName: e.src_name,
+            filePath: e.src_file,
+            startLine: e.src_line ?? null,
+            savedAt: new Date().toISOString(),
+            fingerprint: fingerprintFinding('ambiguous_edge', e.source_id, e.target_id),
+        });
+    }
+    return findings;
+}
+/**
+ * Default baseline path relative to a project directory.
+ * If a name is provided, the file is saved as `baseline-{name}.json`.
+ */
+export function defaultBaselinePath(projectDir, name) {
+    const filename = name ? `baseline-${name}.json` : 'baseline.json';
+    return join(projectDir, '.monomind', filename);
+}
+/**
+ * Domain-aware polarity for each metric.
+ * true  = higher value is better (improving)
+ * false = lower value is better (improving)
+ */
+const METRIC_HIGHER_IS_BETTER = {
+    nodeCount: true,
+    edgeCount: false, // neutral; treated as stable unless large change
+    communityCount: true,
+    godNodeCount: false,
+    surpriseCount: false,
+    hotspotCount: false,
+    unreachableNodeCount: false,
+};
+function trendDirection(metric, delta) {
+    if (delta === 0)
+        return { direction: 'stable', symbol: '→' };
+    // edgeCount is neutral — always stable regardless of direction
+    if (metric === 'edgeCount')
+        return { direction: 'stable', symbol: '→' };
+    const higherIsBetter = METRIC_HIGHER_IS_BETTER[metric] ?? true;
+    if (delta > 0) {
+        return higherIsBetter
+            ? { direction: 'improving', symbol: '↑' }
+            : { direction: 'declining', symbol: '↑' };
+    }
+    // delta < 0
+    return higherIsBetter
+        ? { direction: 'declining', symbol: '↓' }
+        : { direction: 'improving', symbol: '↓' };
+}
+/**
+ * Compute a trend report by comparing two BaselineData snapshots.
+ * Each baseline must carry BaselineVitals fields (nodeCount, edgeCount, …).
+ */
+export function computeTrend(before, after) {
+    const metricKeys = [
+        'nodeCount',
+        'edgeCount',
+        'communityCount',
+        'godNodeCount',
+        'surpriseCount',
+        'hotspotCount',
+        'unreachableNodeCount',
+    ];
+    const metrics = [];
+    for (const key of metricKeys) {
+        const prev = before[key] ?? 0;
+        const curr = after[key] ?? 0;
+        const delta = curr - prev;
+        const { direction, symbol } = trendDirection(key, delta);
+        metrics.push({ metric: key, previous: prev, current: curr, delta, direction, symbol });
+    }
+    // Overall direction: majority vote (improving/declining/stable)
+    const counts = { improving: 0, declining: 0, stable: 0 };
+    for (const m of metrics)
+        counts[m.direction]++;
+    let overallDirection = 'stable';
+    if (counts.improving > counts.declining && counts.improving > counts.stable) {
+        overallDirection = 'improving';
+    }
+    else if (counts.declining > counts.improving && counts.declining > counts.stable) {
+        overallDirection = 'declining';
+    }
+    return { metrics, overallDirection };
+}
+/** Format a list of ComparedFindings as structured text with file:line hints. */
+export function formatComparedFindings(findings, showAll = false) {
+    if (findings.length === 0)
+        return 'Baseline comparison: no findings.';
+    const newFindings = findings.filter(f => f.introduced);
+    const existing = findings.filter(f => !f.introduced);
+    const lines = [
+        `Baseline comparison: ${findings.length} total finding(s) — ${newFindings.length} new, ${existing.length} known.`,
+        '',
+    ];
+    if (newFindings.length > 0) {
+        lines.push(`New findings (${newFindings.length}):`);
+        for (const f of newFindings) {
+            const ref = f.filePath
+                ? f.startLine != null ? `${f.filePath}:${f.startLine}` : f.filePath
+                : f.nodeId;
+            lines.push(`  [${f.type}]  ${f.nodeName}  ${ref}`);
+        }
+        lines.push('');
+    }
+    if (showAll && existing.length > 0) {
+        lines.push(`Known findings (${existing.length}):`);
+        for (const f of existing) {
+            const ref = f.filePath
+                ? f.startLine != null ? `${f.filePath}:${f.startLine}` : f.filePath
+                : f.nodeId;
+            lines.push(`  [${f.type}]  ${f.nodeName}  ${ref}`);
+        }
+        lines.push('');
+    }
+    return lines.join('\n').trimEnd();
+}
+/** Format a TrendReport as structured text for LLM consumption. */
+export function formatTrendReport(report) {
+    const lines = [
+        `Graph trend: ${report.overallDirection}`,
+        '',
+    ];
+    for (const m of report.metrics) {
+        const deltaStr = m.delta > 0 ? `+${m.delta}` : `${m.delta}`;
+        lines.push(`  ${m.symbol} ${m.metric}: ${m.previous} → ${m.current} (${deltaStr})  [${m.direction}]`);
+    }
+    return lines.join('\n');
+}
+//# sourceMappingURL=baseline.js.map

package/dist/src/graph/baseline.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"baseline.js","sourceRoot":"","sources":["../../../src/graph/baseline.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAE,aAAa,EAAE,UAAU,EAAE,MAAM,IAAI,CAAC;AAC7D,OAAO,EAAE,IAAI,EAAE,MAAM,MAAM,CAAC;AAC5B,OAAO,EAAE,UAAU,EAAE,MAAM,QAAQ,CAAC;AAEpC,SAAS,kBAAkB,CAAC,IAAY,EAAE,GAAG,KAAe;IAC1D,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,GAAG,KAAK,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;AAC7F,CAAC;AA0BD;;;;;GAKG;AACH,MAAM,UAAU,YAAY,CAC1B,YAAoB,EACpB,QAA2B,EAC3B,WAAmB;IAEnB,MAAM,IAAI,GAAiB;QACzB,OAAO,EAAE,CAAC;QACV,OAAO,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACjC,WAAW;QACX,QAAQ;KACT,CAAC;IACF,aAAa,CAAC,YAAY,EAAE,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;AACtE,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,YAAY,CAAC,YAAoB;IAC/C,IAAI,CAAC,UAAU,CAAC,YAAY,CAAC;QAAE,OAAO,IAAI,CAAC;IAC3C,IAAI,CAAC;QACH,OAAO,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,YAAY,EAAE,OAAO,CAAC,CAAiB,CAAC;IACzE,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,mBAAmB,CACjC,eAAkC,EAClC,QAA6B;IAE7B,IAAI,CAAC,QAAQ,EAAE,CAAC;QACd,8CAA8C;QAC9C,OAAO,eAAe,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IAChE,CAAC;IACD,6DAA6D;IAC7D,MAAM,oBAAoB,GAAG,IAAI,GAAG,CAClC,QAAQ,CAAC,QAAQ;SACd,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,WAAW,CAAC;SAC1B,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,WAAqB,CAAC,CACrC,CAAC;IACF,iEAAiE;IACjE,MAAM,YAAY,GAAG,IAAI,GAAG,CAAC,QAAQ,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAChE,OAAO,eAAe,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE;QAC7B,IAAI,CAAC,CAAC,WAAW,IAAI,oBAAoB,CAAC,IAAI,GAAG,CAAC,EAAE,CAAC;YACnD,OAAO,EAAE,GAAG,CAAC,EAAE,UAAU,EAAE,CAAC,oBAAoB,CAAC,GAAG,CAAC,CAAC,CAAC,WAAW,CAAC,EAAE,CAAC;QACxE,CAAC;QACD,OAAO,EAAE,GAAG,CAAC,EAAE,UAAU,EAAE,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC;IACxD,CAAC,CAAC,CAAC;AACL,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,qBAAqB,CACnC,EAAqB,EACrB,WAAmB;IAEnB,MAAM,QAAQ,GAAsB,EAAE,CAAC;IAEvC,iDAAiD;IACjD,MAAM,QAAQ,GAAG,EAAE,CAAC,OAAO,CAAC;;;;;;GAM3B,CAAC,CAAC,GAAG,EAAwG,CAAC;IAE/G,KAAK,MAAM,CAAC,IAAI,QAAQ,EAAE,CAAC;QACzB,QAAQ,CAAC,IAAI,CAAC;YACZ,GAAG,EAAE,YAAY,CAAC,CAAC,SAAS,IAAI,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,IAAI,EAAE;YAChD,IAAI,EAAE,eAAe;YACrB,MAAM,EAAE,CAAC,CAAC,EAAE;YACZ,QAAQ,EAAE,CAAC,CAAC,IAAI;YAChB,QAAQ,EAAE,CAAC,CAAC,SAAS;YACrB,SAAS,EAAE,CAAC,CAAC,UAAU,IAAI,IAAI;YAC/B,OAAO,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACjC,WAAW,EAAE,kBAAkB,CAAC,eAAe,EAAE,CAAC,CAAC,SAAS,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC;SAC1E,CAAC,CAAC;IACL,CAAC;IAED,0BAA0B;IAC1B,MAAM,IAAI,GAAG,EAAE,CAAC,OAAO,CAAC;;;;;;;;;;GAUvB,CAAC,CAAC,GAAG,EAAyG,CAAC;IAEhH,KAAK,MAAM,CAAC,IAAI,IAAI,EAAE,CAAC;QACrB,QAAQ,CAAC,IAAI,CAAC;YACZ,GAAG,EAAE,YAAY,CAAC,CAAC,SAAS,IAAI,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,IAAI,EAAE;YAChD,IAAI,EAAE,UAAU;YAChB,MAAM,EAAE,CAAC,CAAC,EAAE;YACZ,QAAQ,EAAE,CAAC,CAAC,IAAI;YAChB,QAAQ,EAAE,CAAC,CAAC,SAAS;YACrB,SAAS,EAAE,CAAC,CAAC,UAAU,IAAI,IAAI;YAC/B,OAAO,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACjC,WAAW,EAAE,kBAAkB,CAAC,UAAU,EAAE,CAAC,CAAC,SAAS,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC;SACrE,CAAC,CAAC;IACL,CAAC;IAED,kFAAkF;IAClF,MAAM,SAAS,GAAG,EAAE,CAAC,OAAO,CAAC;;;;;;;GAO5B,CAAC,CAAC,GAAG,EAAyJ,CAAC;IAEhK,KAAK,MAAM,CAAC,IAAI,SAAS,EAAE,CAAC;QAC1B,QAAQ,CAAC,IAAI,CAAC;YACZ,GAAG,EAAE,YAAY,CAAC,CAAC,SAAS,IAAI,CAAC,CAAC,SAAS,EAAE;YAC7C,IAAI,EAAE,UAAU;YAChB,MAAM,EAAE,CAAC,CAAC,SAAS;YACnB,QAAQ,EAAE,CAAC,CAAC,QAAQ;YACpB,QAAQ,EAAE,CAAC,CAAC,QAAQ;YACpB,SAAS,EAAE,CAAC,CAAC,QAAQ,IAAI,IAAI;YAC7B,OAAO,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACjC,WAAW,EAAE,kBAAkB,CAAC,UAAU,EAAE,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,SAAS,CAAC;SACtE,CAAC,CAAC;IACL,CAAC;IAED,kEAAkE;IAClE,sFAAsF;IACtF,MAAM,OAAO,GAAG,EAAE,CAAC,OAAO,CAAC;;;;;;;;;;;;;GAa1B,CAAC,CAAC,GAAG,EAAyF,CAAC;IAEhG,KAAK,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;QACxB,QAAQ,CAAC,IAAI,CAAC;YACZ,GAAG,EAAE,eAAe,CAAC,CAAC,SAAS,IAAI,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,IAAI,EAAE;YACnD,IAAI,EAAE,aAAa;YACnB,MAAM,EAAE,CAAC,CAAC,EAAE;YACZ,QAAQ,EAAE,CAAC,CAAC,IAAI;YAChB,QAAQ,EAAE,CAAC,CAAC,SAAS;YACrB,SAAS,EAAE,CAAC,CAAC,UAAU,IAAI,IAAI;YAC/B,OAAO,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACjC,WAAW,EAAE,kBAAkB,CAAC,aAAa,EAAE,CAAC,CAAC,SAAS,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC;SACxE,CAAC,CAAC;IACL,CAAC;IAED,6DAA6D;IAC7D,MAAM,kBAAkB,GAAG,EAAE,CAAC,OAAO,CAAC;;;;;;;GAOrC,CAAC,CAAC,GAAG,EAAyF,CAAC;IAEhG,KAAK,MAAM,CAAC,IAAI,kBAAkB,EAAE,CAAC;QACnC,QAAQ,CAAC,IAAI,CAAC;YACZ,GAAG,EAAE,sBAAsB,CAAC,CAAC,SAAS,IAAI,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,IAAI,EAAE;YAC1D,IAAI,EAAE,oBAAoB;YAC1B,MAAM,EAAE,CAAC,CAAC,EAAE;YACZ,QAAQ,EAAE,CAAC,CAAC,IAAI;YAChB,QAAQ,EAAE,CAAC,CAAC,SAAS;YACrB,SAAS,EAAE,CAAC,CAAC,UAAU,IAAI,IAAI;YAC/B,OAAO,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACjC,WAAW,EAAE,kBAAkB,CAAC,oBAAoB,EAAE,CAAC,CAAC,SAAS,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,CAAC;SAC/E,CAAC,CAAC;IACL,CAAC;IAED,mDAAmD;IACnD,MAAM,cAAc,GAAG,EAAE,CAAC,OAAO,CAAC;;;;;;;GAOjC,CAAC,CAAC,GAAG,EAAqI,CAAC;IAE5I,KAAK,MAAM,CAAC,IAAI,cAAc,EAAE,CAAC;QAC/B,QAAQ,CAAC,IAAI,CAAC;YACZ,GAAG,EAAE,kBAAkB,CAAC,CAAC,SAAS,IAAI,CAAC,CAAC,SAAS,EAAE;YACnD,IAAI,EAAE,gBAAgB;YACtB,MAAM,EAAE,CAAC,CAAC,SAAS;YACnB,QAAQ,EAAE,CAAC,CAAC,QAAQ;YACpB,QAAQ,EAAE,CAAC,CAAC,QAAQ;YACpB,SAAS,EAAE,CAAC,CAAC,QAAQ,IAAI,IAAI;YAC7B,OAAO,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACjC,WAAW,EAAE,kBAAkB,CAAC,gBAAgB,EAAE,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,SAAS,CAAC;SAC5E,CAAC,CAAC;IACL,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,mBAAmB,CAAC,UAAkB,EAAE,IAAa;IACnE,MAAM,QAAQ,GAAG,IAAI,CAAC,CAAC,CAAC,YAAY,IAAI,OAAO,CAAC,CAAC,CAAC,eAAe,CAAC;IAClE,OAAO,IAAI,CAAC,UAAU,EAAE,WAAW,EAAE,QAAQ,CAAC,CAAC;AACjD,CAAC;AAkCD;;;;GAIG;AACH,MAAM,uBAAuB,GAA4B;IACvD,SAAS,EAAE,IAAI;IACf,SAAS,EAAE,KAAK,EAAS,iDAAiD;IAC1E,cAAc,EAAE,IAAI;IACpB,YAAY,EAAE,KAAK;IACnB,aAAa,EAAE,KAAK;IACpB,YAAY,EAAE,KAAK;IACnB,oBAAoB,EAAE,KAAK;CAC5B,CAAC;AAEF,SAAS,cAAc,CACrB,MAAc,EACd,KAAa;IAEb,IAAI,KAAK,KAAK,CAAC;QAAE,OAAO,EAAE,SAAS,EAAE,QAAQ,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;IAE7D,+DAA+D;IAC/D,IAAI,MAAM,KAAK,WAAW;QAAE,OAAO,EAAE,SAAS,EAAE,QAAQ,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;IAExE,MAAM,cAAc,GAAG,uBAAuB,CAAC,MAAM,CAAC,IAAI,IAAI,CAAC;IAE/D,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;QACd,OAAO,cAAc;YACnB,CAAC,CAAC,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,EAAE,GAAG,EAAE;YACzC,CAAC,CAAC,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;IAC9C,CAAC;IACD,YAAY;IACZ,OAAO,cAAc;QACnB,CAAC,CAAC,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,EAAE,GAAG,EAAE;QACzC,CAAC,CAAC,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;AAC9C,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,YAAY,CAC1B,MAAqC,EACrC,KAAoC;IAEpC,MAAM,UAAU,GAAgC;QAC9C,WAAW;QACX,WAAW;QACX,gBAAgB;QAChB,cAAc;QACd,eAAe;QACf,cAAc;QACd,sBAAsB;KACvB,CAAC;IAEF,MAAM,OAAO,GAAkB,EAAE,CAAC;IAElC,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAC7B,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAC9B,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAC7B,MAAM,KAAK,GAAG,IAAI,GAAG,IAAI,CAAC;QAC1B,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,GAAG,cAAc,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QACzD,OAAO,CAAC,IAAI,CAAC,EAAE,MAAM,EAAE,GAAG,EAAE,QAAQ,EAAE,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC;IACzF,CAAC;IAED,gEAAgE;IAChE,MAAM,MAAM,GAAG,EAAE,SAAS,EAAE,CAAC,EAAE,SAAS,EAAE,CAAC,EAAE,MAAM,EAAE,CAAC,EAAE,CAAC;IACzD,KAAK,MAAM,CAAC,IAAI,OAAO;QAAE,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC,EAAE,CAAC;IAE/C,IAAI,gBAAgB,GAAmB,QAAQ,CAAC;IAChD,IAAI,MAAM,CAAC,SAAS,GAAG,MAAM,CAAC,SAAS,IAAI,MAAM,CAAC,SAAS,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;QAC5E,gBAAgB,GAAG,WAAW,CAAC;IACjC,CAAC;SAAM,IAAI,MAAM,CAAC,SAAS,GAAG,MAAM,CAAC,SAAS,IAAI,MAAM,CAAC,SAAS,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC;QACnF,gBAAgB,GAAG,WAAW,CAAC;IACjC,CAAC;IAED,OAAO,EAAE,OAAO,EAAE,gBAAgB,EAAE,CAAC;AACvC,CAAC;AAED,iFAAiF;AACjF,MAAM,UAAU,sBAAsB,CAAC,QAA2B,EAAE,OAAO,GAAG,KAAK;IACjF,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,mCAAmC,CAAC;IAEtE,MAAM,WAAW,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IACvD,MAAM,QAAQ,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IAErD,MAAM,KAAK,GAAa;QACtB,wBAAwB,QAAQ,CAAC,MAAM,uBAAuB,WAAW,CAAC,MAAM,SAAS,QAAQ,CAAC,MAAM,SAAS;QACjH,EAAE;KACH,CAAC;IAEF,IAAI,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC3B,KAAK,CAAC,IAAI,CAAC,iBAAiB,WAAW,CAAC,MAAM,IAAI,CAAC,CAAC;QACpD,KAAK,MAAM,CAAC,IAAI,WAAW,EAAE,CAAC;YAC5B,MAAM,GAAG,GAAG,CAAC,CAAC,QAAQ;gBACpB,CAAC,CAAC,CAAC,CAAC,SAAS,IAAI,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,QAAQ,IAAI,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ;gBACnE,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;YACb,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,IAAI,MAAM,CAAC,CAAC,QAAQ,KAAK,GAAG,EAAE,CAAC,CAAC;QACrD,CAAC;QACD,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACjB,CAAC;IAED,IAAI,OAAO,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACnC,KAAK,CAAC,IAAI,CAAC,mBAAmB,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAC;QACnD,KAAK,MAAM,CAAC,IAAI,QAAQ,EAAE,CAAC;YACzB,MAAM,GAAG,GAAG,CAAC,CAAC,QAAQ;gBACpB,CAAC,CAAC,CAAC,CAAC,SAAS,IAAI,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,QAAQ,IAAI,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ;gBACnE,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;YACb,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,IAAI,MAAM,CAAC,CAAC,QAAQ,KAAK,GAAG,EAAE,CAAC,CAAC;QACrD,CAAC;QACD,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACjB,CAAC;IAED,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,CAAC;AACpC,CAAC;AAED,mEAAmE;AACnE,MAAM,UAAU,iBAAiB,CAAC,MAAmB;IACnD,MAAM,KAAK,GAAa;QACtB,gBAAgB,MAAM,CAAC,gBAAgB,EAAE;QACzC,EAAE;KACH,CAAC;IACF,KAAK,MAAM,CAAC,IAAI,MAAM,CAAC,OAAO,EAAE,CAAC;QAC/B,MAAM,QAAQ,GAAG,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,CAAC;QAC5D,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,MAAM,IAAI,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,QAAQ,MAAM,CAAC,CAAC,OAAO,KAAK,QAAQ,OAAO,CAAC,CAAC,SAAS,GAAG,CAAC,CAAC;IACxG,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC"}

package/dist/src/graph/clone-detector.d.ts

@@ -0,0 +1,15 @@
+import type { MonographDb } from '../storage/db.js';
+export interface ClonePair {
+    fileA: string;
+    fileB: string;
+    similarity: number;
+    cloneType: 'exact' | 'renamed' | 'structural';
+    tokenCount: number;
+}
+export interface CloneDetectionResult {
+    pairs: ClonePair[];
+    totalFiles: number;
+    cloneRatio: number;
+}
+export declare function detectClones(db: MonographDb, minSimilarity?: number, minTokens?: number): CloneDetectionResult;
+//# sourceMappingURL=clone-detector.d.ts.map

package/dist/src/graph/clone-detector.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"clone-detector.d.ts","sourceRoot":"","sources":["../../../src/graph/clone-detector.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,kBAAkB,CAAC;AAIpD,MAAM,WAAW,SAAS;IACxB,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,OAAO,GAAG,SAAS,GAAG,YAAY,CAAC;IAC9C,UAAU,EAAE,MAAM,CAAC;CACpB;AAED,MAAM,WAAW,oBAAoB;IACnC,KAAK,EAAE,SAAS,EAAE,CAAC;IACnB,UAAU,EAAE,MAAM,CAAC;IACnB,UAAU,EAAE,MAAM,CAAC;CACpB;AAwCD,wBAAgB,YAAY,CAC1B,EAAE,EAAE,WAAW,EACf,aAAa,SAAM,EACnB,SAAS,SAAI,GACZ,oBAAoB,CAsGtB"}

package/dist/src/graph/clone-detector.js

@@ -0,0 +1,132 @@
+import { insertEdge } from '../storage/edge-store.js';
+import { makeId } from '../types.js';
+/** Tokenize a file path into lowercase alphanumeric tokens. */
+function tokenizePath(filePath) {
+    return filePath
+        .toLowerCase()
+        .split(/[^a-z0-9]+/)
+        .filter(t => t.length > 0);
+}
+/** Extract the base filename (without directory) from a path. */
+function baseName(filePath) {
+    return filePath.replace(/^.*[\\/]/, '');
+}
+/**
+ * Compute both Jaccard similarity and shared token count in a single pass.
+ * Avoids building the countA Map twice (previously done separately in
+ * jaccardSimilarity and sharedTokenCount).
+ */
+function computeSimilarityAndShared(a, b) {
+    const countA = new Map();
+    for (const t of a)
+        countA.set(t, (countA.get(t) ?? 0) + 1);
+    let intersection = 0;
+    const seen = new Map();
+    for (const t of b) {
+        const limit = countA.get(t) ?? 0;
+        const used = seen.get(t) ?? 0;
+        if (used < limit) {
+            intersection++;
+            seen.set(t, used + 1);
+        }
+    }
+    const union = a.length + b.length - intersection;
+    const similarity = union === 0 ? 1 : intersection / union;
+    return { similarity, shared: intersection };
+}
+export function detectClones(db, minSimilarity = 0.8, minTokens = 2) {
+    // Query all File nodes with file_path
+    const rows = db
+        .prepare(`SELECT id, file_path FROM nodes WHERE label = 'File' AND file_path IS NOT NULL`)
+        .all();
+    const totalFiles = rows.length;
+    const pairs = [];
+    const involvedFiles = new Set();
+    // Pre-compute path tokens and name tokens for each file
+    const tokenMap = new Map();
+    for (const row of rows) {
+        const pathTokens = tokenizePath(row.file_path);
+        const nameTokens = tokenizePath(baseName(row.file_path));
+        tokenMap.set(row.file_path, { pathTokens, nameTokens });
+    }
+    // Compare all pairs
+    for (let i = 0; i < rows.length; i++) {
+        for (let j = i + 1; j < rows.length; j++) {
+            const pathA = rows[i].file_path;
+            const pathB = rows[j].file_path;
+            const entryA = tokenMap.get(pathA);
+            const entryB = tokenMap.get(pathB);
+            const { similarity: sim, shared } = computeSimilarityAndShared(entryA.pathTokens, entryB.pathTokens);
+            if (sim < minSimilarity || shared < minTokens) {
+                // Also check near-duplicate file names (same name, different dir)
+                const { similarity: nameSim, shared: nameShared } = computeSimilarityAndShared(entryA.nameTokens, entryB.nameTokens);
+                if (nameSim >= minSimilarity && nameShared >= Math.min(minTokens, entryA.nameTokens.length)) {
+                    // Same-name files in different directories → structural similarity
+                    const pair = {
+                        fileA: pathA,
+                        fileB: pathB,
+                        similarity: nameSim,
+                        cloneType: 'structural',
+                        tokenCount: nameShared,
+                    };
+                    pairs.push(pair);
+                    involvedFiles.add(pathA);
+                    involvedFiles.add(pathB);
+                    // Emit edge
+                    const edgeId = makeId('structurally_similar', rows[i].id, rows[j].id);
+                    try {
+                        insertEdge(db, {
+                            id: edgeId,
+                            sourceId: rows[i].id,
+                            targetId: rows[j].id,
+                            relation: 'STRUCTURALLY_SIMILAR',
+                            confidence: 'INFERRED',
+                            confidenceScore: nameSim,
+                            weight: nameSim,
+                        });
+                    }
+                    catch { /* ignore duplicate edge errors */ }
+                }
+                continue;
+            }
+            // Determine clone type
+            let cloneType;
+            if (sim >= 0.99) {
+                cloneType = 'exact';
+            }
+            else if (baseName(pathA) === baseName(pathB)) {
+                cloneType = 'renamed';
+            }
+            else {
+                cloneType = 'structural';
+            }
+            const pair = {
+                fileA: pathA,
+                fileB: pathB,
+                similarity: sim,
+                cloneType,
+                tokenCount: shared,
+            };
+            pairs.push(pair);
+            involvedFiles.add(pathA);
+            involvedFiles.add(pathB);
+            // Emit STRUCTURALLY_SIMILAR edge into the edges table
+            const edgeId = makeId('structurally_similar', rows[i].id, rows[j].id);
+            try {
+                insertEdge(db, {
+                    id: edgeId,
+                    sourceId: rows[i].id,
+                    targetId: rows[j].id,
+                    relation: 'STRUCTURALLY_SIMILAR',
+                    confidence: 'INFERRED',
+                    confidenceScore: sim,
+                    weight: sim,
+                });
+            }
+            catch { /* ignore duplicate edge errors */ }
+        }
+    }
+    const cloneRatio = totalFiles === 0 ? 0 : involvedFiles.size / totalFiles;
+    return { pairs, totalFiles, cloneRatio };
+}
+//# sourceMappingURL=clone-detector.js.map

package/dist/src/graph/clone-detector.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"clone-detector.js","sourceRoot":"","sources":["../../../src/graph/clone-detector.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,UAAU,EAAE,MAAM,0BAA0B,CAAC;AACtD,OAAO,EAAE,MAAM,EAAE,MAAM,aAAa,CAAC;AAgBrC,+DAA+D;AAC/D,SAAS,YAAY,CAAC,QAAgB;IACpC,OAAO,QAAQ;SACZ,WAAW,EAAE;SACb,KAAK,CAAC,YAAY,CAAC;SACnB,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;AAC/B,CAAC;AAED,iEAAiE;AACjE,SAAS,QAAQ,CAAC,QAAgB;IAChC,OAAO,QAAQ,CAAC,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;AAC1C,CAAC;AAED;;;;GAIG;AACH,SAAS,0BAA0B,CAAC,CAAW,EAAE,CAAW;IAC1D,MAAM,MAAM,GAAG,IAAI,GAAG,EAAkB,CAAC;IACzC,KAAK,MAAM,CAAC,IAAI,CAAC;QAAE,MAAM,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IAE3D,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,MAAM,IAAI,GAAG,IAAI,GAAG,EAAkB,CAAC;IACvC,KAAK,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC;QAClB,MAAM,KAAK,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QACjC,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;QAC9B,IAAI,IAAI,GAAG,KAAK,EAAE,CAAC;YACjB,YAAY,EAAE,CAAC;YACf,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,GAAG,CAAC,CAAC,CAAC;QACxB,CAAC;IACH,CAAC;IAED,MAAM,KAAK,GAAG,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,MAAM,GAAG,YAAY,CAAC;IACjD,MAAM,UAAU,GAAG,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,YAAY,GAAG,KAAK,CAAC;IAC1D,OAAO,EAAE,UAAU,EAAE,MAAM,EAAE,YAAY,EAAE,CAAC;AAC9C,CAAC;AAED,MAAM,UAAU,YAAY,CAC1B,EAAe,EACf,aAAa,GAAG,GAAG,EACnB,SAAS,GAAG,CAAC;IAEb,sCAAsC;IACtC,MAAM,IAAI,GAAG,EAAE;SACZ,OAAO,CAAC,gFAAgF,CAAC;SACzF,GAAG,EAAyC,CAAC;IAEhD,MAAM,UAAU,GAAG,IAAI,CAAC,MAAM,CAAC;IAC/B,MAAM,KAAK,GAAgB,EAAE,CAAC;IAC9B,MAAM,aAAa,GAAG,IAAI,GAAG,EAAU,CAAC;IAExC,wDAAwD;IACxD,MAAM,QAAQ,GAAG,IAAI,GAAG,EAA0D,CAAC;IACnF,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;QACvB,MAAM,UAAU,GAAG,YAAY,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAC/C,MAAM,UAAU,GAAG,YAAY,CAAC,QAAQ,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,CAAC;QACzD,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,SAAS,EAAE,EAAE,UAAU,EAAE,UAAU,EAAE,CAAC,CAAC;IAC1D,CAAC;IAED,oBAAoB;IACpB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACzC,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;YAChC,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;YAChC,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAE,CAAC;YACpC,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAE,CAAC;YAEpC,MAAM,EAAE,UAAU,EAAE,GAAG,EAAE,MAAM,EAAE,GAAG,0BAA0B,CAAC,MAAM,CAAC,UAAU,EAAE,MAAM,CAAC,UAAU,CAAC,CAAC;YAErG,IAAI,GAAG,GAAG,aAAa,IAAI,MAAM,GAAG,SAAS,EAAE,CAAC;gBAC9C,kEAAkE;gBAClE,MAAM,EAAE,UAAU,EAAE,OAAO,EAAE,MAAM,EAAE,UAAU,EAAE,GAAG,0BAA0B,CAAC,MAAM,CAAC,UAAU,EAAE,MAAM,CAAC,UAAU,CAAC,CAAC;gBAErH,IAAI,OAAO,IAAI,aAAa,IAAI,UAAU,IAAI,IAAI,CAAC,GAAG,CAAC,SAAS,EAAE,MAAM,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;oBAC5F,mEAAmE;oBACnE,MAAM,IAAI,GAAc;wBACtB,KAAK,EAAE,KAAK;wBACZ,KAAK,EAAE,KAAK;wBACZ,UAAU,EAAE,OAAO;wBACnB,SAAS,EAAE,YAAY;wBACvB,UAAU,EAAE,UAAU;qBACvB,CAAC;oBACF,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;oBACjB,aAAa,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;oBACzB,aAAa,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;oBAEzB,YAAY;oBACZ,MAAM,MAAM,GAAG,MAAM,CAAC,sBAAsB,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;oBACtE,IAAI,CAAC;wBACH,UAAU,CAAC,EAAE,EAAE;4BACb,EAAE,EAAE,MAAM;4BACV,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE;4BACpB,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE;4BACpB,QAAQ,EAAE,sBAAsB;4BAChC,UAAU,EAAE,UAAU;4BACtB,eAAe,EAAE,OAAO;4BACxB,MAAM,EAAE,OAAO;yBAChB,CAAC,CAAC;oBACL,CAAC;oBAAC,MAAM,CAAC,CAAC,kCAAkC,CAAC,CAAC;gBAChD,CAAC;gBACD,SAAS;YACX,CAAC;YAED,uBAAuB;YACvB,IAAI,SAAiC,CAAC;YACtC,IAAI,GAAG,IAAI,IAAI,EAAE,CAAC;gBAChB,SAAS,GAAG,OAAO,CAAC;YACtB,CAAC;iBAAM,IAAI,QAAQ,CAAC,KAAK,CAAC,KAAK,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;gBAC/C,SAAS,GAAG,SAAS,CAAC;YACxB,CAAC;iBAAM,CAAC;gBACN,SAAS,GAAG,YAAY,CAAC;YAC3B,CAAC;YAED,MAAM,IAAI,GAAc;gBACtB,KAAK,EAAE,KAAK;gBACZ,KAAK,EAAE,KAAK;gBACZ,UAAU,EAAE,GAAG;gBACf,SAAS;gBACT,UAAU,EAAE,MAAM;aACnB,CAAC;YACF,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACjB,aAAa,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YACzB,aAAa,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YAEzB,sDAAsD;YACtD,MAAM,MAAM,GAAG,MAAM,CAAC,sBAAsB,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;YACtE,IAAI,CAAC;gBACH,UAAU,CAAC,EAAE,EAAE;oBACb,EAAE,EAAE,MAAM;oBACV,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE;oBACpB,QAAQ,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE;oBACpB,QAAQ,EAAE,sBAAsB;oBAChC,UAAU,EAAE,UAAU;oBACtB,eAAe,EAAE,GAAG;oBACpB,MAAM,EAAE,GAAG;iBACZ,CAAC,CAAC;YACL,CAAC;YAAC,MAAM,CAAC,CAAC,kCAAkC,CAAC,CAAC;QAChD,CAAC;IACH,CAAC;IAED,MAAM,UAAU,GAAG,UAAU,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,IAAI,GAAG,UAAU,CAAC;IAE1E,OAAO,EAAE,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,CAAC;AAC3C,CAAC"}

package/dist/src/graph/clone-families.d.ts

@@ -0,0 +1,34 @@
+export type RefactoringKind = 'ExtractFunction' | 'ExtractModule' | 'MergeDirectories';
+export declare const MODULE_EXTRACTION_THRESHOLD_LINES = 50;
+export interface RefactoringSuggestion {
+    kind: RefactoringKind;
+    description: string;
+    estimatedLines: number;
+    files: string[];
+}
+export interface CloneInstance {
+    filePath: string;
+    startLine: number;
+    endLine: number;
+    tokenCount?: number;
+}
+export interface CloneGroup {
+    id: number;
+    instances: CloneInstance[];
+    duplicatedLines: number;
+}
+export interface CloneFamily {
+    files: string[];
+    groups: CloneGroup[];
+    totalDuplicatedLines: number;
+    suggestions: RefactoringSuggestion[];
+}
+export declare function groupIntoFamilies(groups: CloneGroup[]): CloneFamily[];
+/** Format clone families as structured text for LLM consumption. */
+export declare function formatCloneFamilies(families: CloneFamily[]): string;
+export declare function cloneFamilySummary(families: CloneFamily[]): {
+    totalFamilies: number;
+    totalDuplicatedLines: number;
+    byKind: Record<RefactoringKind, number>;
+};
+//# sourceMappingURL=clone-families.d.ts.map

package/dist/src/graph/clone-families.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"clone-families.d.ts","sourceRoot":"","sources":["../../../src/graph/clone-families.ts"],"names":[],"mappings":"AAAA,MAAM,MAAM,eAAe,GAAG,iBAAiB,GAAG,eAAe,GAAG,kBAAkB,CAAC;AAEvF,eAAO,MAAM,iCAAiC,KAAK,CAAC;AAEpD,MAAM,WAAW,qBAAqB;IACpC,IAAI,EAAE,eAAe,CAAC;IACtB,WAAW,EAAE,MAAM,CAAC;IACpB,cAAc,EAAE,MAAM,CAAC;IACvB,KAAK,EAAE,MAAM,EAAE,CAAC;CACjB;AAED,MAAM,WAAW,aAAa;IAC5B,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;IAChB,UAAU,CAAC,EAAE,MAAM,CAAC;CACrB;AAED,MAAM,WAAW,UAAU;IACzB,EAAE,EAAE,MAAM,CAAC;IACX,SAAS,EAAE,aAAa,EAAE,CAAC;IAC3B,eAAe,EAAE,MAAM,CAAC;CACzB;AAED,MAAM,WAAW,WAAW;IAC1B,KAAK,EAAE,MAAM,EAAE,CAAC;IAChB,MAAM,EAAE,UAAU,EAAE,CAAC;IACrB,oBAAoB,EAAE,MAAM,CAAC;IAC7B,WAAW,EAAE,qBAAqB,EAAE,CAAC;CACtC;AA+CD,wBAAgB,iBAAiB,CAAC,MAAM,EAAE,UAAU,EAAE,GAAG,WAAW,EAAE,CAsBrE;AAED,oEAAoE;AACpE,wBAAgB,mBAAmB,CAAC,QAAQ,EAAE,WAAW,EAAE,GAAG,MAAM,CAUnE;AAED,wBAAgB,kBAAkB,CAAC,QAAQ,EAAE,WAAW,EAAE,GAAG;IAC3D,aAAa,EAAE,MAAM,CAAC;IACtB,oBAAoB,EAAE,MAAM,CAAC;IAC7B,MAAM,EAAE,MAAM,CAAC,eAAe,EAAE,MAAM,CAAC,CAAC;CACzC,CAcA"}

package/dist/src/graph/clone-families.js

@@ -0,0 +1,100 @@
+export const MODULE_EXTRACTION_THRESHOLD_LINES = 50;
+function fileSetKey(files) {
+    // Avoid spread allocation: sort a copy directly
+    return files.slice().sort().join('|');
+}
+function generateSuggestions(family) {
+    const suggestions = [];
+    const lines = family.totalDuplicatedLines;
+    if (lines < MODULE_EXTRACTION_THRESHOLD_LINES) {
+        suggestions.push({
+            kind: 'ExtractFunction',
+            description: `Extract ${lines} duplicated lines into a shared function`,
+            estimatedLines: lines,
+            files: family.files,
+        });
+    }
+    else {
+        // Check if all files are in the same directory → MergeDirectories
+        // Use Set directly to avoid spread+map intermediate array
+        const dirSet = new Set();
+        for (const f of family.files) {
+            const slash = f.replace(/\\/g, '/');
+            const lastSlash = slash.lastIndexOf('/');
+            dirSet.add(lastSlash >= 0 ? slash.slice(0, lastSlash) : '.');
+        }
+        const dirs = [...dirSet];
+        if (dirs.length > 1 && family.files.length >= 3) {
+            suggestions.push({
+                kind: 'MergeDirectories',
+                description: `Consider merging ${dirs.length} directories with ${lines} shared lines`,
+                estimatedLines: lines,
+                files: family.files,
+            });
+        }
+        else {
+            suggestions.push({
+                kind: 'ExtractModule',
+                description: `Extract ${lines} duplicated lines into a shared module`,
+                estimatedLines: lines,
+                files: family.files,
+            });
+        }
+    }
+    return suggestions;
+}
+export function groupIntoFamilies(groups) {
+    const familyMap = new Map();
+    for (const group of groups) {
+        const files = group.instances.map(i => i.filePath);
+        const key = fileSetKey(files);
+        if (!familyMap.has(key)) {
+            familyMap.set(key, { files: new Set(files), groups: [] });
+        }
+        familyMap.get(key).groups.push(group);
+    }
+    // Build output array without intermediate spread + map allocation
+    const result = [];
+    for (const { files, groups } of familyMap.values()) {
+        const fileList = [...files].sort();
+        let totalDuplicatedLines = 0;
+        for (const g of groups)
+            totalDuplicatedLines += g.duplicatedLines;
+        const base = { files: fileList, groups, totalDuplicatedLines };
+        result.push({ ...base, suggestions: generateSuggestions(base) });
+    }
+    return result.sort((a, b) => b.totalDuplicatedLines - a.totalDuplicatedLines);
+}
+/** Format clone families as structured text for LLM consumption. */
+export function formatCloneFamilies(families) {
+    if (families.length === 0)
+        return 'No clone families detected.';
+    const lines = [`Clone Families (${families.length} total):`];
+    for (const f of families.slice(0, 20)) {
+        lines.push(`\n  ${f.totalDuplicatedLines} duplicated lines across ${f.files.length} files:`);
+        for (const file of f.files)
+            lines.push(`    ${file}`);
+        for (const s of f.suggestions)
+            lines.push(`  -> [${s.kind}] ${s.description}`);
+    }
+    if (families.length > 20)
+        lines.push(`\n  ... and ${families.length - 20} more families`);
+    return lines.join('\n');
+}
+export function cloneFamilySummary(families) {
+    const byKind = {
+        ExtractFunction: 0, ExtractModule: 0, MergeDirectories: 0,
+    };
+    let totalDuplicatedLines = 0;
+    for (const f of families) {
+        totalDuplicatedLines += f.totalDuplicatedLines;
+        for (const s of f.suggestions)
+            byKind[s.kind]++;
+    }
+    return {
+        totalFamilies: families.length,
+        totalDuplicatedLines,
+        byKind,
+    };
+}
+//# sourceMappingURL=clone-families.js.map

package/dist/src/graph/clone-families.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"clone-families.js","sourceRoot":"","sources":["../../../src/graph/clone-families.ts"],"names":[],"mappings":"AAEA,MAAM,CAAC,MAAM,iCAAiC,GAAG,EAAE,CAAC;AA6BpD,SAAS,UAAU,CAAC,KAAe;IACjC,gDAAgD;IAChD,OAAO,KAAK,CAAC,KAAK,EAAE,CAAC,IAAI,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AACxC,CAAC;AAED,SAAS,mBAAmB,CAAC,MAAwC;IACnE,MAAM,WAAW,GAA4B,EAAE,CAAC;IAChD,MAAM,KAAK,GAAG,MAAM,CAAC,oBAAoB,CAAC;IAE1C,IAAI,KAAK,GAAG,iCAAiC,EAAE,CAAC;QAC9C,WAAW,CAAC,IAAI,CAAC;YACf,IAAI,EAAE,iBAAiB;YACvB,WAAW,EAAE,WAAW,KAAK,0CAA0C;YACvE,cAAc,EAAE,KAAK;YACrB,KAAK,EAAE,MAAM,CAAC,KAAK;SACpB,CAAC,CAAC;IACL,CAAC;SAAM,CAAC;QACN,kEAAkE;QAClE,0DAA0D;QAC1D,MAAM,MAAM,GAAG,IAAI,GAAG,EAAU,CAAC;QACjC,KAAK,MAAM,CAAC,IAAI,MAAM,CAAC,KAAK,EAAE,CAAC;YAC7B,MAAM,KAAK,GAAG,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;YACpC,MAAM,SAAS,GAAG,KAAK,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;YACzC,MAAM,CAAC,GAAG,CAAC,SAAS,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;QAC/D,CAAC;QACD,MAAM,IAAI,GAAG,CAAC,GAAG,MAAM,CAAC,CAAC;QACzB,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;YAChD,WAAW,CAAC,IAAI,CAAC;gBACf,IAAI,EAAE,kBAAkB;gBACxB,WAAW,EAAE,oBAAoB,IAAI,CAAC,MAAM,qBAAqB,KAAK,eAAe;gBACrF,cAAc,EAAE,KAAK;gBACrB,KAAK,EAAE,MAAM,CAAC,KAAK;aACpB,CAAC,CAAC;QACL,CAAC;aAAM,CAAC;YACN,WAAW,CAAC,IAAI,CAAC;gBACf,IAAI,EAAE,eAAe;gBACrB,WAAW,EAAE,WAAW,KAAK,wCAAwC;gBACrE,cAAc,EAAE,KAAK;gBACrB,KAAK,EAAE,MAAM,CAAC,KAAK;aACpB,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IACD,OAAO,WAAW,CAAC;AACrB,CAAC;AAED,MAAM,UAAU,iBAAiB,CAAC,MAAoB;IACpD,MAAM,SAAS,GAAG,IAAI,GAAG,EAAwD,CAAC;IAElF,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;QAC3B,MAAM,KAAK,GAAG,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC;QACnD,MAAM,GAAG,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC;QAC9B,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YACxB,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,EAAE,KAAK,EAAE,IAAI,GAAG,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,EAAE,EAAE,CAAC,CAAC;QAC5D,CAAC;QACD,SAAS,CAAC,GAAG,CAAC,GAAG,CAAE,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACzC,CAAC;IAED,kEAAkE;IAClE,MAAM,MAAM,GAAkB,EAAE,CAAC;IACjC,KAAK,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,SAAS,CAAC,MAAM,EAAE,EAAE,CAAC;QACnD,MAAM,QAAQ,GAAG,CAAC,GAAG,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;QACnC,IAAI,oBAAoB,GAAG,CAAC,CAAC;QAC7B,KAAK,MAAM,CAAC,IAAI,MAAM;YAAE,oBAAoB,IAAI,CAAC,CAAC,eAAe,CAAC;QAClE,MAAM,IAAI,GAAG,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,EAAE,oBAAoB,EAAE,CAAC;QAC/D,MAAM,CAAC,IAAI,CAAC,EAAE,GAAG,IAAI,EAAE,WAAW,EAAE,mBAAmB,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACnE,CAAC;IACD,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,oBAAoB,GAAG,CAAC,CAAC,oBAAoB,CAAC,CAAC;AAChF,CAAC;AAED,oEAAoE;AACpE,MAAM,UAAU,mBAAmB,CAAC,QAAuB;IACzD,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,6BAA6B,CAAC;IAChE,MAAM,KAAK,GAAa,CAAC,mBAAmB,QAAQ,CAAC,MAAM,UAAU,CAAC,CAAC;IACvE,KAAK,MAAM,CAAC,IAAI,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC;QACtC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,oBAAoB,4BAA4B,CAAC,CAAC,KAAK,CAAC,MAAM,SAAS,CAAC,CAAC;QAC7F,KAAK,MAAM,IAAI,IAAI,CAAC,CAAC,KAAK;YAAE,KAAK,CAAC,IAAI,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC;QACtD,KAAK,MAAM,CAAC,IAAI,CAAC,CAAC,WAAW;YAAE,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC;IACjF,CAAC;IACD,IAAI,QAAQ,CAAC,MAAM,GAAG,EAAE;QAAE,KAAK,CAAC,IAAI,CAAC,eAAe,QAAQ,CAAC,MAAM,GAAG,EAAE,gBAAgB,CAAC,CAAC;IAC1F,OAAO,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1B,CAAC;AAED,MAAM,UAAU,kBAAkB,CAAC,QAAuB;IAKxD,MAAM,MAAM,GAAoC;QAC9C,eAAe,EAAE,CAAC,EAAE,aAAa,EAAE,CAAC,EAAE,gBAAgB,EAAE,CAAC;KAC1D,CAAC;IACF,IAAI,oBAAoB,GAAG,CAAC,CAAC;IAC7B,KAAK,MAAM,CAAC,IAAI,QAAQ,EAAE,CAAC;QACzB,oBAAoB,IAAI,CAAC,CAAC,oBAAoB,CAAC;QAC/C,KAAK,MAAM,CAAC,IAAI,CAAC,CAAC,WAAW;YAAE,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC;IAClD,CAAC;IACD,OAAO;QACL,aAAa,EAAE,QAAQ,CAAC,MAAM;QAC9B,oBAAoB;QACpB,MAAM;KACP,CAAC;AACJ,CAAC"}