@nahisaho/musubix-pattern-mcp 1.0.0

package/dist/extractor/pattern-extractor.js

@@ -0,0 +1,201 @@
+/**
+ * @fileoverview AST-based pattern extraction
+ * @traceability TSK-PATTERN-001, REQ-PATTERN-001-F001
+ */
+import { TypeScriptParser } from './typescript-parser.js';
+import { SubtreeFinder } from './subtree-finder.js';
+import { AntiUnifier } from './anti-unifier.js';
+/**
+ * Pattern extractor using AST analysis
+ *
+ * @description
+ * Extracts recurring patterns from source code by:
+ * 1. Parsing code into AST using TypeScript Compiler API
+ * 2. Detecting repeated subtrees via structural hashing
+ * 3. Abstracting common structures via anti-unification
+ */
+export class PatternExtractor {
+    options;
+    parser;
+    finder;
+    antiUnifier;
+    constructor(options = {}) {
+        this.options = {
+            language: options.language ?? 'typescript',
+            minFrequency: options.minFrequency ?? 2,
+            maxDepth: options.maxDepth ?? 10,
+        };
+        this.parser = new TypeScriptParser({ maxDepth: this.options.maxDepth });
+        this.finder = new SubtreeFinder({
+            minFrequency: this.options.minFrequency,
+            maxDepth: this.options.maxDepth,
+        });
+        this.antiUnifier = new AntiUnifier();
+    }
+    /**
+     * Extract patterns from source code
+     * @param source - Source code string
+     * @returns Extracted patterns
+     */
+    async extract(source) {
+        // Phase 1: Parse to AST
+        const ast = this.parser.parse(source);
+        // Phase 2: Find repeated subtrees
+        const candidates = this.finder.find(ast);
+        // Phase 3: Convert to patterns
+        const patterns = candidates.map(c => this.createPattern(c));
+        return patterns;
+    }
+    /**
+     * Extract patterns from multiple source files
+     */
+    async extractFromMultiple(sources) {
+        // Parse all sources
+        const asts = sources.map(s => this.parser.parse(s));
+        // Find patterns in combined AST space
+        const allPatterns = [];
+        const patternMap = new Map();
+        for (const ast of asts) {
+            const candidates = this.finder.find(ast);
+            for (const candidate of candidates) {
+                const existing = patternMap.get(candidate.hash);
+                if (existing) {
+                    existing.frequency += candidate.frequency;
+                    existing.locations.push(...candidate.locations);
+                }
+                else {
+                    patternMap.set(candidate.hash, { ...candidate });
+                }
+            }
+        }
+        // Filter by global frequency
+        const filtered = Array.from(patternMap.values())
+            .filter(c => c.frequency >= this.options.minFrequency);
+        // Abstract similar patterns
+        const abstractedPatterns = this.abstractSimilarPatterns(filtered);
+        for (const candidate of abstractedPatterns) {
+            allPatterns.push(this.createPattern(candidate));
+        }
+        return allPatterns;
+    }
+    /**
+     * Abstract similar patterns using anti-unification
+     */
+    abstractSimilarPatterns(candidates) {
+        if (candidates.length < 2)
+            return candidates;
+        // Group by similar structure
+        const groups = [];
+        const processed = new Set();
+        for (let i = 0; i < candidates.length; i++) {
+            if (processed.has(i))
+                continue;
+            const group = [candidates[i]];
+            processed.add(i);
+            for (let j = i + 1; j < candidates.length; j++) {
+                if (processed.has(j))
+                    continue;
+                const similarity = this.antiUnifier.calculateSimilarity(candidates[i].ast, candidates[j].ast);
+                if (similarity > 0.7) {
+                    group.push(candidates[j]);
+                    processed.add(j);
+                }
+            }
+            groups.push(group);
+        }
+        // Abstract each group
+        const result = [];
+        for (const group of groups) {
+            if (group.length === 1) {
+                result.push(group[0]);
+            }
+            else {
+                // Anti-unify the group
+                const asts = group.map(c => c.ast);
+                const unified = this.antiUnifier.antiUnifyAll(asts);
+                result.push({
+                    ast: unified.pattern,
+                    hash: `abstract-${Date.now()}`,
+                    frequency: group.reduce((sum, c) => sum + c.frequency, 0),
+                    depth: Math.max(...group.map(c => c.depth)),
+                    locations: group.flatMap(c => c.locations),
+                });
+            }
+        }
+        return result;
+    }
+    /**
+     * Create pattern from candidate
+     */
+    createPattern(candidate) {
+        const id = `PAT-${Date.now()}-${Math.random().toString(36).slice(2, 8)}`;
+        const now = new Date().toISOString();
+        // Extract holes from AST if it was anti-unified
+        const holes = this.extractHoles(candidate.ast);
+        return {
+            id,
+            name: this.generatePatternName(candidate.ast),
+            language: this.options.language,
+            ast: candidate.ast,
+            holes,
+            frequency: candidate.frequency,
+            createdAt: now,
+            updatedAt: now,
+        };
+    }
+    /**
+     * Generate a descriptive pattern name
+     */
+    generatePatternName(ast) {
+        const typeNames = {
+            'FunctionDeclaration': 'function_pattern',
+            'MethodDeclaration': 'method_pattern',
+            'ClassDeclaration': 'class_pattern',
+            'IfStatement': 'conditional_pattern',
+            'ForStatement': 'loop_pattern',
+            'ForOfStatement': 'iteration_pattern',
+            'ForInStatement': 'enumeration_pattern',
+            'WhileStatement': 'while_loop_pattern',
+            'TryStatement': 'error_handling_pattern',
+            'SwitchStatement': 'switch_pattern',
+            'ArrowFunction': 'lambda_pattern',
+            'CallExpression': 'call_pattern',
+            'VariableDeclaration': 'variable_pattern',
+            'PropertyDeclaration': 'property_pattern',
+        };
+        const baseName = typeNames[ast.type] ?? `${ast.type.toLowerCase()}_pattern`;
+        return `${baseName}_${Math.random().toString(36).slice(2, 6)}`;
+    }
+    /**
+     * Extract holes from anti-unified AST
+     */
+    extractHoles(ast) {
+        const holes = [];
+        const traverse = (node) => {
+            if (node.type === 'Hole' && node.value) {
+                holes.push({
+                    id: node.value,
+                    type: 'any',
+                });
+            }
+            for (const child of node.children) {
+                traverse(child);
+            }
+        };
+        traverse(ast);
+        return holes;
+    }
+    /**
+     * Get parser for direct AST access
+     */
+    getParser() {
+        return this.parser;
+    }
+    /**
+     * Get anti-unifier for direct access
+     */
+    getAntiUnifier() {
+        return this.antiUnifier;
+    }
+}
+//# sourceMappingURL=pattern-extractor.js.map

package/dist/extractor/pattern-extractor.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"pattern-extractor.js","sourceRoot":"","sources":["../../src/extractor/pattern-extractor.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,OAAO,EAAE,gBAAgB,EAAE,MAAM,wBAAwB,CAAC;AAC1D,OAAO,EAAE,aAAa,EAAyB,MAAM,qBAAqB,CAAC;AAC3E,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAEhD;;;;;;;;GAQG;AACH,MAAM,OAAO,gBAAgB;IACnB,OAAO,CAAoB;IAC3B,MAAM,CAAmB;IACzB,MAAM,CAAgB;IACtB,WAAW,CAAc;IAEjC,YAAY,UAAsC,EAAE;QAClD,IAAI,CAAC,OAAO,GAAG;YACb,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,YAAY;YAC1C,YAAY,EAAE,OAAO,CAAC,YAAY,IAAI,CAAC;YACvC,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,EAAE;SACjC,CAAC;QAEF,IAAI,CAAC,MAAM,GAAG,IAAI,gBAAgB,CAAC,EAAE,QAAQ,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC,CAAC;QACxE,IAAI,CAAC,MAAM,GAAG,IAAI,aAAa,CAAC;YAC9B,YAAY,EAAE,IAAI,CAAC,OAAO,CAAC,YAAY;YACvC,QAAQ,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ;SAChC,CAAC,CAAC;QACH,IAAI,CAAC,WAAW,GAAG,IAAI,WAAW,EAAE,CAAC;IACvC,CAAC;IAED;;;;OAIG;IACH,KAAK,CAAC,OAAO,CAAC,MAAc;QAC1B,wBAAwB;QACxB,MAAM,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QAEtC,kCAAkC;QAClC,MAAM,UAAU,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAEzC,+BAA+B;QAC/B,MAAM,QAAQ,GAAG,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC,CAAC;QAE5D,OAAO,QAAQ,CAAC;IAClB,CAAC;IAED;;OAEG;IACH,KAAK,CAAC,mBAAmB,CAAC,OAAiB;QACzC,oBAAoB;QACpB,MAAM,IAAI,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;QAEpD,sCAAsC;QACtC,MAAM,WAAW,GAAc,EAAE,CAAC;QAClC,MAAM,UAAU,GAAG,IAAI,GAAG,EAA4B,CAAC;QAEvD,KAAK,MAAM,GAAG,IAAI,IAAI,EAAE,CAAC;YACvB,MAAM,UAAU,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACzC,KAAK,MAAM,SAAS,IAAI,UAAU,EAAE,CAAC;gBACnC,MAAM,QAAQ,GAAG,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;gBAChD,IAAI,QAAQ,EAAE,CAAC;oBACb,QAAQ,CAAC,SAAS,IAAI,SAAS,CAAC,SAAS,CAAC;oBAC1C,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,SAAS,CAAC,SAAS,CAAC,CAAC;gBAClD,CAAC;qBAAM,CAAC;oBACN,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,EAAE,EAAE,GAAG,SAAS,EAAE,CAAC,CAAC;gBACnD,CAAC;YACH,CAAC;QACH,CAAC;QAED,6BAA6B;QAC7B,MAAM,QAAQ,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC;aAC7C,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC;QAEzD,4BAA4B;QAC5B,MAAM,kBAAkB,GAAG,IAAI,CAAC,uBAAuB,CAAC,QAAQ,CAAC,CAAC;QAElE,KAAK,MAAM,SAAS,IAAI,kBAAkB,EAAE,CAAC;YAC3C,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,SAAS,CAAC,CAAC,CAAC;QAClD,CAAC;QAED,OAAO,WAAW,CAAC;IACrB,CAAC;IAED;;OAEG;IACK,uBAAuB,CAAC,UAA8B;QAC5D,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC;YAAE,OAAO,UAAU,CAAC;QAE7C,6BAA6B;QAC7B,MAAM,MAAM,GAAyB,EAAE,CAAC;QACxC,MAAM,SAAS,GAAG,IAAI,GAAG,EAAU,CAAC;QAEpC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3C,IAAI,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC;gBAAE,SAAS;YAE/B,MAAM,KAAK,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;YAC9B,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;YAEjB,KAAK,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC/C,IAAI,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC;oBAAE,SAAS;gBAE/B,MAAM,UAAU,GAAG,IAAI,CAAC,WAAW,CAAC,mBAAmB,CACrD,UAAU,CAAC,CAAC,CAAC,CAAC,GAAG,EACjB,UAAU,CAAC,CAAC,CAAC,CAAC,GAAG,CAClB,CAAC;gBAEF,IAAI,UAAU,GAAG,GAAG,EAAE,CAAC;oBACrB,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;oBAC1B,SAAS,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;gBACnB,CAAC;YACH,CAAC;YAED,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACrB,CAAC;QAED,sBAAsB;QACtB,MAAM,MAAM,GAAuB,EAAE,CAAC;QAEtC,KAAK,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;YAC3B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;gBACvB,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;YACxB,CAAC;iBAAM,CAAC;gBACN,uBAAuB;gBACvB,MAAM,IAAI,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;gBACnC,MAAM,OAAO,GAAG,IAAI,CAAC,WAAW,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;gBAEpD,MAAM,CAAC,IAAI,CAAC;oBACV,GAAG,EAAE,OAAO,CAAC,OAAO;oBACpB,IAAI,EAAE,YAAY,IAAI,CAAC,GAAG,EAAE,EAAE;oBAC9B,SAAS,EAAE,KAAK,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,GAAG,CAAC,CAAC,SAAS,EAAE,CAAC,CAAC;oBACzD,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;oBAC3C,SAAS,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC;iBAC3C,CAAC,CAAC;YACL,CAAC;QACH,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;OAEG;IACK,aAAa,CAAC,SAA2B;QAC/C,MAAM,EAAE,GAAG,OAAO,IAAI,CAAC,GAAG,EAAE,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC;QACzE,MAAM,GAAG,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;QAErC,gDAAgD;QAChD,MAAM,KAAK,GAAG,IAAI,CAAC,YAAY,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC;QAE/C,OAAO;YACL,EAAE;YACF,IAAI,EAAE,IAAI,CAAC,mBAAmB,CAAC,SAAS,CAAC,GAAG,CAAC;YAC7C,QAAQ,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ;YAC/B,GAAG,EAAE,SAAS,CAAC,GAAG;YAClB,KAAK;YACL,SAAS,EAAE,SAAS,CAAC,SAAS;YAC9B,SAAS,EAAE,GAAG;YACd,SAAS,EAAE,GAAG;SACf,CAAC;IACJ,CAAC;IAED;;OAEG;IACK,mBAAmB,CAAC,GAAY;QACtC,MAAM,SAAS,GAA2B;YACxC,qBAAqB,EAAE,kBAAkB;YACzC,mBAAmB,EAAE,gBAAgB;YACrC,kBAAkB,EAAE,eAAe;YACnC,aAAa,EAAE,qBAAqB;YACpC,cAAc,EAAE,cAAc;YAC9B,gBAAgB,EAAE,mBAAmB;YACrC,gBAAgB,EAAE,qBAAqB;YACvC,gBAAgB,EAAE,oBAAoB;YACtC,cAAc,EAAE,wBAAwB;YACxC,iBAAiB,EAAE,gBAAgB;YACnC,eAAe,EAAE,gBAAgB;YACjC,gBAAgB,EAAE,cAAc;YAChC,qBAAqB,EAAE,kBAAkB;YACzC,qBAAqB,EAAE,kBAAkB;SAC1C,CAAC;QAEF,MAAM,QAAQ,GAAG,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC,WAAW,EAAE,UAAU,CAAC;QAC5E,OAAO,GAAG,QAAQ,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC;IACjE,CAAC;IAED;;OAEG;IACK,YAAY,CAAC,GAAY;QAC/B,MAAM,KAAK,GAAqB,EAAE,CAAC;QAEnC,MAAM,QAAQ,GAAG,CAAC,IAAa,EAAQ,EAAE;YACvC,IAAI,IAAI,CAAC,IAAI,KAAK,MAAM,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;gBACvC,KAAK,CAAC,IAAI,CAAC;oBACT,EAAE,EAAE,IAAI,CAAC,KAAK;oBACd,IAAI,EAAE,KAAK;iBACZ,CAAC,CAAC;YACL,CAAC;YACD,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;gBAClC,QAAQ,CAAC,KAAK,CAAC,CAAC;YAClB,CAAC;QACH,CAAC,CAAC;QAEF,QAAQ,CAAC,GAAG,CAAC,CAAC;QACd,OAAO,KAAK,CAAC;IACf,CAAC;IAED;;OAEG;IACH,SAAS;QACP,OAAO,IAAI,CAAC,MAAM,CAAC;IACrB,CAAC;IAED;;OAEG;IACH,cAAc;QACZ,OAAO,IAAI,CAAC,WAAW,CAAC;IAC1B,CAAC;CACF"}

package/dist/extractor/subtree-finder.d.ts

@@ -0,0 +1,81 @@
+/**
+ * @fileoverview Repeated subtree detection in AST
+ * @traceability TSK-PATTERN-001, REQ-PATTERN-001-F002
+ */
+import type { ASTNode } from '../types.js';
+/**
+ * Subtree candidate with frequency information
+ */
+export interface SubtreeCandidate {
+    ast: ASTNode;
+    hash: string;
+    frequency: number;
+    depth: number;
+    locations: SubtreeLocation[];
+}
+/**
+ * Location of a subtree in source
+ */
+export interface SubtreeLocation {
+    startRow: number;
+    startColumn: number;
+    endRow: number;
+    endColumn: number;
+}
+/**
+ * Subtree finder options
+ */
+export interface SubtreeFinderOptions {
+    /** Minimum frequency to be considered a pattern */
+    minFrequency: number;
+    /** Minimum subtree depth to consider */
+    minDepth: number;
+    /** Maximum subtree depth to consider */
+    maxDepth: number;
+    /** Node types to ignore */
+    ignoreTypes: string[];
+}
+/**
+ * Finds repeated subtrees in an AST
+ *
+ * @description
+ * Uses structural hashing to identify repeated code patterns.
+ * Implements anti-unification to find common structures.
+ */
+export declare class SubtreeFinder {
+    private options;
+    constructor(options?: Partial<SubtreeFinderOptions>);
+    /**
+     * Find all repeated subtrees in AST
+     */
+    find(ast: ASTNode): SubtreeCandidate[];
+    /**
+     * Traverse AST and collect subtree hashes
+     */
+    private traverse;
+    /**
+     * Hash a leaf node
+     */
+    private hashLeaf;
+    /**
+     * Hash a subtree
+     */
+    private hashSubtree;
+    /**
+     * Remove subsumed patterns (patterns contained in larger patterns)
+     */
+    private removeSubsumed;
+    /**
+     * Check if candidate is subsumed by any existing pattern
+     */
+    private isSubsumedByAny;
+    /**
+     * Check if AST contains a subtree with given hash
+     */
+    private containsHash;
+    /**
+     * Compute hash for AST node (for containment check)
+     */
+    private computeHash;
+}
+//# sourceMappingURL=subtree-finder.d.ts.map

package/dist/extractor/subtree-finder.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"subtree-finder.d.ts","sourceRoot":"","sources":["../../src/extractor/subtree-finder.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,aAAa,CAAC;AAE3C;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAC/B,GAAG,EAAE,OAAO,CAAC;IACb,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,MAAM,CAAC;IAClB,KAAK,EAAE,MAAM,CAAC;IACd,SAAS,EAAE,eAAe,EAAE,CAAC;CAC9B;AAED;;GAEG;AACH,MAAM,WAAW,eAAe;IAC9B,QAAQ,EAAE,MAAM,CAAC;IACjB,WAAW,EAAE,MAAM,CAAC;IACpB,MAAM,EAAE,MAAM,CAAC;IACf,SAAS,EAAE,MAAM,CAAC;CACnB;AAED;;GAEG;AACH,MAAM,WAAW,oBAAoB;IACnC,mDAAmD;IACnD,YAAY,EAAE,MAAM,CAAC;IACrB,wCAAwC;IACxC,QAAQ,EAAE,MAAM,CAAC;IACjB,wCAAwC;IACxC,QAAQ,EAAE,MAAM,CAAC;IACjB,2BAA2B;IAC3B,WAAW,EAAE,MAAM,EAAE,CAAC;CACvB;AAED;;;;;;GAMG;AACH,qBAAa,aAAa;IACxB,OAAO,CAAC,OAAO,CAAuB;gBAE1B,OAAO,GAAE,OAAO,CAAC,oBAAoB,CAAM;IAgBvD;;OAEG;IACH,IAAI,CAAC,GAAG,EAAE,OAAO,GAAG,gBAAgB,EAAE;IAgBtC;;OAEG;IACH,OAAO,CAAC,QAAQ;IAsDhB;;OAEG;IACH,OAAO,CAAC,QAAQ;IAKhB;;OAEG;IACH,OAAO,CAAC,WAAW;IAKnB;;OAEG;IACH,OAAO,CAAC,cAAc;IAmBtB;;OAEG;IACH,OAAO,CAAC,eAAe;IAgBvB;;OAEG;IACH,OAAO,CAAC,YAAY;IAYpB;;OAEG;IACH,OAAO,CAAC,WAAW;CAIpB"}

package/dist/extractor/subtree-finder.js

@@ -0,0 +1,155 @@
+/**
+ * @fileoverview Repeated subtree detection in AST
+ * @traceability TSK-PATTERN-001, REQ-PATTERN-001-F002
+ */
+/**
+ * Finds repeated subtrees in an AST
+ *
+ * @description
+ * Uses structural hashing to identify repeated code patterns.
+ * Implements anti-unification to find common structures.
+ */
+export class SubtreeFinder {
+    options;
+    constructor(options = {}) {
+        this.options = {
+            minFrequency: options.minFrequency ?? 2,
+            minDepth: options.minDepth ?? 2,
+            maxDepth: options.maxDepth ?? 10,
+            ignoreTypes: options.ignoreTypes ?? [
+                'EndOfFileToken',
+                'SemicolonToken',
+                'OpenBraceToken',
+                'CloseBraceToken',
+                'OpenParenToken',
+                'CloseParenToken',
+            ],
+        };
+    }
+    /**
+     * Find all repeated subtrees in AST
+     */
+    find(ast) {
+        const subtreeMap = new Map();
+        this.traverse(ast, subtreeMap, 0);
+        // Filter by minimum frequency and depth
+        const candidates = Array.from(subtreeMap.values())
+            .filter(c => c.frequency >= this.options.minFrequency &&
+            c.depth >= this.options.minDepth);
+        // Remove subsumed patterns (smaller patterns contained in larger ones)
+        return this.removeSubsumed(candidates);
+    }
+    /**
+     * Traverse AST and collect subtree hashes
+     */
+    traverse(node, subtreeMap, depth) {
+        // Skip ignored node types
+        if (this.options.ignoreTypes.includes(node.type)) {
+            return { hash: '', depth: 0 };
+        }
+        // Skip if max depth exceeded
+        if (depth > this.options.maxDepth) {
+            return { hash: this.hashLeaf(node), depth: 1 };
+        }
+        // Calculate children hashes
+        const childResults = node.children.map(child => this.traverse(child, subtreeMap, depth + 1));
+        // Build hash for this subtree
+        const hash = this.hashSubtree(node, childResults.map(r => r.hash));
+        const subtreeDepth = 1 + Math.max(0, ...childResults.map(r => r.depth));
+        // Skip trivial subtrees
+        if (subtreeDepth < this.options.minDepth) {
+            return { hash, depth: subtreeDepth };
+        }
+        // Register or update in map
+        const location = {
+            startRow: node.startPosition.row,
+            startColumn: node.startPosition.column,
+            endRow: node.endPosition.row,
+            endColumn: node.endPosition.column,
+        };
+        const existing = subtreeMap.get(hash);
+        if (existing) {
+            existing.frequency++;
+            existing.locations.push(location);
+        }
+        else {
+            subtreeMap.set(hash, {
+                ast: node,
+                hash,
+                frequency: 1,
+                depth: subtreeDepth,
+                locations: [location],
+            });
+        }
+        return { hash, depth: subtreeDepth };
+    }
+    /**
+     * Hash a leaf node
+     */
+    hashLeaf(node) {
+        // Include type and optionally value for identifiers
+        return `L:${node.type}`;
+    }
+    /**
+     * Hash a subtree
+     */
+    hashSubtree(node, childHashes) {
+        const childPart = childHashes.filter(h => h).join(',');
+        return `N:${node.type}[${childPart}]`;
+    }
+    /**
+     * Remove subsumed patterns (patterns contained in larger patterns)
+     */
+    removeSubsumed(candidates) {
+        // Sort by depth (largest first)
+        const sorted = [...candidates].sort((a, b) => b.depth - a.depth);
+        const result = [];
+        const seenHashes = new Set();
+        for (const candidate of sorted) {
+            // Check if this subtree's locations are all covered by a larger pattern
+            const isSubsumed = this.isSubsumedByAny(candidate, result);
+            if (!isSubsumed && !seenHashes.has(candidate.hash)) {
+                result.push(candidate);
+                seenHashes.add(candidate.hash);
+            }
+        }
+        return result;
+    }
+    /**
+     * Check if candidate is subsumed by any existing pattern
+     */
+    isSubsumedByAny(candidate, existing) {
+        for (const pattern of existing) {
+            // Skip if same or smaller depth
+            if (pattern.depth <= candidate.depth)
+                continue;
+            // Check if candidate's hash appears in pattern's AST
+            if (this.containsHash(pattern.ast, candidate.hash)) {
+                return true;
+            }
+        }
+        return false;
+    }
+    /**
+     * Check if AST contains a subtree with given hash
+     */
+    containsHash(ast, targetHash) {
+        const currentHash = this.computeHash(ast);
+        if (currentHash === targetHash)
+            return true;
+        for (const child of ast.children) {
+            if (this.containsHash(child, targetHash)) {
+                return true;
+            }
+        }
+        return false;
+    }
+    /**
+     * Compute hash for AST node (for containment check)
+     */
+    computeHash(node) {
+        const childHashes = node.children.map(c => this.computeHash(c));
+        return this.hashSubtree(node, childHashes);
+    }
+}
+//# sourceMappingURL=subtree-finder.js.map

package/dist/extractor/subtree-finder.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"subtree-finder.js","sourceRoot":"","sources":["../../src/extractor/subtree-finder.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAuCH;;;;;;GAMG;AACH,MAAM,OAAO,aAAa;IAChB,OAAO,CAAuB;IAEtC,YAAY,UAAyC,EAAE;QACrD,IAAI,CAAC,OAAO,GAAG;YACb,YAAY,EAAE,OAAO,CAAC,YAAY,IAAI,CAAC;YACvC,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,CAAC;YAC/B,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,EAAE;YAChC,WAAW,EAAE,OAAO,CAAC,WAAW,IAAI;gBAClC,gBAAgB;gBAChB,gBAAgB;gBAChB,gBAAgB;gBAChB,iBAAiB;gBACjB,gBAAgB;gBAChB,iBAAiB;aAClB;SACF,CAAC;IACJ,CAAC;IAED;;OAEG;IACH,IAAI,CAAC,GAAY;QACf,MAAM,UAAU,GAAG,IAAI,GAAG,EAA4B,CAAC;QAEvD,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC;QAElC,wCAAwC;QACxC,MAAM,UAAU,GAAG,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC;aAC/C,MAAM,CAAC,CAAC,CAAC,EAAE,CACV,CAAC,CAAC,SAAS,IAAI,IAAI,CAAC,OAAO,CAAC,YAAY;YACxC,CAAC,CAAC,KAAK,IAAI,IAAI,CAAC,OAAO,CAAC,QAAQ,CACjC,CAAC;QAEJ,uEAAuE;QACvE,OAAO,IAAI,CAAC,cAAc,CAAC,UAAU,CAAC,CAAC;IACzC,CAAC;IAED;;OAEG;IACK,QAAQ,CACd,IAAa,EACb,UAAyC,EACzC,KAAa;QAEb,0BAA0B;QAC1B,IAAI,IAAI,CAAC,OAAO,CAAC,WAAW,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YACjD,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC;QAChC,CAAC;QAED,6BAA6B;QAC7B,IAAI,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC;YAClC,OAAO,EAAE,IAAI,EAAE,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,CAAC;QACjD,CAAC;QAED,4BAA4B;QAC5B,MAAM,YAAY,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAC7C,IAAI,CAAC,QAAQ,CAAC,KAAK,EAAE,UAAU,EAAE,KAAK,GAAG,CAAC,CAAC,CAC5C,CAAC;QAEF,8BAA8B;QAC9B,MAAM,IAAI,GAAG,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;QACnE,MAAM,YAAY,GAAG,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,GAAG,YAAY,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC;QAExE,wBAAwB;QACxB,IAAI,YAAY,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC;YACzC,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC;QACvC,CAAC;QAED,4BAA4B;QAC5B,MAAM,QAAQ,GAAoB;YAChC,QAAQ,EAAE,IAAI,CAAC,aAAa,CAAC,GAAG;YAChC,WAAW,EAAE,IAAI,CAAC,aAAa,CAAC,MAAM;YACtC,MAAM,EAAE,IAAI,CAAC,WAAW,CAAC,GAAG;YAC5B,SAAS,EAAE,IAAI,CAAC,WAAW,CAAC,MAAM;SACnC,CAAC;QAEF,MAAM,QAAQ,GAAG,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QACtC,IAAI,QAAQ,EAAE,CAAC;YACb,QAAQ,CAAC,SAAS,EAAE,CAAC;YACrB,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QACpC,CAAC;aAAM,CAAC;YACN,UAAU,CAAC,GAAG,CAAC,IAAI,EAAE;gBACnB,GAAG,EAAE,IAAI;gBACT,IAAI;gBACJ,SAAS,EAAE,CAAC;gBACZ,KAAK,EAAE,YAAY;gBACnB,SAAS,EAAE,CAAC,QAAQ,CAAC;aACtB,CAAC,CAAC;QACL,CAAC;QAED,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC;IACvC,CAAC;IAED;;OAEG;IACK,QAAQ,CAAC,IAAa;QAC5B,oDAAoD;QACpD,OAAO,KAAK,IAAI,CAAC,IAAI,EAAE,CAAC;IAC1B,CAAC;IAED;;OAEG;IACK,WAAW,CAAC,IAAa,EAAE,WAAqB;QACtD,MAAM,SAAS,GAAG,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QACvD,OAAO,KAAK,IAAI,CAAC,IAAI,IAAI,SAAS,GAAG,CAAC;IACxC,CAAC;IAED;;OAEG;IACK,cAAc,CAAC,UAA8B;QACnD,gCAAgC;QAChC,MAAM,MAAM,GAAG,CAAC,GAAG,UAAU,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;QACjE,MAAM,MAAM,GAAuB,EAAE,CAAC;QACtC,MAAM,UAAU,GAAG,IAAI,GAAG,EAAU,CAAC;QAErC,KAAK,MAAM,SAAS,IAAI,MAAM,EAAE,CAAC;YAC/B,wEAAwE;YACxE,MAAM,UAAU,GAAG,IAAI,CAAC,eAAe,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC;YAE3D,IAAI,CAAC,UAAU,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;gBACnD,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;gBACvB,UAAU,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC;YACjC,CAAC;QACH,CAAC;QAED,OAAO,MAAM,CAAC;IAChB,CAAC;IAED;;OAEG;IACK,eAAe,CACrB,SAA2B,EAC3B,QAA4B;QAE5B,KAAK,MAAM,OAAO,IAAI,QAAQ,EAAE,CAAC;YAC/B,gCAAgC;YAChC,IAAI,OAAO,CAAC,KAAK,IAAI,SAAS,CAAC,KAAK;gBAAE,SAAS;YAE/C,qDAAqD;YACrD,IAAI,IAAI,CAAC,YAAY,CAAC,OAAO,CAAC,GAAG,EAAE,SAAS,CAAC,IAAI,CAAC,EAAE,CAAC;gBACnD,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;QACD,OAAO,KAAK,CAAC;IACf,CAAC;IAED;;OAEG;IACK,YAAY,CAAC,GAAY,EAAE,UAAkB;QACnD,MAAM,WAAW,GAAG,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;QAC1C,IAAI,WAAW,KAAK,UAAU;YAAE,OAAO,IAAI,CAAC;QAE5C,KAAK,MAAM,KAAK,IAAI,GAAG,CAAC,QAAQ,EAAE,CAAC;YACjC,IAAI,IAAI,CAAC,YAAY,CAAC,KAAK,EAAE,UAAU,CAAC,EAAE,CAAC;gBACzC,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;QACD,OAAO,KAAK,CAAC;IACf,CAAC;IAED;;OAEG;IACK,WAAW,CAAC,IAAa;QAC/B,MAAM,WAAW,GAAG,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC;QAChE,OAAO,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,WAAW,CAAC,CAAC;IAC7C,CAAC;CACF"}

package/dist/extractor/typescript-parser.d.ts

@@ -0,0 +1,91 @@
+/**
+ * @fileoverview TypeScript AST parser using TypeScript Compiler API
+ * @traceability TSK-PATTERN-001, REQ-PATTERN-001-F001
+ */
+import type { ASTNode } from '../types.js';
+/**
+ * TypeScript parser options
+ */
+export interface TypeScriptParserOptions {
+    /** Include position information */
+    includePositions: boolean;
+    /** Include node values (identifiers, literals) */
+    includeValues: boolean;
+    /** Maximum depth to parse */
+    maxDepth: number;
+}
+/**
+ * TypeScript AST parser
+ *
+ * @description
+ * Parses TypeScript/JavaScript source code into a normalized AST
+ * using the TypeScript Compiler API.
+ */
+export declare class TypeScriptParser {
+    private options;
+    constructor(options?: Partial<TypeScriptParserOptions>);
+    /**
+     * Parse TypeScript/JavaScript source code to AST
+     */
+    parse(source: string, fileName?: string): ASTNode;
+    /**
+     * Convert TypeScript AST node to normalized ASTNode
+     */
+    private convertNode;
+    /**
+     * Get human-readable node type name
+     */
+    private getNodeTypeName;
+    /**
+     * Extract value from node (for identifiers, literals)
+     */
+    private extractValue;
+    /**
+     * Convert position offset to row/column
+     */
+    private getPosition;
+    /**
+     * Get all function declarations in source
+     */
+    getFunctions(source: string): FunctionInfo[];
+    /**
+     * Get all class declarations in source
+     */
+    getClasses(source: string): ClassInfo[];
+    /**
+     * Get imports from source
+     */
+    getImports(source: string): ImportInfo[];
+}
+/**
+ * Function information extracted from AST
+ */
+export interface FunctionInfo {
+    name: string;
+    parameters: {
+        name: string;
+        type: string;
+    }[];
+    returnType: string;
+    startLine: number;
+    endLine: number;
+}
+/**
+ * Class information extracted from AST
+ */
+export interface ClassInfo {
+    name: string;
+    methods: string[];
+    properties: string[];
+    startLine: number;
+    endLine: number;
+}
+/**
+ * Import information extracted from AST
+ */
+export interface ImportInfo {
+    module: string;
+    namedImports: string[];
+    defaultImport?: string;
+}
+//# sourceMappingURL=typescript-parser.d.ts.map

package/dist/extractor/typescript-parser.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"typescript-parser.d.ts","sourceRoot":"","sources":["../../src/extractor/typescript-parser.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,OAAO,KAAK,EAAE,OAAO,EAAY,MAAM,aAAa,CAAC;AAErD;;GAEG;AACH,MAAM,WAAW,uBAAuB;IACtC,mCAAmC;IACnC,gBAAgB,EAAE,OAAO,CAAC;IAC1B,kDAAkD;IAClD,aAAa,EAAE,OAAO,CAAC;IACvB,6BAA6B;IAC7B,QAAQ,EAAE,MAAM,CAAC;CAClB;AAED;;;;;;GAMG;AACH,qBAAa,gBAAgB;IAC3B,OAAO,CAAC,OAAO,CAA0B;gBAE7B,OAAO,GAAE,OAAO,CAAC,uBAAuB,CAAM;IAQ1D;;OAEG;IACH,KAAK,CAAC,MAAM,EAAE,MAAM,EAAE,QAAQ,SAAc,GAAG,OAAO;IAYtD;;OAEG;IACH,OAAO,CAAC,WAAW;IA2BnB;;OAEG;IACH,OAAO,CAAC,eAAe;IAIvB;;OAEG;IACH,OAAO,CAAC,YAAY;IAgBpB;;OAEG;IACH,OAAO,CAAC,WAAW;IAkBnB;;OAEG;IACH,YAAY,CAAC,MAAM,EAAE,MAAM,GAAG,YAAY,EAAE;IA+D5C;;OAEG;IACH,UAAU,CAAC,MAAM,EAAE,MAAM,GAAG,SAAS,EAAE;IAyCvC;;OAEG;IACH,UAAU,CAAC,MAAM,EAAE,MAAM,GAAG,UAAU,EAAE;CA6CzC;AAED;;GAEG;AACH,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,UAAU,EAAE;QAAE,IAAI,EAAE,MAAM,CAAC;QAAC,IAAI,EAAE,MAAM,CAAA;KAAE,EAAE,CAAC;IAC7C,UAAU,EAAE,MAAM,CAAC;IACnB,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;CACjB;AAED;;GAEG;AACH,MAAM,WAAW,SAAS;IACxB,IAAI,EAAE,MAAM,CAAC;IACb,OAAO,EAAE,MAAM,EAAE,CAAC;IAClB,UAAU,EAAE,MAAM,EAAE,CAAC;IACrB,SAAS,EAAE,MAAM,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;CACjB;AAED;;GAEG;AACH,MAAM,WAAW,UAAU;IACzB,MAAM,EAAE,MAAM,CAAC;IACf,YAAY,EAAE,MAAM,EAAE,CAAC;IACvB,aAAa,CAAC,EAAE,MAAM,CAAC;CACxB"}