driftdetect-core 0.4.1 → 0.4.3

package/dist/boundaries/boundary-scanner.js

@@ -0,0 +1,801 @@
+/**
+ * Boundary Scanner - Detects data access patterns in source code
+ *
+ * Uses a two-phase approach following Drift's philosophy:
+ * 1. LEARN: First pass discovers patterns from YOUR codebase
+ * 2. DETECT: Second pass uses learned patterns, regex as fallback
+ *
+ * This ensures we capture how YOUR code accesses data, not hardcoded assumptions.
+ */
+import * as fs from 'node:fs/promises';
+import * as path from 'node:path';
+import { minimatch } from 'minimatch';
+import { BoundaryStore, createBoundaryStore } from './boundary-store.js';
+import { DataAccessLearner, createDataAccessLearner } from './data-access-learner.js';
+// ============================================================================
+// Language Detection
+// ============================================================================
+function getLanguage(filePath) {
+    const ext = path.extname(filePath).toLowerCase();
+    switch (ext) {
+        case '.py':
+        case '.pyw':
+            return 'python';
+        case '.ts':
+        case '.tsx':
+            return 'typescript';
+        case '.js':
+        case '.jsx':
+        case '.mjs':
+        case '.cjs':
+            return 'javascript';
+        case '.cs':
+            return 'csharp';
+        case '.php':
+            return 'php';
+        case '.java':
+            return 'java';
+        default:
+            return null;
+    }
+}
+function isDataAccessFile(filePath, content) {
+    // Skip test files for boundary detection
+    if (filePath.includes('.test.') || filePath.includes('.spec.') ||
+        filePath.includes('__tests__') || filePath.includes('/tests/')) {
+        return false;
+    }
+    // ORM patterns
+    const ormPatterns = [
+        'DbSet', 'DbContext', 'Entity', 'Table', 'Column',
+        'models.Model', 'CharField', 'ForeignKey', 'ManyToMany',
+        'declarative_base', 'relationship',
+        'prisma.', '@prisma/client',
+        '@Entity', '@Column', 'getRepository',
+        'sequelize.define', 'DataTypes',
+        'drizzle-orm',
+        'SELECT', 'INSERT', 'UPDATE', 'DELETE', 'FROM',
+        'execute', 'query', 'rawQuery',
+        'supabase', '.from(', '.select(', '.insert(', '.update(',
+    ];
+    return ormPatterns.some(pattern => content.includes(pattern));
+}
+// ============================================================================
+// Sensitive Field Detection
+// ============================================================================
+const SENSITIVE_PATTERNS = {
+    pii: [
+        /\bssn\b/i, /\bsocial_security\b/i, /\bdate_of_birth\b/i,
+        /\bdob\b/i, /\baddress\b/i, /\bphone\b/i, /\bphone_number\b/i,
+        /\bemail\b/i, /\bfull_name\b/i, /\bfirst_name\b/i, /\blast_name\b/i,
+    ],
+    credentials: [
+        /\bpassword\b/i, /\bpassword_hash\b/i, /\bsecret\b/i, /\btoken\b/i,
+        /\bapi_key\b/i, /\bprivate_key\b/i, /\bhash\b/i, /\bsalt\b/i,
+        /\brefresh_token\b/i, /\baccess_token\b/i,
+    ],
+    financial: [
+        /\bcredit_card\b/i, /\bcard_number\b/i, /\bcvv\b/i,
+        /\bbank_account\b/i, /\bsalary\b/i, /\bincome\b/i,
+        /\bpayment\b/i, /\bbalance\b/i,
+    ],
+    health: [
+        /\bdiagnosis\b/i, /\bprescription\b/i, /\bmedical\b/i, /\bhealth\b/i,
+    ],
+};
+function detectSensitiveFields(content, file) {
+    const fields = [];
+    const lines = content.split('\n');
+    for (let i = 0; i < lines.length; i++) {
+        const line = lines[i];
+        if (!line)
+            continue;
+        // Skip comments
+        const trimmed = line.trim();
+        if (trimmed.startsWith('//') || trimmed.startsWith('#') ||
+            trimmed.startsWith('*') || trimmed.startsWith('/*')) {
+            continue;
+        }
+        for (const [type, patterns] of Object.entries(SENSITIVE_PATTERNS)) {
+            for (const pattern of patterns) {
+                const match = line.match(pattern);
+                if (match) {
+                    // Try to extract table name from context
+                    let table = null;
+                    // Look for class/model name in surrounding lines
+                    for (let j = Math.max(0, i - 10); j < i; j++) {
+                        const prevLine = lines[j];
+                        if (!prevLine)
+                            continue;
+                        const classMatch = prevLine.match(/class\s+(\w+)/);
+                        const modelMatch = prevLine.match(/model\s+(\w+)/);
+                        const tableMatch = prevLine.match(/Table\s*\(\s*["'](\w+)["']/);
+                        const fromMatch = prevLine.match(/\.from\s*\(\s*["'](\w+)["']/);
+                        if (classMatch?.[1])
+                            table = classMatch[1];
+                        if (modelMatch?.[1])
+                            table = modelMatch[1];
+                        if (tableMatch?.[1])
+                            table = tableMatch[1];
+                        if (fromMatch?.[1])
+                            table = fromMatch[1];
+                    }
+                    fields.push({
+                        field: match[0],
+                        table,
+                        sensitivityType: type,
+                        file,
+                        line: i + 1,
+                        confidence: 0.8,
+                    });
+                    break;
+                }
+            }
+        }
+    }
+    return fields;
+}
+// ============================================================================
+// ORM Model Detection
+// ============================================================================
+function detectORMModels(content, file) {
+    const models = [];
+    // Detect EF Core DbSet patterns
+    const dbSetPattern = /DbSet<(\w+)>\s+(\w+)/g;
+    let match;
+    while ((match = dbSetPattern.exec(content)) !== null) {
+        const modelName = match[1];
+        const propName = match[2];
+        if (modelName && propName) {
+            const lineNum = content.substring(0, match.index).split('\n').length;
+            models.push({
+                name: modelName,
+                tableName: propName.toLowerCase(),
+                fields: [],
+                file,
+                line: lineNum,
+                framework: 'efcore',
+                confidence: 0.9,
+            });
+        }
+    }
+    // Detect Django models
+    const djangoModelPattern = /class\s+(\w+)\s*\([^)]*models\.Model[^)]*\)/g;
+    while ((match = djangoModelPattern.exec(content)) !== null) {
+        const modelName = match[1];
+        if (modelName) {
+            const lineNum = content.substring(0, match.index).split('\n').length;
+            models.push({
+                name: modelName,
+                tableName: modelName.toLowerCase() + 's',
+                fields: [],
+                file,
+                line: lineNum,
+                framework: 'django',
+                confidence: 0.85,
+            });
+        }
+    }
+    // Detect SQLAlchemy models
+    const sqlalchemyPattern = /class\s+(\w+)\s*\([^)]*(?:Base|DeclarativeBase)[^)]*\)/g;
+    while ((match = sqlalchemyPattern.exec(content)) !== null) {
+        const modelName = match[1];
+        if (modelName) {
+            const lineNum = content.substring(0, match.index).split('\n').length;
+            models.push({
+                name: modelName,
+                tableName: modelName.toLowerCase() + 's',
+                fields: [],
+                file,
+                line: lineNum,
+                framework: 'sqlalchemy',
+                confidence: 0.85,
+            });
+        }
+    }
+    // Detect TypeORM entities
+    const typeormPattern = /@Entity\s*\([^)]*\)\s*(?:export\s+)?class\s+(\w+)/g;
+    while ((match = typeormPattern.exec(content)) !== null) {
+        const modelName = match[1];
+        if (modelName) {
+            const lineNum = content.substring(0, match.index).split('\n').length;
+            models.push({
+                name: modelName,
+                tableName: modelName.toLowerCase() + 's',
+                fields: [],
+                file,
+                line: lineNum,
+                framework: 'typeorm',
+                confidence: 0.9,
+            });
+        }
+    }
+    // Detect Prisma models (from schema.prisma)
+    if (file.endsWith('.prisma')) {
+        const prismaPattern = /model\s+(\w+)\s*\{/g;
+        while ((match = prismaPattern.exec(content)) !== null) {
+            const modelName = match[1];
+            if (modelName) {
+                const lineNum = content.substring(0, match.index).split('\n').length;
+                models.push({
+                    name: modelName,
+                    tableName: modelName.toLowerCase() + 's',
+                    fields: [],
+                    file,
+                    line: lineNum,
+                    framework: 'prisma',
+                    confidence: 0.95,
+                });
+            }
+        }
+    }
+    // Detect Sequelize models
+    const sequelizePattern = /sequelize\.define\s*\(\s*['"](\w+)['"]/g;
+    while ((match = sequelizePattern.exec(content)) !== null) {
+        const modelName = match[1];
+        if (modelName) {
+            const lineNum = content.substring(0, match.index).split('\n').length;
+            models.push({
+                name: modelName,
+                tableName: modelName.toLowerCase() + 's',
+                fields: [],
+                file,
+                line: lineNum,
+                framework: 'sequelize',
+                confidence: 0.85,
+            });
+        }
+    }
+    return models;
+}
+// ============================================================================
+// Query Access Detection
+// ============================================================================
+/**
+ * Common table name patterns - helps identify tables from variable names
+ * Maps common suffixes/patterns to likely table names
+ */
+const TABLE_NAME_HINTS = {
+    // User-related
+    'user': 'users',
+    'users': 'users',
+    'account': 'accounts',
+    'accounts': 'accounts',
+    'profile': 'profiles',
+    'profiles': 'profiles',
+    'auth': 'auth',
+    'session': 'sessions',
+    'sessions': 'sessions',
+    'token': 'tokens',
+    'tokens': 'tokens',
+    // Content
+    'post': 'posts',
+    'posts': 'posts',
+    'comment': 'comments',
+    'comments': 'comments',
+    'article': 'articles',
+    'articles': 'articles',
+    'document': 'documents',
+    'documents': 'documents',
+    'file': 'files',
+    'files': 'files',
+    'image': 'images',
+    'images': 'images',
+    'media': 'media',
+    // E-commerce
+    'order': 'orders',
+    'orders': 'orders',
+    'product': 'products',
+    'products': 'products',
+    'cart': 'carts',
+    'carts': 'carts',
+    'item': 'items',
+    'items': 'items',
+    'payment': 'payments',
+    'payments': 'payments',
+    'transaction': 'transactions',
+    'transactions': 'transactions',
+    'invoice': 'invoices',
+    'invoices': 'invoices',
+    // Organization
+    'company': 'companies',
+    'companies': 'companies',
+    'organization': 'organizations',
+    'organizations': 'organizations',
+    'team': 'teams',
+    'teams': 'teams',
+    'member': 'members',
+    'members': 'members',
+    'role': 'roles',
+    'roles': 'roles',
+    'permission': 'permissions',
+    'permissions': 'permissions',
+    // Communication
+    'message': 'messages',
+    'messages': 'messages',
+    'notification': 'notifications',
+    'notifications': 'notifications',
+    'email': 'emails',
+    'emails': 'emails',
+    // Analytics
+    'event': 'events',
+    'events': 'events',
+    'log': 'logs',
+    'logs': 'logs',
+    'audit': 'audit_logs',
+    'metric': 'metrics',
+    'metrics': 'metrics',
+    // Settings
+    'setting': 'settings',
+    'settings': 'settings',
+    'config': 'configs',
+    'preference': 'preferences',
+    'preferences': 'preferences',
+    // Relationships
+    'subscription': 'subscriptions',
+    'subscriptions': 'subscriptions',
+    'follower': 'followers',
+    'followers': 'followers',
+    'friend': 'friends',
+    'friends': 'friends',
+    'connection': 'connections',
+    'connections': 'connections',
+    // Generic
+    'record': 'records',
+    'records': 'records',
+    'entry': 'entries',
+    'entries': 'entries',
+    'data': 'data',
+    'result': 'results',
+    'results': 'results',
+};
+/**
+ * Common field patterns for different data types
+ */
+const COMMON_FIELD_PATTERNS = [
+    // ID fields
+    { pattern: /\bid\b/i, fields: ['id'] },
+    { pattern: /\buser_id\b/i, fields: ['user_id'] },
+    { pattern: /\baccount_id\b/i, fields: ['account_id'] },
+    { pattern: /\borganization_id\b/i, fields: ['organization_id'] },
+    // Timestamps
+    { pattern: /\bcreated_at\b/i, fields: ['created_at'] },
+    { pattern: /\bupdated_at\b/i, fields: ['updated_at'] },
+    { pattern: /\bdeleted_at\b/i, fields: ['deleted_at'] },
+    // User fields
+    { pattern: /\bemail\b/i, fields: ['email'] },
+    { pattern: /\busername\b/i, fields: ['username'] },
+    { pattern: /\bpassword\b/i, fields: ['password'] },
+    { pattern: /\bname\b/i, fields: ['name'] },
+    { pattern: /\bfirst_name\b/i, fields: ['first_name'] },
+    { pattern: /\blast_name\b/i, fields: ['last_name'] },
+    { pattern: /\bphone\b/i, fields: ['phone'] },
+    { pattern: /\baddress\b/i, fields: ['address'] },
+    // Status fields
+    { pattern: /\bstatus\b/i, fields: ['status'] },
+    { pattern: /\bstate\b/i, fields: ['state'] },
+    { pattern: /\bactive\b/i, fields: ['active'] },
+    { pattern: /\benabled\b/i, fields: ['enabled'] },
+    // Content fields
+    { pattern: /\btitle\b/i, fields: ['title'] },
+    { pattern: /\bdescription\b/i, fields: ['description'] },
+    { pattern: /\bcontent\b/i, fields: ['content'] },
+    { pattern: /\bbody\b/i, fields: ['body'] },
+    // Financial
+    { pattern: /\bamount\b/i, fields: ['amount'] },
+    { pattern: /\bprice\b/i, fields: ['price'] },
+    { pattern: /\btotal\b/i, fields: ['total'] },
+    { pattern: /\bbalance\b/i, fields: ['balance'] },
+];
+/**
+ * Extract table name from a line of code using multiple strategies
+ */
+function extractTableName(line, context) {
+    // Strategy 1: Supabase .from('table_name')
+    const supabaseFrom = line.match(/\.from\s*\(\s*["']([a-zA-Z_][a-zA-Z0-9_]*)["']/i);
+    if (supabaseFrom?.[1])
+        return supabaseFrom[1];
+    // Strategy 2: Supabase with template literal .from(`table_name`)
+    const supabaseTemplate = line.match(/\.from\s*\(\s*`([a-zA-Z_][a-zA-Z0-9_]*)`/i);
+    if (supabaseTemplate?.[1])
+        return supabaseTemplate[1];
+    // Strategy 3: Prisma prisma.tableName.
+    const prisma = line.match(/prisma\.([a-zA-Z_][a-zA-Z0-9_]*)\./i);
+    if (prisma?.[1])
+        return prisma[1];
+    // Strategy 4: Drizzle db.select().from(tableName) or db.insert(tableName)
+    const drizzleFrom = line.match(/\.from\s*\(\s*([a-zA-Z_][a-zA-Z0-9_]*)\s*\)/i);
+    if (drizzleFrom?.[1] && !drizzleFrom[1].startsWith('"') && !drizzleFrom[1].startsWith("'")) {
+        const hint = TABLE_NAME_HINTS[drizzleFrom[1].toLowerCase()];
+        if (hint)
+            return hint;
+        return drizzleFrom[1];
+    }
+    // Strategy 5: SQL FROM clause
+    const sqlFrom = line.match(/FROM\s+["'`]?([a-zA-Z_][a-zA-Z0-9_]*)["'`]?/i);
+    if (sqlFrom?.[1])
+        return sqlFrom[1];
+    // Strategy 6: SQL INTO clause
+    const sqlInto = line.match(/INTO\s+["'`]?([a-zA-Z_][a-zA-Z0-9_]*)["'`]?/i);
+    if (sqlInto?.[1])
+        return sqlInto[1];
+    // Strategy 7: SQL UPDATE clause
+    const sqlUpdate = line.match(/UPDATE\s+["'`]?([a-zA-Z_][a-zA-Z0-9_]*)["'`]?/i);
+    if (sqlUpdate?.[1])
+        return sqlUpdate[1];
+    // Strategy 8: SQL DELETE FROM clause
+    const sqlDelete = line.match(/DELETE\s+FROM\s+["'`]?([a-zA-Z_][a-zA-Z0-9_]*)["'`]?/i);
+    if (sqlDelete?.[1])
+        return sqlDelete[1];
+    // Strategy 9: Django Model.objects
+    const djangoModel = line.match(/([A-Z][a-zA-Z0-9]*)\.objects/);
+    if (djangoModel?.[1])
+        return djangoModel[1].toLowerCase() + 's';
+    // Strategy 10: SQLAlchemy session.query(Model)
+    const sqlalchemyQuery = line.match(/\.query\s*\(\s*([A-Z][a-zA-Z0-9]*)\s*\)/);
+    if (sqlalchemyQuery?.[1])
+        return sqlalchemyQuery[1].toLowerCase() + 's';
+    // Strategy 11: TypeORM getRepository(Entity)
+    const typeormRepo = line.match(/getRepository\s*\(\s*([A-Z][a-zA-Z0-9]*)\s*\)/);
+    if (typeormRepo?.[1])
+        return typeormRepo[1].toLowerCase() + 's';
+    // Strategy 12: Sequelize Model.findAll/create/etc
+    const sequelizeModel = line.match(/([A-Z][a-zA-Z0-9]*)\.(?:find|create|update|destroy|bulkCreate)/);
+    if (sequelizeModel?.[1])
+        return sequelizeModel[1].toLowerCase() + 's';
+    // Strategy 13: Mongoose Model.find/save/etc
+    const mongooseModel = line.match(/([A-Z][a-zA-Z0-9]*)\.(?:find|findOne|findById|create|save|updateOne|deleteOne)/);
+    if (mongooseModel?.[1])
+        return mongooseModel[1].toLowerCase() + 's';
+    // Strategy 14: Knex table('name')
+    const knexTable = line.match(/\.table\s*\(\s*["']([a-zA-Z_][a-zA-Z0-9_]*)["']/i);
+    if (knexTable?.[1])
+        return knexTable[1];
+    // Strategy 15: Generic ORM pattern - look for variable assignment with table hint
+    const varAssignment = line.match(/(?:const|let|var)\s+(\w+)\s*=.*\.(?:from|table|query)/i);
+    if (varAssignment?.[1]) {
+        const hint = TABLE_NAME_HINTS[varAssignment[1].toLowerCase()];
+        if (hint)
+            return hint;
+    }
+    // Strategy 16: Look in surrounding context for table hints
+    for (const ctxLine of context) {
+        const tableHint = ctxLine.match(/table\s*[=:]\s*["']([a-zA-Z_][a-zA-Z0-9_]*)["']/i);
+        if (tableHint?.[1])
+            return tableHint[1];
+        const collectionHint = ctxLine.match(/collection\s*[=:]\s*["']([a-zA-Z_][a-zA-Z0-9_]*)["']/i);
+        if (collectionHint?.[1])
+            return collectionHint[1];
+    }
+    // Strategy 17: Infer from common variable names in the line
+    for (const [hint, table] of Object.entries(TABLE_NAME_HINTS)) {
+        const varPattern = new RegExp(`\\b${hint}(?:s|Data|Result|Response|Query)?\\b`, 'i');
+        if (varPattern.test(line)) {
+            return table;
+        }
+    }
+    return 'unknown';
+}
+/**
+ * Extract field names from a line of code
+ */
+function extractFields(line) {
+    const fields = [];
+    // Extract from .select('field1, field2')
+    const selectMatch = line.match(/\.select\s*\(\s*["'`]([^"'`]+)["'`]/i);
+    if (selectMatch?.[1]) {
+        const selectFields = selectMatch[1].split(/\s*,\s*/);
+        fields.push(...selectFields.map(f => f.trim()).filter(f => f && f !== '*'));
+    }
+    // Extract from .eq('field', value) or .match({ field: value })
+    const eqMatch = line.match(/\.eq\s*\(\s*["'](\w+)["']/i);
+    if (eqMatch?.[1])
+        fields.push(eqMatch[1]);
+    // Extract from object keys in .insert({ field: value })
+    const insertMatch = line.match(/\.(?:insert|update|upsert)\s*\(\s*\{([^}]+)\}/i);
+    if (insertMatch?.[1]) {
+        const keyMatches = insertMatch[1].matchAll(/(\w+)\s*:/g);
+        for (const m of keyMatches) {
+            if (m[1])
+                fields.push(m[1]);
+        }
+    }
+    // Extract common fields from patterns
+    for (const { pattern, fields: patternFields } of COMMON_FIELD_PATTERNS) {
+        if (pattern.test(line)) {
+            fields.push(...patternFields);
+        }
+    }
+    // Deduplicate
+    return [...new Set(fields)];
+}
+// ============================================================================
+// Boundary Scanner Class
+// ============================================================================
+/**
+ * Boundary Scanner - Scans source files for data access patterns
+ *
+ * Uses a two-phase approach:
+ * 1. LEARN: First pass discovers patterns from the codebase
+ * 2. DETECT: Second pass uses learned patterns + regex fallback
+ */
+export class BoundaryScanner {
+    config;
+    store;
+    learner;
+    learnedConventions = null;
+    constructor(config) {
+        this.config = config;
+        this.store = createBoundaryStore({ rootDir: config.rootDir });
+        this.learner = createDataAccessLearner();
+    }
+    async initialize() {
+        await this.store.initialize();
+    }
+    /**
+     * Scan files for data access patterns using learning approach
+     */
+    async scanFiles(files) {
+        const startTime = Date.now();
+        let filesScanned = 0;
+        const allModels = [];
+        const allAccessPoints = [];
+        const allSensitiveFields = [];
+        // Reset learner for fresh scan
+        this.learner.reset();
+        // ========================================================================
+        // PHASE 1: LEARN - Discover patterns from the codebase
+        // ========================================================================
+        if (!this.config.skipLearning) {
+            if (this.config.verbose) {
+                console.log('Phase 1: Learning data access patterns from codebase...');
+            }
+            for (const file of files) {
+                const filePath = path.join(this.config.rootDir, file);
+                const language = getLanguage(file);
+                if (!language && !file.endsWith('.prisma'))
+                    continue;
+                try {
+                    const content = await fs.readFile(filePath, 'utf-8');
+                    if (isDataAccessFile(file, content) || file.endsWith('.prisma')) {
+                        this.learner.learnFromFile(content, file);
+                    }
+                }
+                catch {
+                    // Skip files that can't be read during learning
+                }
+            }
+            // Finalize learning
+            this.learnedConventions = this.learner.finalizeLearning(files.length);
+            if (this.config.verbose) {
+                console.log(`  Learned ${this.learnedConventions.tables.size} tables`);
+                console.log(`  Primary framework: ${this.learnedConventions.primaryFramework ?? 'mixed'}`);
+                console.log(`  Naming convention: ${this.learnedConventions.tableNamingConvention}`);
+            }
+        }
+        // ========================================================================
+        // PHASE 2: DETECT - Use learned patterns + regex fallback
+        // ========================================================================
+        if (this.config.verbose) {
+            console.log('Phase 2: Detecting data access with learned patterns...');
+        }
+        for (const file of files) {
+            const filePath = path.join(this.config.rootDir, file);
+            const language = getLanguage(file);
+            if (!language && !file.endsWith('.prisma'))
+                continue;
+            try {
+                const content = await fs.readFile(filePath, 'utf-8');
+                if (isDataAccessFile(file, content) || file.endsWith('.prisma')) {
+                    filesScanned++;
+                    allModels.push(...detectORMModels(content, file));
+                    // Use learned patterns for detection, with regex fallback
+                    const accessPoints = this.detectWithLearning(content, file);
+                    allAccessPoints.push(...accessPoints);
+                    allSensitiveFields.push(...detectSensitiveFields(content, file));
+                }
+            }
+            catch (error) {
+                if (this.config.verbose) {
+                    console.error(`Error scanning ${file}:`, error.message);
+                }
+            }
+        }
+        for (const model of allModels) {
+            this.store.addModel(model);
+        }
+        for (const accessPoint of allAccessPoints) {
+            this.store.addAccessPoint(accessPoint);
+        }
+        for (const field of allSensitiveFields) {
+            this.store.addSensitiveField(field);
+        }
+        const accessMap = this.store.getAccessMap();
+        await this.store.saveAccessMap(accessMap);
+        const violations = this.store.checkAllViolations();
+        const duration = Date.now() - startTime;
+        return {
+            accessMap,
+            violations,
+            stats: {
+                filesScanned,
+                tablesFound: Object.keys(accessMap.tables).length,
+                accessPointsFound: Object.keys(accessMap.accessPoints).length,
+                sensitiveFieldsFound: accessMap.sensitiveFields.length,
+                violationsFound: violations.length,
+                scanDurationMs: duration,
+            },
+        };
+    }
+    /**
+     * Detect data access using learned patterns first, regex as fallback
+     */
+    detectWithLearning(content, file) {
+        const accessPoints = [];
+        const lines = content.split('\n');
+        for (let i = 0; i < lines.length; i++) {
+            const line = lines[i];
+            if (!line)
+                continue;
+            // Skip comments
+            const trimmed = line.trim();
+            if (trimmed.startsWith('//') || trimmed.startsWith('#') ||
+                trimmed.startsWith('*') || trimmed.startsWith('/*')) {
+                continue;
+            }
+            // Get surrounding context
+            const contextStart = Math.max(0, i - 5);
+            const contextEnd = Math.min(lines.length, i + 5);
+            const context = lines.slice(contextStart, contextEnd).filter(l => l !== undefined);
+            // Check if this line has a data access pattern
+            const operation = this.detectOperation(line);
+            if (!operation)
+                continue;
+            // Try to extract table name using multiple strategies
+            let table = 'unknown';
+            let confidence = 0.5;
+            // Strategy 1: Use learned patterns (highest confidence)
+            if (this.learnedConventions?.hasEnoughData) {
+                const learnedTable = this.extractTableWithLearning(line, context);
+                if (learnedTable) {
+                    table = learnedTable;
+                    confidence = 0.95; // High confidence for learned patterns
+                }
+            }
+            // Strategy 2: Regex fallback (lower confidence)
+            if (table === 'unknown') {
+                table = extractTableName(line, context);
+                confidence = table === 'unknown' ? 0.3 : 0.7;
+            }
+            const fields = extractFields(line);
+            const id = `${file}:${i + 1}:0:${table}`;
+            if (!accessPoints.find(ap => ap.id === id)) {
+                accessPoints.push({
+                    id,
+                    table,
+                    fields,
+                    operation,
+                    file,
+                    line: i + 1,
+                    column: 0,
+                    context: trimmed.slice(0, 100),
+                    isRawSql: /\b(SELECT|INSERT|UPDATE|DELETE)\b/i.test(line),
+                    confidence,
+                });
+            }
+        }
+        return accessPoints;
+    }
+    /**
+     * Extract table name using learned patterns
+     */
+    extractTableWithLearning(line, context) {
+        if (!this.learnedConventions)
+            return null;
+        // Check if any learned table pattern matches
+        for (const [tableName] of this.learnedConventions.tables) {
+            // Check if the table name appears in the line
+            const tableRegex = new RegExp(`['"\`]${tableName}['"\`]|\\b${tableName}\\b`, 'i');
+            if (tableRegex.test(line)) {
+                return tableName;
+            }
+        }
+        // Try to infer from variable names
+        const varMatch = line.match(/(?:const|let|var)\s+(\w+)|(\w+)\s*=\s*(?:await\s+)?/);
+        if (varMatch) {
+            const varName = (varMatch[1] || varMatch[2])?.toLowerCase();
+            if (varName) {
+                const inferredTable = this.learner.inferTableFromVariable(varName);
+                if (inferredTable)
+                    return inferredTable;
+            }
+        }
+        // Check context for learned tables
+        for (const ctxLine of context) {
+            for (const [tableName] of this.learnedConventions.tables) {
+                const tableRegex = new RegExp(`['"\`]${tableName}['"\`]`, 'i');
+                if (tableRegex.test(ctxLine)) {
+                    return tableName;
+                }
+            }
+        }
+        return null;
+    }
+    /**
+     * Detect the operation type from a line
+     */
+    detectOperation(line) {
+        // Read operations
+        if (/\.find\w*\s*\(|\.get\w*\s*\(|\.where\s*\(|\.select\s*\(|\.filter\s*\(|\.all\s*\(|\.objects\.|\.query\s*\(|\.from\s*\(|\bSELECT\b/i.test(line)) {
+            return 'read';
+        }
+        // Write operations
+        if (/\.create\s*\(|\.save\s*\(|\.update\s*\(|\.insert\s*\(|\.upsert\s*\(|\bINSERT\b|\bUPDATE\b/i.test(line)) {
+            return 'write';
+        }
+        // Delete operations
+        if (/\.delete\s*\(|\.remove\s*\(|\.destroy\s*\(|\bDELETE\b/i.test(line)) {
+            return 'delete';
+        }
+        return null;
+    }
+    /**
+     * Scan directory with glob patterns
+     */
+    async scanDirectory(options = {}) {
+        const patterns = options.patterns ?? ['**/*.ts', '**/*.tsx', '**/*.js', '**/*.jsx', '**/*.py'];
+        const ignorePatterns = options.ignorePatterns ?? ['node_modules', '.git', 'dist', 'build', '__pycache__', '.drift'];
+        const files = await this.findFiles(patterns, ignorePatterns);
+        return this.scanFiles(files);
+    }
+    /**
+     * Get learned conventions (for debugging/inspection)
+     */
+    getLearnedConventions() {
+        return this.learnedConventions;
+    }
+    /**
+     * Get the underlying store
+     */
+    getStore() {
+        return this.store;
+    }
+    /**
+     * Find files matching patterns
+     */
+    async findFiles(patterns, ignorePatterns) {
+        const files = [];
+        const walk = async (dir, relativePath = '') => {
+            const entries = await fs.readdir(dir, { withFileTypes: true });
+            for (const entry of entries) {
+                const fullPath = path.join(dir, entry.name);
+                const relPath = relativePath ? `${relativePath}/${entry.name}` : entry.name;
+                if (entry.isDirectory()) {
+                    if (!ignorePatterns.includes(entry.name) && !entry.name.startsWith('.')) {
+                        await walk(fullPath, relPath);
+                    }
+                }
+                else if (entry.isFile()) {
+                    for (const pattern of patterns) {
+                        if (minimatch(relPath, pattern)) {
+                            files.push(relPath);
+                            break;
+                        }
+                    }
+                }
+            }
+        };
+        await walk(this.config.rootDir);
+        return files;
+    }
+}
+// ============================================================================
+// Factory Function
+// ============================================================================
+/**
+ * Create a new BoundaryScanner instance
+ */
+export function createBoundaryScanner(config) {
+    return new BoundaryScanner(config);
+}
+//# sourceMappingURL=boundary-scanner.js.map