fix(arch-scan): 修复多语言扫描器并添加 DELETE 清空表

- scanner.ts: 添加 DELETE 清空旧数据语句，避免重复插入
- ts-scanner: 改用 regex 替代 ts-morph 解析（避免解析失败）
- go-scanner/py-scanner/proto-scanner: 完善实现

scripts/arch-scan/scanner.ts

@@ -1,17 +1,22 @@
 #!/usr/bin/env tsx
-import { createDb } from './schema.js';
+import { createDb } from "./schema.js";
-import { scanTypeScript } from './scanners/ts-scanner.js';
+import { scanTypeScript } from "./scanners/ts-scanner.js";
-import { scanGo } from './scanners/go-scanner.js';
+import { scanGo } from "./scanners/go-scanner.js";
-import { scanPython } from './scanners/py-scanner.js';
+import { scanPython } from "./scanners/py-scanner.js";
-import { scanProtobuf } from './scanners/proto-scanner.js';
+import { scanProtobuf } from "./scanners/proto-scanner.js";
-import path from 'node:path';
+import path from "node:path";
 
 const ROOT = process.cwd();
 
 function main(): void {
-  const db = createDb(path.join(ROOT, 'arch.db'));
+  const db = createDb(path.join(ROOT, "arch.db"));
 
-  console.log('🔍 arch-scan: 开始多语言扫描...');
+  // 清空旧数据（避免重跑产生重复符号）
+  db.exec(
+    "DELETE FROM calls; DELETE FROM dependencies; DELETE FROM symbols; DELETE FROM contracts; DELETE FROM events; DELETE FROM modules;",
+  );
+
+  console.log("🔍 arch-scan: 开始多语言扫描...");
 
   // 串行扫描（并行会导致 FOREIGN KEY 错误）
   const tsStats = scanTypeScript(db, ROOT);
@@ -26,8 +31,8 @@ function main(): void {
   const protoStats = scanProtobuf(db, ROOT);
   console.log(`  Protobuf: ${protoStats.contracts} 契约`);
 
-  console.log('✅ arch-scan 完成');
+  console.log("✅ arch-scan 完成");
   db.close();
 }
 
-main();
+main();

scripts/arch-scan/scanners/go-scanner.ts

@@ -1,12 +1,133 @@
-import type { Database as DBType } from 'better-sqlite3';
+import path from "node:path";
+import fs from "node:fs";
+import type { Database as DBType } from "better-sqlite3";
 
 interface ScanStats {
   modules: number;
   symbols: number;
 }
 
+/** 递归收集目录下匹配扩展名的所有文件 */
+function walkDir(dir: string, exts: string[]): string[] {
+  if (!fs.existsSync(dir)) return [];
+  const results: string[] = [];
+  const stack: string[] = [dir];
+  while (stack.length > 0) {
+    const current = stack.pop()!;
+    let entries: fs.Dirent[];
+    try {
+      entries = fs.readdirSync(current, { withFileTypes: true });
+    } catch {
+      continue;
+    }
+    for (const entry of entries) {
+      const fullPath = path.join(current, entry.name);
+      if (entry.isDirectory()) {
+        if (["node_modules", "dist", ".git", "vendor"].includes(entry.name))
+          continue;
+        stack.push(fullPath);
+      } else if (entry.isFile()) {
+        const ext = path.extname(entry.name);
+        if (exts.includes(ext)) results.push(fullPath);
+      }
+    }
+  }
+  return results;
+}
+
+// Go 符号正则：匹配导出函数、类型、结构体、接口
+const GO_FUNC_RE = /^func\s+(?:\([^)]*\)\s+)?([A-Z]\w*)\s*\(/gm;
+const GO_TYPE_RE = /^type\s+([A-Z]\w*)\s+/gm;
+const GO_STRUCT_RE = /^type\s+([A-Z]\w*)\s+struct\s*\{/gm;
+const GO_INTERFACE_RE = /^type\s+([A-Z]\w*)\s+interface\s*\{/gm;
+
 export function scanGo(db: DBType, root: string): ScanStats {
-  // Go 扫描器骨架：P1 后期用 tree-sitter-go 实现
+  const insertModule = db.prepare(
-  // 当前仅扫描 go.mod 识别模块
+    "INSERT OR IGNORE INTO modules (name, path, language, service, type) VALUES (?, ?, ?, ?, ?)",
-  return { modules: 0, symbols: 0 };
+  );
-}
+  const insertSymbol = db.prepare(
+    "INSERT INTO symbols (module_id, name, kind, language, file_path, line_start, line_end, is_exported) VALUES (?, ?, ?, ?, ?, ?, ?, ?)",
+  );
+  const getModuleId = db.prepare("SELECT id FROM modules WHERE name = ?") as {
+    get: (name: string) => { id: number } | undefined;
+  };
+
+  let modules = 0;
+  let symbols = 0;
+
+  // 扫描 services/* 下有 go.mod 的服务
+  const servicesDir = path.join(root, "services");
+  if (!fs.existsSync(servicesDir)) return { modules, symbols };
+
+  const entries = fs.readdirSync(servicesDir, { withFileTypes: true });
+  for (const entry of entries) {
+    if (!entry.isDirectory()) continue;
+    const serviceName = entry.name;
+    const servicePath = path.join(servicesDir, serviceName);
+    const hasGoMod = fs.existsSync(path.join(servicePath, "go.mod"));
+    if (!hasGoMod) continue;
+
+    insertModule.run(serviceName, servicePath, "go", serviceName, "service");
+    const modRow = getModuleId.get(serviceName);
+    if (modRow) {
+      modules++;
+      // 扫描 .go 文件
+      const goFiles = walkDir(servicePath, [".go"]);
+      for (const filePath of goFiles) {
+        // 跳过测试文件
+        if (filePath.endsWith("_test.go")) continue;
+        let content: string;
+        try {
+          content = fs.readFileSync(filePath, "utf-8");
+        } catch {
+          continue;
+        }
+        const lines = content.split("\n");
+
+        // 函数
+        let match: RegExpExecArray | null;
+        const funcRe = new RegExp(GO_FUNC_RE);
+        while ((match = funcRe.exec(content)) !== null) {
+          const name = match[1];
+          const lineNum = content.slice(0, match.index).split("\n").length;
+          insertSymbol.run(
+            modRow.id,
+            name,
+            "function",
+            "go",
+            filePath,
+            lineNum,
+            lineNum,
+            1,
+          );
+          symbols++;
+        }
+
+        // 类型（struct / interface / type alias）
+        const typeRe = new RegExp(GO_TYPE_RE);
+        while ((match = typeRe.exec(content)) !== null) {
+          const name = match[1];
+          const lineNum = content.slice(0, match.index).split("\n").length;
+          // 判断是 struct 还是 interface 还是普通 type
+          const line = lines[lineNum - 1] || "";
+          let kind = "type";
+          if (line.includes("struct")) kind = "struct";
+          else if (line.includes("interface")) kind = "interface";
+          insertSymbol.run(
+            modRow.id,
+            name,
+            kind,
+            "go",
+            filePath,
+            lineNum,
+            lineNum,
+            1,
+          );
+          symbols++;
+        }
+      }
+    }
+  }
+
+  return { modules, symbols };
+}

scripts/arch-scan/scanners/proto-scanner.ts

@@ -1,10 +1,92 @@
-import type { Database as DBType } from 'better-sqlite3';
+import path from "node:path";
+import fs from "node:fs";
+import type { Database as DBType } from "better-sqlite3";
 
 interface ScanStats {
   contracts: number;
 }
 
+// Proto 符号正则
+const PROTO_SERVICE_RE = /^service\s+(\w+)\s*\{/gm;
+const PROTO_MESSAGE_RE = /^message\s+(\w+)\s*\{/gm;
+const PROTO_RPC_RE = /^\s*rpc\s+(\w+)\s*\(/gm;
+
 export function scanProtobuf(db: DBType, root: string): ScanStats {
-  // Protobuf 扫描器骨架：扫描 packages/shared-proto/proto/*.proto
+  const insertContract = db.prepare(
-  return { contracts: 0 };
+    "INSERT OR IGNORE INTO contracts (name, type, file_path, service, content) VALUES (?, ?, ?, ?, ?)",
-}
+  );
+
+  let contracts = 0;
+
+  // 扫描 packages/shared-proto/proto/*.proto
+  const protoDir = path.join(root, "packages", "shared-proto", "proto");
+  if (!fs.existsSync(protoDir)) return { contracts };
+
+  const entries = fs.readdirSync(protoDir, { withFileTypes: true });
+  for (const entry of entries) {
+    if (!entry.isFile() || !entry.name.endsWith(".proto")) continue;
+    const filePath = path.join(protoDir, entry.name);
+    const protoName = path.basename(entry.name, ".proto");
+    const serviceName = protoName.replace("_", "-");
+
+    let content: string;
+    try {
+      content = fs.readFileSync(filePath, "utf-8");
+    } catch {
+      continue;
+    }
+
+    // 提取 package 名
+    const pkgMatch = content.match(/^package\s+(\S+);/m);
+    const pkgName = pkgMatch ? pkgMatch[1] : "";
+
+    // 记录 proto 文件作为契约
+    insertContract.run(protoName, "proto-file", filePath, serviceName, pkgName);
+    contracts++;
+
+    // 提取 service 定义
+    let match: RegExpExecArray | null;
+    const serviceRe = new RegExp(PROTO_SERVICE_RE);
+    while ((match = serviceRe.exec(content)) !== null) {
+      const name = match[1];
+      insertContract.run(
+        `${protoName}.${name}`,
+        "service",
+        filePath,
+        serviceName,
+        name,
+      );
+      contracts++;
+    }
+
+    // 提取 message 定义
+    const messageRe = new RegExp(PROTO_MESSAGE_RE);
+    while ((match = messageRe.exec(content)) !== null) {
+      const name = match[1];
+      insertContract.run(
+        `${protoName}.${name}`,
+        "message",
+        filePath,
+        serviceName,
+        name,
+      );
+      contracts++;
+    }
+
+    // 提取 rpc 方法
+    const rpcRe = new RegExp(PROTO_RPC_RE);
+    while ((match = rpcRe.exec(content)) !== null) {
+      const name = match[1];
+      insertContract.run(
+        `${protoName}.rpc.${name}`,
+        "rpc",
+        filePath,
+        serviceName,
+        name,
+      );
+      contracts++;
+    }
+  }
+
+  return { contracts };
+}

scripts/arch-scan/scanners/py-scanner.ts

@@ -1,11 +1,142 @@
-import type { Database as DBType } from 'better-sqlite3';
+import path from "node:path";
+import fs from "node:fs";
+import type { Database as DBType } from "better-sqlite3";
 
 interface ScanStats {
   modules: number;
   symbols: number;
 }
 
+/** 递归收集目录下匹配扩展名的所有文件 */
+function walkDir(dir: string, exts: string[]): string[] {
+  if (!fs.existsSync(dir)) return [];
+  const results: string[] = [];
+  const stack: string[] = [dir];
+  while (stack.length > 0) {
+    const current = stack.pop()!;
+    let entries: fs.Dirent[];
+    try {
+      entries = fs.readdirSync(current, { withFileTypes: true });
+    } catch {
+      continue;
+    }
+    for (const entry of entries) {
+      const fullPath = path.join(current, entry.name);
+      if (entry.isDirectory()) {
+        if (
+          [
+            "node_modules",
+            "dist",
+            ".git",
+            "__pycache__",
+            ".venv",
+            "venv",
+          ].includes(entry.name)
+        )
+          continue;
+        stack.push(fullPath);
+      } else if (entry.isFile()) {
+        const ext = path.extname(entry.name);
+        if (exts.includes(ext)) results.push(fullPath);
+      }
+    }
+  }
+  return results;
+}
+
+// Python 符号正则
+const PY_FUNC_RE = /^(?:async\s+)?def\s+(\w+)\s*\(/gm;
+const PY_CLASS_RE = /^class\s+(\w+)\s*[\(:]/gm;
+
 export function scanPython(db: DBType, root: string): ScanStats {
-  // Python 扫描器骨架：P1 后期用 tree-sitter-python 实现
+  const insertModule = db.prepare(
-  return { modules: 0, symbols: 0 };
+    "INSERT OR IGNORE INTO modules (name, path, language, service, type) VALUES (?, ?, ?, ?, ?)",
-}
+  );
+  const insertSymbol = db.prepare(
+    "INSERT INTO symbols (module_id, name, kind, language, file_path, line_start, line_end, is_exported) VALUES (?, ?, ?, ?, ?, ?, ?, ?)",
+  );
+  const getModuleId = db.prepare("SELECT id FROM modules WHERE name = ?") as {
+    get: (name: string) => { id: number } | undefined;
+  };
+
+  let modules = 0;
+  let symbols = 0;
+
+  // 扫描 services/* 下有 pyproject.toml 的服务
+  const servicesDir = path.join(root, "services");
+  if (!fs.existsSync(servicesDir)) return { modules, symbols };
+
+  const entries = fs.readdirSync(servicesDir, { withFileTypes: true });
+  for (const entry of entries) {
+    if (!entry.isDirectory()) continue;
+    const serviceName = entry.name;
+    const servicePath = path.join(servicesDir, serviceName);
+    const hasPyproject = fs.existsSync(
+      path.join(servicePath, "pyproject.toml"),
+    );
+    if (!hasPyproject) continue;
+
+    insertModule.run(
+      serviceName,
+      servicePath,
+      "python",
+      serviceName,
+      "service",
+    );
+    const modRow = getModuleId.get(serviceName);
+    if (modRow) {
+      modules++;
+      // 扫描 .py 文件
+      const pyFiles = walkDir(servicePath, [".py"]);
+      for (const filePath of pyFiles) {
+        let content: string;
+        try {
+          content = fs.readFileSync(filePath, "utf-8");
+        } catch {
+          continue;
+        }
+
+        // 函数（含 async def）
+        let match: RegExpExecArray | null;
+        const funcRe = new RegExp(PY_FUNC_RE);
+        while ((match = funcRe.exec(content)) !== null) {
+          const name = match[1];
+          const lineNum = content.slice(0, match.index).split("\n").length;
+          // Python 中以 _ 开头的为私有
+          const isExported = !name.startsWith("_") ? 1 : 0;
+          insertSymbol.run(
+            modRow.id,
+            name,
+            "function",
+            "python",
+            filePath,
+            lineNum,
+            lineNum,
+            isExported,
+          );
+          symbols++;
+        }
+
+        // 类
+        const classRe = new RegExp(PY_CLASS_RE);
+        while ((match = classRe.exec(content)) !== null) {
+          const name = match[1];
+          const lineNum = content.slice(0, match.index).split("\n").length;
+          insertSymbol.run(
+            modRow.id,
+            name,
+            "class",
+            "python",
+            filePath,
+            lineNum,
+            lineNum,
+            1,
+          );
+          symbols++;
+        }
+      }
+    }
+  }
+
+  return { modules, symbols };
+}

scripts/arch-scan/scanners/ts-scanner.ts

@@ -1,38 +1,211 @@
-import { Project, SyntaxKind } from 'ts-morph';
+import path from "node:path";
-import path from 'node:path';
+import fs from "node:fs";
-import type { Database as DBType } from 'better-sqlite3';
+import type { Database as DBType } from "better-sqlite3";
 
 interface ScanStats {
   modules: number;
   symbols: number;
 }
 
-export function scanTypeScript(db: DBType, root: string): ScanStats {
+/** 递归收集目录下匹配扩展名的所有文件 */
-  const project = new Project({
+function walkDir(dir: string, exts: string[]): string[] {
-    tsConfigFilePath: undefined,
+  if (!fs.existsSync(dir)) return [];
-    skipAddingFilesFromTsConfig: true,
+  const results: string[] = [];
-    compilerOptions: {
+  const stack: string[] = [dir];
-      allowJs: true,
+  while (stack.length > 0) {
-      declaration: false,
+    const current = stack.pop()!;
-      resolveJsonModule: true,
+    let entries: fs.Dirent[];
-    },
+    try {
-  });
+      entries = fs.readdirSync(current, { withFileTypes: true });
+    } catch {
+      continue;
+    }
+    for (const entry of entries) {
+      const fullPath = path.join(current, entry.name);
+      if (entry.isDirectory()) {
+        // 跳过 node_modules / dist / .git
+        if (
+          ["node_modules", "dist", ".git", ".next", "build"].includes(
+            entry.name,
+          )
+        )
+          continue;
+        stack.push(fullPath);
+      } else if (entry.isFile()) {
+        const ext = path.extname(entry.name);
+        if (exts.includes(ext)) results.push(fullPath);
+      }
+    }
+  }
+  return results;
+}
+
+export function scanTypeScript(db: DBType, root: string): ScanStats {
+  const insertModule = db.prepare(
+    "INSERT OR IGNORE INTO modules (name, path, language, service, type) VALUES (?, ?, ?, ?, ?)",
+  );
+  const insertSymbol = db.prepare(
+    "INSERT INTO symbols (module_id, name, kind, language, file_path, line_start, line_end, is_exported) VALUES (?, ?, ?, ?, ?, ?, ?, ?)",
+  );
+  const getModuleId = db.prepare("SELECT id FROM modules WHERE name = ?") as {
+    get: (name: string) => { id: number } | undefined;
+  };
 
-  const patterns = ['services/*/src/**/*.ts', 'apps/*/src/**/*.ts', 'packages/*/src/**/*.ts'];
   let modules = 0;
   let symbols = 0;
 
-  const insertModule = db.prepare(
+  // 扫描 services/* 和 apps/* 和 packages/* 下的 TS 服务
-    'INSERT OR IGNORE INTO modules (name, path, language, service, type) VALUES (?, ?, ?, ?, ?)'
+  const serviceDirs = ["services", "apps"];
-  );
+  const moduleList: { name: string; path: string; service: string }[] = [];
-  const insertSymbol = db.prepare(
-    'INSERT INTO symbols (module_id, name, kind, language, file_path, line_start, line_end, is_exported) VALUES (?, ?, ?, ?, ?, ?, ?, ?)'
-  );
 
-  // 简化实现：扫描 services/*/src 目录作为模块
+  for (const dir of serviceDirs) {
-  const servicesDir = path.join(root, 'services');
+    const baseDir = path.join(root, dir);
-  // 实际实现用 fast-glob 模式匹配
+    if (!fs.existsSync(baseDir)) continue;
-  // 此处为骨架，P1 后期补全
+    const entries = fs.readdirSync(baseDir, { withFileTypes: true });
+    for (const entry of entries) {
+      if (!entry.isDirectory()) continue;
+      const serviceName = entry.name;
+      const servicePath = path.join(baseDir, serviceName);
+      // 确认是 TS 服务（有 src 目录或 package.json）
+      const hasSrc = fs.existsSync(path.join(servicePath, "src"));
+      const hasPkg = fs.existsSync(path.join(servicePath, "package.json"));
+      if (!hasSrc && !hasPkg) continue;
+
+      const moduleName = serviceName;
+      insertModule.run(moduleName, servicePath, "ts", serviceName, "service");
+      const row = getModuleId.get(moduleName);
+      if (row) {
+        modules++;
+        moduleList.push({
+          name: moduleName,
+          path: servicePath,
+          service: serviceName,
+        });
+      }
+    }
+  }
+
+  // packages/* 下的 TS 包
+  const packagesDir = path.join(root, "packages");
+  if (fs.existsSync(packagesDir)) {
+    const entries = fs.readdirSync(packagesDir, { withFileTypes: true });
+    for (const entry of entries) {
+      if (!entry.isDirectory()) continue;
+      const pkgName = entry.name;
+      const pkgPath = path.join(packagesDir, pkgName);
+      const hasSrc = fs.existsSync(path.join(pkgPath, "src"));
+      const hasPkg = fs.existsSync(path.join(pkgPath, "package.json"));
+      const hasProto = fs.existsSync(path.join(pkgPath, "proto"));
+      if (!hasSrc && !hasPkg && !hasProto) continue;
+      const moduleName = pkgName;
+      insertModule.run(moduleName, pkgPath, "ts", pkgName, "package");
+      const row = getModuleId.get(moduleName);
+      if (row) {
+        modules++;
+        moduleList.push({ name: moduleName, path: pkgPath, service: pkgName });
+      }
+    }
+  }
+
+  // 收集所有 .ts 文件并用 regex 提取符号（ts-morph 对未安装依赖的文件可能解析失败）
+  const TS_FUNC_RE = /(?:export\s+)?(?:async\s+)?function\s+(\w+)\s*\(/g;
+  const TS_CLASS_RE = /(?:export\s+)?(?:abstract\s+)?class\s+(\w+)/g;
+  const TS_INTERFACE_RE = /(?:export\s+)?interface\s+(\w+)/g;
+  const TS_CONST_RE = /(?:export\s+)?const\s+([A-Z][A-Z0-9_]+)\s*=/g;
+
+  for (const mod of moduleList) {
+    const srcDir = path.join(mod.path, "src");
+    if (!fs.existsSync(srcDir)) continue;
+    const tsFiles = walkDir(srcDir, [".ts", ".tsx"]);
+    const modRow = getModuleId.get(mod.name);
+    if (!modRow) continue;
+
+    for (const filePath of tsFiles) {
+      let content: string;
+      try {
+        content = fs.readFileSync(filePath, "utf-8");
+      } catch {
+        continue;
+      }
+
+      // 函数
+      let match: RegExpExecArray | null;
+      const funcRe = new RegExp(TS_FUNC_RE);
+      while ((match = funcRe.exec(content)) !== null) {
+        const name = match[1];
+        const lineNum = content.slice(0, match.index).split("\n").length;
+        const isExported = match[0].includes("export") ? 1 : 0;
+        insertSymbol.run(
+          modRow.id,
+          name,
+          "function",
+          "ts",
+          filePath,
+          lineNum,
+          lineNum,
+          isExported,
+        );
+        symbols++;
+      }
+
+      // 类
+      const classRe = new RegExp(TS_CLASS_RE);
+      while ((match = classRe.exec(content)) !== null) {
+        const name = match[1];
+        const lineNum = content.slice(0, match.index).split("\n").length;
+        const isExported = match[0].includes("export") ? 1 : 0;
+        insertSymbol.run(
+          modRow.id,
+          name,
+          "class",
+          "ts",
+          filePath,
+          lineNum,
+          lineNum,
+          isExported,
+        );
+        symbols++;
+      }
+
+      // 接口
+      const ifaceRe = new RegExp(TS_INTERFACE_RE);
+      while ((match = ifaceRe.exec(content)) !== null) {
+        const name = match[1];
+        const lineNum = content.slice(0, match.index).split("\n").length;
+        const isExported = match[0].includes("export") ? 1 : 0;
+        insertSymbol.run(
+          modRow.id,
+          name,
+          "interface",
+          "ts",
+          filePath,
+          lineNum,
+          lineNum,
+          isExported,
+        );
+        symbols++;
+      }
+
+      // 常量（UPPER_CASE）
+      const constRe = new RegExp(TS_CONST_RE);
+      while ((match = constRe.exec(content)) !== null) {
+        const name = match[1];
+        const lineNum = content.slice(0, match.index).split("\n").length;
+        const isExported = match[0].includes("export") ? 1 : 0;
+        insertSymbol.run(
+          modRow.id,
+          name,
+          "const",
+          "ts",
+          filePath,
+          lineNum,
+          lineNum,
+          isExported,
+        );
+        symbols++;
+      }
+    }
+  }
 
   return { modules, symbols };
-}
+}