回传改为TCVB

Author: yefansky

src/cvbManager.ts

@@ -18,6 +18,7 @@ const languageMapping: { [key: string]: string } = {
   'js': 'javascript'
 };
 
+// ================== CVB 核心类 ==================
 export class Cvb {
   private content: string;
   private metadata: Record<string, string>;
@@ -34,6 +35,10 @@ export class Cvb {
     return this.metadata;
   }
 
+  setMetaData(key: string, metadata : string) {
+    this.metadata[key] = metadata;
+  }
+
   getFiles(): Record<string, string> {
     return this.files;
   }
@@ -116,20 +121,357 @@ export class Cvb {
       files,
     };
   }
+
+  static getFormatDescription(): string {
+    return `
+  CVB 格式介绍:
+  - 文件以 "## BEGIN_CVB" 开头，以 "## END_CVB" 结尾。
+  - 元数据部分以 "## META" 开头，以 "## END_META" 结尾，包含用户需求和时间戳。
+  - 每个文件以 "## FILE:文件路径" 开头，紧接着是 Markdown 格式的代码块，包含文件内容。
+  - 多个文件按顺序拼接在一起。
+    `;
+  }
 }
 
-/**
- * 返回 CVB 格式介绍的静态字符串
- * @returns CVB 格式介绍
- */
-export function getCvbFormatDescription(): string {
-  return `
-CVB 格式介绍:
-- 文件以 "## BEGIN_CVB" 开头，以 "## END_CVB" 结尾。
-- 元数据部分以 "## META" 开头，以 "## END_META" 结尾，包含用户需求和时间戳。
-- 每个文件以 "## FILE:文件路径" 开头，紧接着是 Markdown 格式的代码块，包含文件内容。
-- 多个文件按顺序拼接在一起。
-  `;
+// ================== TCVB 差量格式 ==================
+abstract class CvbOperation {
+  constructor(
+    public readonly filePath: string,
+    public readonly type: 'replace' | 'insert' | 'delete'
+  ) {}
+}
+
+class ReplaceOperation extends CvbOperation {
+  constructor(
+    filePath: string,
+    public readonly beforeAnchor: string,
+    public readonly afterAnchor: string,
+    public readonly oldContent: string,
+    public readonly newContent: string
+  ) {
+    super(filePath, 'replace');
+  }
+}
+
+class InsertOperation extends CvbOperation {
+  constructor(
+    filePath: string,
+    public readonly beforeAnchor: string,
+    public readonly afterAnchor: string,
+    public readonly content: string
+  ) {
+    super(filePath, 'insert');
+  }
+}
+
+class DeleteOperation extends CvbOperation {
+  constructor(
+    filePath: string,
+    public readonly beforeAnchor: string,
+    public readonly afterAnchor: string,
+    public readonly oldContent: string
+  ) {
+    super(filePath, 'delete');
+  }
+}
+
+export class TCVB {
+  private operations: CvbOperation[] = [];
+
+  constructor(tcvbContent: string) {
+    this.parse(tcvbContent);
+  }
+
+  private parse(content: string) {
+    const fileBlockRegex = /^## FILE:(.*?)\n([\s\S]*?)(?=^## FILE:|^## END_TCVB)/gm;
+    let fileMatch: RegExpExecArray | null;
+    
+    while ((fileMatch = fileBlockRegex.exec(content)) !== null) {
+      const filePath = filePathNormalize(fileMatch[1]);
+      const operationsBlock = fileMatch[2];
+      
+      const operationRegex = /^## OPERATION:(\w+)(?:\s+FILE:(.*?))?\n([\s\S]*?)(?=^## OPERATION:|^## FILE:|^## END_TCVB)/gm;
+      let opMatch: RegExpExecArray | null;
+
+      while ((opMatch = operationRegex.exec(operationsBlock)) !== null) {
+        const type = opMatch[1].toLowerCase();
+        const explicitFilePath = opMatch[2] ? filePathNormalize(opMatch[2]) : null;
+        const operationContent = opMatch[3].trim();
+
+        const finalFilePath = explicitFilePath || filePath;
+        this.parseOperation(finalFilePath, type, operationContent);
+      }
+    }
+  }
+
+  private parseOperation(filePath: string, type: string, content: string) {
+    try {
+      switch (type) {
+        case 'replace':
+          this.parseReplace(filePath, content);
+          break;
+        case 'insert':
+          this.parseInsert(filePath, content);
+          break;
+        case 'delete':
+          this.parseDelete(filePath, content);
+          break;
+        default:
+          throw new Error(`Unknown operation type: ${type}`);
+      }
+    } catch (e) {
+      console.error(`Failed to parse ${type} operation for ${filePath}: ${e}`);
+    }
+  }
+
+  private parseReplace(filePath: string, content: string) {
+    const sections = this.parseSections(content, ['BEFORE_ANCHOR', 'AFTER_ANCHOR', 'OLD_CONTENT', 'NEW_CONTENT']);
+    this.operations.push(new ReplaceOperation(
+      filePath,
+      sections.BEFORE_ANCHOR,
+      sections.AFTER_ANCHOR,
+      sections.OLD_CONTENT,
+      sections.NEW_CONTENT
+    ));
+  }
+
+  private parseInsert(filePath: string, content: string) {
+    const sections = this.parseSections(content, ['BEFORE_ANCHOR', 'AFTER_ANCHOR', 'INSERT_CONTENT']);
+    this.operations.push(new InsertOperation(
+      filePath,
+      sections.BEFORE_ANCHOR,
+      sections.AFTER_ANCHOR,
+      sections.INSERT_CONTENT
+    ));
+  }
+
+  private parseDelete(filePath: string, content: string) {
+    const sections = this.parseSections(content, ['BEFORE_ANCHOR', 'AFTER_ANCHOR', 'DELETE_CONTENT']);
+    this.operations.push(new DeleteOperation(
+      filePath,
+      sections.BEFORE_ANCHOR,
+      sections.AFTER_ANCHOR,
+      sections.DELETE_CONTENT
+    ));
+  }
+
+  private parseSections(content: string, expectedSections: string[]): Record<string, string> {
+    const result: Record<string, string> = {};
+    let currentSection: string | null = null;
+    let buffer: string[] = [];
+
+    for (const line of content.split('\n')) {
+      const sectionMatch = line.match(/^## ([A-Z_]+)/);
+      if (sectionMatch) {
+        if (currentSection) {
+          result[currentSection] = buffer.join('\n').trim();
+          buffer = [];
+        }
+        currentSection = sectionMatch[1];
+        if (!expectedSections.includes(currentSection)) {
+          throw new Error(`Unexpected section: ${currentSection}`);
+        }
+      } else if (currentSection) {
+        buffer.push(line);
+      }
+    }
+
+    if (currentSection) {
+      result[currentSection] = buffer.join('\n').trim();
+    }
+
+    // Validate required sections
+    for (const section of expectedSections) {
+      if (!(section in result)) {
+        throw new Error(`Missing required section: ${section}`);
+      }
+    }
+
+    return result;
+  }
+
+  getOperations(): CvbOperation[] {
+    return [...this.operations];
+  }
+
+  static getFormatDescription(): string {
+    return `
+TCVB 格式规范（版本2.0）：
+
+## BEGIN_TCVB
+[文件块1]
+[文件块2]
+...
+## END_TCVB
+
+文件块格式：
+## FILE:<文件路径>
+[操作1]
+[操作2]
+...
+
+操作类型：
+1. 替换操作（REPLACE）:
+## OPERATION:REPLACE
+## BEFORE_ANCHOR
+[前锚点内容]
+## AFTER_ANCHOR
+[后锚点内容]
+## OLD_CONTENT
+[被替换内容]
+## NEW_CONTENT
+[新内容]
+
+2. 插入操作（INSERT）:
+## OPERATION:INSERT
+## BEFORE_ANCHOR
+[插入位置前锚点]
+## AFTER_ANCHOR
+[插入位置后锚点]
+## INSERT_CONTENT
+[插入内容]
+
+3. 删除操作（DELETE）:
+## OPERATION:DELETE
+## BEFORE_ANCHOR
+[被删内容前锚点]
+## AFTER_ANCHOR
+[被删内容后锚点]
+## DELETE_CONTENT
+[被删除内容]
+
+高级特性：
+1. 文件路径复用：同一文件下的多个操作共享FILE声明
+2. 混合操作：允许在文件块内任意顺序组合操作类型
+3. 精准锚点：使用至少3行唯一文本作为锚点
+4. 跨文件操作：可通过## OPERATION:TYPE FILE:path 临时指定其他文件
+
+示例：
+## BEGIN_TCVB
+## FILE:src/app.js
+## OPERATION:REPLACE
+## BEFORE_ANCHOR
+function legacy() {
+  console.log('old');
+## AFTER_ANCHOR
+}
+
+## OLD_CONTENT
+  return 100;
+## NEW_CONTENT
+  return 200;
+
+## OPERATION:INSERT
+## BEFORE_ANCHOR
+// == 配置开始 ==
+## AFTER_ANCHOR
+// == 配置结束 ==
+## INSERT_CONTENT
+  timeout: 3000,
+
+## FILE:README.md
+## OPERATION:DELETE
+## BEFORE_ANCHOR
+<!-- DEPRECATED SECTION -->
+## AFTER_ANCHOR
+<!-- END DEPRECATED -->
+## DELETE_CONTENT
+...旧内容...
+## END_TCVB
+`;
+  }
+}
+
+// ================== 合并函数 ==================
+export function mergeCvb(baseCvb: Cvb, tcvb: TCVB): Cvb {
+  const mergedFiles = new Map<string, string>(Object.entries(baseCvb.getFiles()));
+
+  // 按文件分组操作
+  const operationsByFile = new Map<string, CvbOperation[]>();
+  for (const op of tcvb.getOperations()) {
+    if (!operationsByFile.has(op.filePath)) {
+      operationsByFile.set(op.filePath, []);
+    }
+    operationsByFile.get(op.filePath)!.push(op);
+  }
+
+  // 处理每个文件的修改
+  for (const [filePath, operations] of operationsByFile) {
+    let content = mergedFiles.get(filePath) || '';
+    
+    // 按操作顺序执行修改
+    for (const op of operations) {
+      if (op instanceof ReplaceOperation) {
+        content = applyReplace(content, op);
+      } else if (op instanceof InsertOperation) {
+        content = applyInsert(content, op);
+      } else if (op instanceof DeleteOperation) {
+        content = applyDelete(content, op);
+      }
+    }
+    
+    mergedFiles.set(filePath, content);
+  }
+
+  // 重新生成CVB内容
+  return rebuildCvb(baseCvb, mergedFiles);
+}
+
+function applyReplace(content: string, op: ReplaceOperation): string {
+  const pattern = buildPattern(op.beforeAnchor, op.oldContent, op.afterAnchor);
+  const replacement = `${op.beforeAnchor}${op.newContent}${op.afterAnchor}`;
+  return content.replace(pattern, replacement);
+}
+
+function applyInsert(content: string, op: InsertOperation): string {
+  const pattern = buildPattern(op.beforeAnchor, '', op.afterAnchor);
+  const replacement = `${op.beforeAnchor}${op.content}${op.afterAnchor}`;
+  return content.replace(pattern, replacement);
+}
+
+function applyDelete(content: string, op: DeleteOperation): string {
+  const pattern = buildPattern(op.beforeAnchor, op.oldContent, op.afterAnchor);
+  return content.replace(pattern, `${op.beforeAnchor}${op.afterAnchor}`);
+}
+
+function buildPattern(before: string, content: string, after: string): RegExp {
+  return new RegExp(
+    `${escapeRegExp(before)}${escapeRegExp(content)}${escapeRegExp(after)}`,
+    'gs' // 使用dotall模式匹配换行
+  );
+}
+
+function rebuildCvb(baseCvb: Cvb, files: Map<string, string>): Cvb {
+  let newContent = `## BEGIN_CVB\n## META\n`;
+  
+  // 保留元数据
+  const metadata = baseCvb.getMetadata();
+  for (const [key, value] of Object.entries(metadata)) {
+    newContent += `${key}: ${value}\n`;
+  }
+  newContent += `## END_META\n\n`;
+
+  // 重建文件内容
+  for (const [filePath, content] of files) {
+    const ext = path.extname(filePath).slice(1).toLowerCase();
+    const lang = languageMapping[ext] || 'text';
+    newContent += `## FILE:${filePath}\n\`\`\`${lang}\n${content}\n\`\`\`\n\n`;
+  }
+
+  newContent += `## END_CVB`;
+  const cvb = new Cvb(newContent);
+
+  cvb.setMetaData("时间戳", generateTimestamp());
+  return cvb;
+}
+
+// ================== 工具函数 ==================
+function escapeRegExp(str: string): string {
+  return str.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
+}
+
+function filePathNormalize(rawPath: string): string {
+  return path.normalize(rawPath.replace(/^[\\/]+/, ''));
 }
 
 /**