resume-parser-mcp/src/tool/resume-parser.ts

To install dependencies:

import fs from 'fs';
import path from 'path';
import { Tool } from 'fastmcp';
import { z } from 'zod';
import { downloadFile } from '../parser/downloadFileFromUrl';
import {
  extractTextFromDocx,
  extractTextFromPdf,
} from '../parser/extractFileText';
import { parseTextToResume } from '../parser/parseTextToResume';
import { useLogger } from '../utils/logger';

const fsPromises = fs.promises;
const name = 'resume_parser';
const description = 'Parses a resume and returns a JSON object';
const parameters = z.object({
  resume_url: z.string(),
});
const ResumeParser: Tool<any, z.ZodType<typeof parameters._type>> = {
  name,
  description,
  parameters,
  execute: async ({ resume_url }, { log }) => {
    const logger = useLogger(log);
    let tempFilePath = null;

    try {
      // 1. 下载文件到临时目录
      tempFilePath = await downloadFile(resume_url);
      logger.info('文件下载成功:', tempFilePath);

      // 2. 根据文件扩展名提取文本
      let text = '';
      const fileExtension = path.extname(tempFilePath).toLowerCase();

      if (fileExtension === '.pdf') {
        text = await extractTextFromPdf(tempFilePath);
      } else if (fileExtension === '.docx') {
        text = await extractTextFromDocx(tempFilePath);
      } else {
        throw new Error(
          `不支持的文件格式: ${fileExtension}。请使用PDF或DOCX格式。`
        );
      }

      if (!text) {
        throw new Error('无法从文件中提取文本内容');
      }

      // 3. 解析文本为JSONResume格式
      const resume = parseTextToResume(text);
      return resume;
    } catch (error: any) {
      logger.error('简历解析失败:', error?.message);
      throw error;
    } finally {
      // 4. 清理临时文件
      if (tempFilePath) {
        try {
          await fsPromises.unlink(tempFilePath);
          logger.info('临时文件已清理');
        } catch (cleanupError: any) {
          logger.warn('临时文件清理失败:', cleanupError?.message);
        }
      }
    }
  },
};

export { name, description, parameters, ResumeParser };
export default ResumeParser;