resume-parser-mcp/src/parser/extractFileText/extractTextFromPdf.ts

To install dependencies:

const fs = require('fs');
const fsPromises = fs.promises;
import pdfParse from 'pdf-parse'; // 导入pdf-parse库
import { useLogger } from '../../utils/logger';

export async function extractTextFromPdf(filePath: string) {
  const logger = useLogger();
  try {
    const dataBuffer = await fsPromises.readFile(filePath);
    const data = await pdfParse(dataBuffer);
    return data.text;
  } catch (error: any) {
    logger.error('PDF文本提取失败:', error?.message);
    throw new Error(`无法提取PDF文本: ${error?.message}`);
  }
}