使用pdf2docx转换PDF至DOCX失败，求助无本地存储解决方案

错误原因 #

你遇到的TypeError: bad filename是因为pdf2docx.Converter（底层依赖PyMuPDF的fitz.Document）无法直接处理原始字节串。代码中cv = Converter(pdf_stream.read())传入的是pdf_stream.read()返回的字节内容，而它需要的是文件路径、已打开的文件对象或BytesIO流对象。

修复步骤 #

1. 修正Converter的参数 #

直接传入BytesIO对象而非字节串，无需调用read()：

# 原错误代码
cv = Converter(pdf_stream.read())

# 修改后代码
cv = Converter(pdf_stream)

注意：刚创建的BytesIO(response.content)默认文件指针在起始位置，无需额外调整。

2. 可选优化建议 #

• 改用异步HTTP客户端：在FastAPI异步函数中使用同步requests.get会阻塞事件循环，换成aiohttp提升性能：

  import aiohttp
  
  # 替换原requests.get逻辑
  async with aiohttp.ClientSession() as session:
      async with session.get(urls) as response:
          pdf_content = await response.read()
          pdf_stream = BytesIO(pdf_content)
  

• 提前初始化Firebase：不要在每次请求中重复初始化Firebase应用，将初始化代码放在函数外部：

  # 文件顶部、函数外执行
  firebase = pyrebase.initialize_app(firebase_config)
  storage = firebase.storage()
  

• 无需手动关闭BytesIO：内存中的BytesIO对象无需手动调用close()，Python会自动回收资源。

修改后的完整代码 #

from fastapi import Request
from pydantic import BaseModel
from typing import Union, List, Dict
import os
from io import BytesIO
import aiohttp
import pyrebase
from pdf2docx import Converter

# 提前初始化Firebase应用
firebase_config = {}  # 填入你的Firebase配置
firebase = pyrebase.initialize_app(firebase_config)
storage = firebase.storage()

class PDFManipulationRequest(BaseModel):
    urls: Union[str, List[str]]
    user: Dict = {}
    rename_file: str = ""
    pages: List[int] = []
    rotation: Dict = {}
    password: str = ""
    metadata: Dict

async def pdf_to_word(request: Request, data: PDFManipulationRequest):
    urls = data.urls
    metadata = data.metadata
    filename = metadata["name"]
    filename_without_extension = os.path.splitext(filename)[0]

    # 异步下载PDF文件
    async with aiohttp.ClientSession() as session:
        async with session.get(urls) as response:
            pdf_content = await response.read()
            pdf_stream = BytesIO(pdf_content)

    # 初始化转换器，传入BytesIO流对象
    cv = Converter(pdf_stream)
    output_doc_stream = BytesIO()

    cv.convert(output_doc_stream, start=0, end=None)
    cv.close()

    # 设置Firebase存储路径
    folder_name = "pdf-to-word/modified"
    destination_path = f"{folder_name}/{filename_without_extension}.docx"

    # 上传转换后的DOCX文件
    storage.child(destination_path).put(output_doc_stream.getvalue())

    # 获取文件下载URL
    docx_url = storage.child(destination_path).get_url(None)

    return docx_url

内容的提问来源于stack exchange，提问作者Rayyan Alam