使用 Streamlit file_uploader 后出现“TypeError：预期的 str、bytes 或 os.PathLike 对象，而不是 NoneType”

Question

我正在尝试使用streamlit和gpt索引创建某种虚拟测试网站。我不断收到同样的错误。

from gpt_index import GPTListIndex, readers, GPTSimpleVectorIndex, LLMPredictor, PromptHelper, download_loader
from langchain import OpenAI
import logging
import sys
import os
import streamlit as st

os.environ['OPENAI_API_KEY'] = "I am deleting my api key from this post"



fileup = st.file_uploader(label=" ")


SimpleDirectoryReader = download_loader("SimpleDirectoryReader")
loader = SimpleDirectoryReader(fileup)
documents = loader.load_data()
index = GPTSimpleVectorIndex(documents)
index.save_to_disk('index.json')
question = st.text_input("What do you want me to do with the file uploaded?")
response = index.query(question)
st.write(response)

为什么我不断收到“TypeError：预期的 str、bytes 或 os.PathLike 对象，而不是 NoneType”？

这是完整的错误：

Traceback (most recent call last):
  File "/home/appuser/venv/lib/python3.9/site-packages/streamlit/runtime/scriptrunner/script_runner.py", line 565, in _run_script
    exec(code, module.__dict__)
  File "/app/indextest/streamlit_app.py", line 16, in <module>
    loader = SimpleDirectoryReader(fileup)
  File ".modules/file.py", line 67, in __init__
    self.input_dir = Path(input_dir)
  File "/usr/local/lib/python3.9/pathlib.py", line 1082, in __new__
    self = cls._from_parts(args, init=False)
  File "/usr/local/lib/python3.9/pathlib.py", line 707, in _from_parts
    drv, root, parts = self._parse_args(args)
  File "/usr/local/lib/python3.9/pathlib.py", line 691, in _parse_args
    a = os.fspath(a)
TypeError: expected str, bytes or os.PathLike object, not NoneType

Answer 1

根据 Streamlit 文档，

st.file_uploader

要么返回

None

，要么返回

UploadedFileObject

（可以用作“类似文件”的对象）：

https://docs.streamlit.io/library/api-reference/widgets/st.file_uploader

它没有明确提及它，但如果没有上传文件，它可能会返回

None

。我们无法告诉您为什么没有上传文件。

在您的情况下，这意味着

fileup

变量是

None

，这是

SimpleDirectoryReader

的无效参数，如错误消息所示。

文档中的示例包含一个

if

语句，仅当返回值不是

None

时才在脚本中继续。你可能也应该这样做。

Answer 2

fileuploader 返回以下类

class UploadedFile(io.BytesIO):
    """A mutable uploaded file.

    This class extends BytesIO, which has copy-on-write semantics when
    initialized with `bytes`.
    """

    def __init__(self, record: UploadedFileRec, file_urls: FileURLsProto):
        # BytesIO's copy-on-write semantics doesn't seem to be mentioned in
        # the Python docs - possibly because it's a CPython-only optimization
        # and not guaranteed to be in other Python runtimes. But it's detailed
        # here: https://hg.python.org/cpython/rev/79a5fbe2c78f
        super().__init__(record.data)
        self.file_id = record.file_id
        self.name = record.name
        self.type = record.type
        self.size = len(record.data)
        self._file_urls = file_urls

    def __eq__(self, other: object) -> bool:
        if not isinstance(other, UploadedFile):
            return NotImplemented
        return self.file_id == other.file_id

    def __repr__(self) -> str:
        return util.repr_(self)

假设您尝试上传此文件，您首先必须保存该文件，然后上传它，因为它要求路径而不是（io.BytesIO）。

save_path = Path(save_folder, uploaded_file.name)
    with open(save_path, mode='wb') as w:
        w.write(uploaded_file.getvalue())
loader = SimpleDirectoryReader(save_path)

这将在您的系统上创建本地副本，并提供路径作为 SimpleDirectoryReader 函数的输入参数。