Not support doc format

[Panweitong]

Confirm this is an issue with the Python library and not an underlying OpenAI API

• This is an issue with the Python library

Describe the bug

OpenAI Docs:

Python library:

To Reproduce

Use Files API to Upload doc file，and use "assistants" for Assistants

Code snippets

import openai,io
import requests

openai.api_key = "xxxxxxxxxxxxxxxxxxx"

class FileLike(io.BytesIO):
  def __init__(self, _bytes, filename=None):
    super().__init__(_bytes)
    self.name = filename

url = "https://ccrb.s3.cn-northwest-1.amazonaws.com.cn/%E7%BB%B3%E8%88%9E%E9%A3%9E%E6%89%AC%E6%B4%BB%E5%8A%9B%E7%BB%BD%E6%94%BE.doc"

r = requests.get(url)
fileName = url.split("https://ccrb.s3.cn-northwest-1.amazonaws.com.cn/")[1]

bytes_io = io.BytesIO(r.content)
file_bytes = bytes_io.read()

res = openai.files.create(
  file=FileLike(file_bytes, fileName), purpose="assistants"
)
if res.id and res.status == "processed":
  file = openai.files.retrieve(res.id)
  print(file)

OS

Ubuntu

Python version

Python v3.10.12

Library version

openai v1.51.0

[RobertCraigie]

Thanks for the report, can you share an example snippet to reproduce the issue?

[Panweitong]

Thanks for the report, can you share an example snippet to reproduce the issue?

OK，I will share an example snippet later.

[Panweitong]

Thanks for the report, can you share an example snippet to reproduce the issue?

import openai,io
import requests

openai.api_key = "xxxxxxxxxxxxxxxxxxx"

class FileLike(io.BytesIO):
  def __init__(self, _bytes, filename=None):
    super().__init__(_bytes)
    self.name = filename

url = "https://ccrb.s3.cn-northwest-1.amazonaws.com.cn/%E7%BB%B3%E8%88%9E%E9%A3%9E%E6%89%AC%E6%B4%BB%E5%8A%9B%E7%BB%BD%E6%94%BE.doc"

r = requests.get(url)
fileName = url.split("https://ccrb.s3.cn-northwest-1.amazonaws.com.cn/")[1]

bytes_io = io.BytesIO(r.content)
file_bytes = bytes_io.read()

res = openai.files.create(
  file=FileLike(file_bytes, fileName), purpose="assistants"
)
if res.id and res.status == "processed":
  file = openai.files.retrieve(res.id)
  print(file)

[Panweitong]

Thanks for the report, can you share an example snippet to reproduce the issue?

I tried to test it with the API, but it still reported an error, but the documentation says it supports doc format

[ganeshhnalla]

@Panweitong OpenAI accepts files of various formats (e.g., .txt, .csv, .json, .pdf, .docx, .doc, etc.), but it's always good to ensure that the version of the API you're using indeed supports doc files.

If you're attempting to use doc files, ensure the content is properly extracted, as doc is a binary format, and OpenAI may have trouble directly interpreting the binary content.

[RobertCraigie]

Really sorry for the delayed response. As this is an issue with the underlying OpenAI API and not the SDK, I'm going to go ahead and close this issue.

Would you mind reposting at community.openai.com?