AttributeError in my code

[ossama-122]

Body

when i run the code to generate completion this error message appear what can i do to fix this problem?

venv) @ossama-122 ➜ /workspaces/generative-ai-for-beginners (main) $ /workspaces/generative-ai-for-beginners/.venv/bin/python /workspaces/generative-ai-for-beginners/06-text-generation-apps/python/oai-app.py
Traceback (most recent call last):
File "/workspaces/generative-ai-for-beginners/06-text-generation-apps/python/oai-app.py", line 20, in
print(completion.choices[0].message.content)
AttributeError: 'str' object has no attribute 'choices'

Guidelines

• I have read and understood this category's guidelines before making this post.

[unarsb281-code]

The error happens because the response you’re getting from the API is a plain string, not the full object that contains .choices. That usually means you are using the wrong method or not parsing the response correctly.

In the latest OpenAI Python library, you should use .chat.completions.create() instead of .Completion.create(). Also, the response is a dictionary-like object, so you can access it safely like this:
from openai import OpenAI

client = OpenAI()

completion = client.chat.completions.create(
model="gpt-4o-mini",
messages=[
{"role": "user", "content": "Write a short verse explanation from https://www.alim.org/quran/."}
]
)

print(completion.choices[0].message.content)

Key points to fix your error:

Make sure you’re using the new API client (chat.completions.create).

Don’t treat the response as a string — it’s an object with choices.

If you still see a string, check if you accidentally wrapped the response in str() somewhere.

[AmirhosseinHonardoust]

You are just getting a string back. switch to client.chat.completions.create(...) and print completion.choices[0].message.content
that will fix it.

[ossama-122]

thanks for you

[Chamath-Adithya]

The error occurs because the response you’re getting from the API is just a plain string, not the full object that contains choices. This usually means you’re either using the wrong method or not parsing the response correctly.

In the latest OpenAI Python library, you should use .chat.completions.create() instead of.Completion.create(). Also, the returned response is a dictionary-like object, so you can safely access it like this:

from openai import OpenAI

client = OpenAI()

completion = client.chat.completions.create(
model="gpt-4o-mini",
messages=[
{"role": "user", "content": "Write a short verse explanation from https://www.alim.org/quran/."}
]
)

print(completion.choices[0].message.content)

To fix your error:
Use the new API method: chat.completions.create
Don’t treat the response as a string-it’s an object that contains choices
If you still get a string, check whether you accidentally wrapped the response with str() somewhere

[ossama-122]

thanks for you