Using Assistant + Completions API

API
,
1

I’m trying to make a file_search assistant and use chat completions to allow interaction with the assistant. I have the following code, but am receiving an error I don’t quite understand.

from openai import OpenAI 

client = OpenAI(api_key='xxx')

assistant = client.beta.assistants.create(
    name = "CDA File Reader",
    instructions = "Read the instructions at the top of the file. Be cordial and conversational.", 
    model = "gpt-4o", 
    tools = [{"type": "file_search"}],
)

vector_store = client.beta.vector_stores.create(name = "Training Documents")

file_paths = ["/Users/kishanpatel/Library/CloudStorage/OneDrive-CHAMPSSoftwareInc/CDA/test.txt"]
file_streams = [open(path, "rb") for path in file_paths]

file_batch = client.beta.vector_stores.file_batches.upload_and_poll(
    vector_store_id = vector_store.id, files = file_streams
)

print(file_batch.status)
print(file_batch.file_counts) 

assistant = client.beta.assistants.update(
    assistant_id = assistant.id, 
    tool_resources = {"file_search": {"vector_store_ids": [vector_store.id]}}
)

def chat_with_assistant(user_input): 
    chat_message = [
        {"role": "system", "content": "You are a data assistant. Read the information at the top of the provided information, and respond accordingly."},
        {"role": "user", "content": user_input}
    ]

    response = client.chat.completions.create(
        model="gpt-4",  
        messages=chat_message,
        tools=[{"type": "file_search", "resource_id": vector_store.id}]
    )

    assistant_reply = response.choices[0].message.content
    return assistant_reply

user_input = input("Input: ") 
assistant_response = chat_with_assistant(user_input) 
print("Output: " + assistant_response)

The error that I'm getting is this: 

Traceback (most recent call last):
  File "/Users/kishanpatel/Library/CloudStorage/OneDrive-CHAMPSSoftwareInc/CDA/testing.py", line 6, in <module>
    assistant = client.beta.assistants.create(
        name="CDA File Reader",
    ...<7 lines>...
        }
    )
  File "/Library/Frameworks/Python.framework/Versions/3.13/lib/python3.13/site-packages/openai/resources/beta/assistants.py", line 146, in create
    return self._post(
           ~~~~~~~~~~^
        "/assistants",
        ^^^^^^^^^^^^^^
    ...<18 lines>...
        cast_to=Assistant,
        ^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/Library/Frameworks/Python.framework/Versions/3.13/lib/python3.13/site-packages/openai/_base_client.py", line 1278, in post
    return cast(ResponseT, self.request(cast_to, opts, stream=stream, stream_cls=stream_cls))
                           ~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Library/Frameworks/Python.framework/Versions/3.13/lib/python3.13/site-packages/openai/_base_client.py", line 955, in request
    return self._request(
           ~~~~~~~~~~~~~^
        cast_to=cast_to,
        ^^^^^^^^^^^^^^^^
    ...<3 lines>...
        retries_taken=retries_taken,
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    )
    ^
  File "/Library/Frameworks/Python.framework/Versions/3.13/lib/python3.13/site-packages/openai/_base_client.py", line 1059, in _request
    raise self._make_status_error_from_response(err.response) from None
openai.BadRequestError: Error code: 400 - {'error': {'message': "Missing required parameter: 'tools[0].function'.", 'type': 'invalid_request_error', 'param': 'tools[0].function', 'code': 'missing_required_parameter'}}

How can I get around this? Any help is appreciated. Thank you!

2

You do not interact with an assistant through chat completions.

You must create a thread with messages.

Then you initiate a run, using the thread ID and assistant ID.

Then wait for the status to change to completed.

Then retrieve the latest message produced into the thread.

3

  1. don’t create an Assistant in your code - you only need to create it once.

  2. Check out this part of the docs:
    https://platform.openai.com/docs/api-reference/assistants/createAssistant
    and check the tools section.
    and then as @arata says - you would follow up creating a thread and messages and a run.

There are a lot of examples out there to use. I wrote a fastapi version. GitHub - jlvanhulst/fastapi-assistant