Build a Stateful Agent with the OpenAI Assistants API

1. 1
   Set Up Your Environment

   Install the OpenAI Python library and configure your API key as an environment variable. The client will automatically use the `OPENAI_API_KEY` variable.
2. 2
   Create an Assistant

   Define your assistant's behavior. Specify instructions, a model, and enable tools. Here, we create a 'Math Tutor' that uses the Code Interpreter tool to solve math problems.
3. 3
   Create a Conversation Thread

   A Thread represents a single conversation and stores its message history. Create a new, empty thread for each user conversation you want to start.
4. 4
   Add a Message and Run the Assistant

   Add the user's message to the thread. Then, create a 'Run' to process the entire thread. The OpenAI Assistants API will execute the necessary steps, like calling the Code Interpreter.
5. 5
   Check Run Status & Get the Response

   Runs are asynchronous. Poll the run's status until it is 'completed'. Once complete, you can retrieve the list of messages from the thread, which will now include the assistant's response.

import os
import time
from openai import OpenAI

# Ensure your OPENAI_API_KEY is set as an environment variable
client = OpenAI()

# 1. Create an Assistant
assistant = client.beta.assistants.create(
    name="Math Tutor",
    instructions="You are a personal math tutor. Write and run Python code to answer math questions.",
    tools=[{"type": "code_interpreter"}],
    model="gpt-4o",
)

# 2. Create a Thread
thread = client.beta.threads.create()

# 3. Add a Message to the Thread
message = client.beta.threads.messages.create(
    thread_id=thread.id,
    role="user",
    content="What is 9 factorial? Show your work."
)

# 4. Create a Run
run = client.beta.threads.runs.create(
    thread_id=thread.id,
    assistant_id=assistant.id,
    instructions="Please address the user as Jane Doe. The user has a premium account."
)

print(f"Run created with ID: {run.id}")

# 5. Poll for the Run to complete
while run.status in ['queued', 'in_progress', 'cancelling']:
    time.sleep(1) # Wait for 1 second
    run = client.beta.threads.runs.retrieve(
        thread_id=thread.id,
        run_id=run.id
    )
    print(f"Current run status: {run.status}")

# 6. Retrieve and print the Assistant's response
if run.status == 'completed':
    messages = client.beta.threads.messages.list(
        thread_id=thread.id
    )
    # The latest message is at the beginning of the list
    assistant_message = messages.data[0].content[0].text.value
    print("\n--- Assistant's Response ---")
    print(assistant_message)
else:
    print(f"Run failed with status: {run.status}")