#308 OpenAI with Python
The basics of using ChatGPT/OpenAI APIs with Python
Notes
Ironically, I asked AI to write some OpenAI examples for python .. and it gave me outdated code that no longer works!
Here’s a quick outline and example of using the openai-python library to make basic Open API calls.
Installation
This example uses the openai python client. Install with pip:
pip install -r requirements.txt
To use the API, an API key is required from https://platform.openai.com/settings/organization/api-keys. Set the defautl environement variable for the following examples to work:
export OPENAI_API_KEY='your-key'
Once installed, the library actually provides a command line tool that can be used to make one-off queries, for example:
$ openai api chat.completions.create -m gpt-4o-mini -g user "Hello, who are you?"
Hello! I’m an AI language model created by OpenAI. I'm here to help answer your questions, provide information, and assist you with a variety of topics. How can I help you today?
Python Usage
The openai-python library provides easy access to the OpenAI aPI. Here’s a basic example:
Create a client with optional configuration details. By default, will use API key from OPENAI_API_KEY environment variable
client = OpenAI()
Call the chat completion API with a prompt, and selected model:
response = client.chat.completions.create(
messages=[
{
"role": "user",
"content": "Hello, who are you?",
}
],
model="gpt-4o-mini"
)
Parse the response and do something with it.
return response.choices[0].message.content.strip()
Note: response data will follow the JSON schema of the API being called. In this case a chat completion object.
Example Run
The chatgpt_query.py is a simple wrapper around the API calls:
$ ./chatgpt_query.py "hello world"
Hello! How can I assist you today?