​Getting Started with GPT APIs: A Step-by-Step Guide

Image generated by Midjourney based on my prompt

Are you fascinated by the world of AI and machine learning? If so, you’re likely familiar with the Generative Pretrained Transformer model, GPT, which is the state-of-the-art offering from OpenAI, a leading Large Language Model (LLM). If you’re interested in delving deep into these technologies, I highly recommend reading my article “The Generative AI Revolution — A Primer.” It provides an extensive exploration of GPT, other models, and key related concepts.

The GPT model is renowned for its ability to generate text that is so human-like it can easily be mistaken for human-produced content. In this blog post, my goal is to help you understand how to use OpenAI APIs by walking you through a simple API call implementation.

We’ll begin by setting up the Python environment and discussing how to configure your OpenAI account. And to add a touch of enjoyment, we’ll also have some fun by writing code. Let’s dive right in!

Setting Up the Python Environment

Before you interact with the OpenAI API, you’ll need to set up a Python environment, you can choose your preferred language:

1. Install Python: First, you’ll need to have Python installed on your computer. You can download it from the official Python website. Python 3.6 or later is recommended.
2. Install OpenAI package: OpenAI provides a Python client library for their API. It can be installed via pip (an acronym of “pip Install Packages”) is today the standard tool for installing Python packages and their dependencies. Most recent distributions of Python come with pip preinstalled. Open your command prompt or terminal and run the following command:

$ pip install openai 

This command installs the `openai` Python library which provides a Python interface to the OpenAI API — if you don’t find ‘pip’ in your setup, try ‘pip3’ instead. This library needs to be imported inside your code.

Configuring Your OpenAI Account

To use OpenAI’s API, you’ll first need to sign up for an account on their website. Once you’ve signed up and logged in, you can get your API key from the OpenAI interface. Remember, a paid membership to ChatGPT is designed for end users who interact with the AI model directly whereas a paid membership for the OpenAI API is typically for developers or businesses who want to integrate AI capabilities into their applications — so two different memberships.

Here are the steps to get your API key:

1. After signing into the OpenAI website, click on the “View API keys” under your profile picture.

2. Here, you will see your API key. This is the key that you will use in your Python code to authenticate your requests.

Remember to keep your API key safe and don’t share it with anyone. Your API key is linked to your account and will be used to track your API usage and billing.

N.B.: OpenAI’s API is not free, and they charge based on the number of tokens processed by your requests.

Writing Code to Interact with the OpenAI API

Once your environment is set up, you can start writing Python code to interact with the OpenAI API. Here is a very simple sample code snippet:

import os
import openai
# Set your API Key
openai.api_key = 'your-api-key'
prompt = "Boston is the capital of"
# Generate text
response = openai.Completion.create(
  model="text-davinci-003",
  prompt=prompt,
  temperature=0,
  max_tokens=100,
  top_p=1,
  frequency_penalty=0.0,
  presence_penalty=0.0,
  stop=["\n"]
)
print(response.choices[0].text.strip())

In this simple example, this is where the text is actually generated. The openai.Completion.create() method is called with several arguments:

• model: The ID of the language model to use. In this case, it’s “text-davinci-003”, which is a powerful language model developed by OpenAI.
• prompt: The text that the language model should generate a continuation for.
• temperature: This controls the randomness of the model’s output. A temperature of 0 makes the output deterministic, while higher temperatures make the output more random.
• max_tokens: The maximum length of the generated text, in tokens. A token can be as short as one character or as long as one word.
• top_p: This parameter, also known as nucleus sampling, controls the diversity of the model’s output. It takes a value between 0 and 1. In this case, it’s set to 1, meaning the model can choose from the full set of possible tokens at each step.
• frequency_penalty and presence_penalty: These parameters control the penalties for using frequent or infrequent words, respectively.
• stop: This is a list of tokens at which the model should stop generating text. In this case, it’s set to stop generating text at the first newline character.

Remember to replace `’your-api-key’` with your actual API key, which you will get when you sign up for the OpenAI API.

At this point, you should be good to run that Python script above using your preferred IDE or other tool like a simple curl command as seen below. The curl command would look something like this:

curl https://api.openai.com/v1/engines/text-davinci-003/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your-api-key" \
-d '{
   "prompt": "Boston is the capital of",
   "max_tokens": 100,
   "temperature": 0,
   "top_p": 1,
   "frequency_penalty": 0.0,
   "presence_penalty": 0.0,
   "stop": ["\n"]
}'

Chat completions response format

In the example above chat completions API response looks as follows:

{
 "id": "cmpl-7Wz4HDP38eV4oxcpFrjPRlbur6ouI",
 "object": "text_completion",
 "created": 1688095597,
 "model": "text-davinci-003",
 "choices": [
 {
 "text": " Massachusetts",
 "index": 0,
 "logprobs": null,
 "finish_reason": "stop"
 }
 ],
 "usage": {
 "prompt_tokens": 5,
 "completion_tokens": 1,
 "total_tokens": 6
 }
}

In Python, the output can be extracted with response.choices[0].text.strip() as seen in the last line of the prompt code. This returns “Massachusetts” based on the above.

Please note: You should always read and adhere to OpenAI’s use case policy before integrating the API into your applications.

You’re done!

Final Thoughts

OpenAI’s GPT-X is a powerful tool that opens up new opportunities for creating innovative and smart applications. With the right setup and coding practices, these tools can be effectively harnessed. They are transforming the landscape of coding by enabling AI-driven code generation and analysis. The future holds even more exciting prospects like automated test generation and intelligent code explanations. The future of coding is promising and filled with anticipation for the new developments to come.

We hope this simple guide has been helpful as you embark on your journey with AI-powered code development. Happy coding!