Accessing OpenAI’s API Using R and httr2

A simple code to get started with OpenAI’s API

In this blog post, I will guide you through the process of accessing OpenAI’s GPT-4 API using R and the httr2 package. By the end, you’ll be able to make requests to the API and retrieve responses programmatically.

Prerequisites

Before we start, make sure you have done the following:

1. Add Billing Information: Ensure your OpenAI account has billing information added. You can do this here.
2. Create Project API Keys: Generate your API keys from OpenAI. You can do this here.

Useful Links

• OpenAI Models
• API Key Usage

Step-by-Step Instructions

Let’s walk through the code to make an API request using R.

Step 1: Install and Load the httr2 Package

First, ensure you have the httr2 package installed and loaded in your R environment:

install.packages('httr2')
library(httr2)

Step 2: Set Up Your API Key

Make sure to set up your API key as an environment variable named OPENAI_API_KEY. You can do this by adding the following line to your .Renviron file or setting it directly in your R session:

Sys.setenv(OPENAI_API_KEY = 'YOUR_API_KEY_HERE')

Step 3: Create the Request Function

Here is a function that constructs and sends a request to the GPT-4 API. The function takes a prompt as input and returns the response from the API.

# Request to ChatGPT.
chatgpt_request <- function(prompt) {
  
  if(!is.character(prompt)) stop('The prompt must be a character string.')
  
  api_key <- Sys.getenv('OPENAI_API_KEY')
  
  # Create the request
  request <- request('https://api.openai.com/v1/chat/completions') |>
    req_headers(
      Authorization = paste('Bearer', api_key),
      `Content-Type` = 'application/json'
    ) |>
    req_body_json(list(
      model = 'gpt-4o-2024-05-13',
      messages = list(list(role = 'user', content = prompt)),
      max_tokens = 150
    ))
  
  # Perform the request
  response <- request |>
    req_perform()
  
  # Parse and return the response
  resp_body_json(response)
}

Step 4: Make a Request

Use the function to make a request to the API. For example:

prompt <- 'Please create a poem about R, OpenAI, API, and httr2'
response <- chatgpt_request(prompt)
cat(response$choices[[1]]$message$content)

This code will send the prompt ‘Please create a poem about R, OpenAI, API, and httr2’ to the GPT-4 API and print the response.

Here is the response!

In the land of data, vast and wide,
Where secrets in the numbers hide,
A tool arose with power great,
To analyze, to innovate.

R, the language of the wise,
With syntax clear, it did surprise.
From stats to plots, it paved the way,
For insights fresh, both night and day.

But seeking more, the coders yearn,
For AI’s magic, they discern.
OpenAI, with knowledge bright,
Brought forth APIs, pure delight.

With queries swift, and answers neat,
The world of data felt complete.
Conversing with a bot so smart,
Transforming data, artful heart.

Yet bridging these, a task in hand,
To fetch, to send, at our command.
httr.

Conclusion

By following these steps, you can easily access OpenAI’s GPT-4 API using R and the httr2 package. This allows you to integrate powerful language model capabilities into your R applications. For more details on the models and API usage, refer to the OpenAI Models and API Key Usage documentation.