Use GPT-3.5 (davinci-003) to extract a bibtex entry from a URL

url-to-bibtex.py

import sys
import openai
import subprocess

# Replace YOUR_API_KEY with your actual OpenAI API key
openai.api_key = "YOUR_API_KEY"

def get_page_contents(url):
  try:
    result = subprocess.run(["lynx", "-dump", url], capture_output=True, check=True)
    content = result.stdout.decode("utf-8", errors="replace")
    return content[:1000]
  except subprocess.CalledProcessError as e:
    print(f"Error: Failed to retrieve content from {url}")
    sys.exit(1)

def generate_bibtex_entry(url, content):
  prompt = f"I want to cite this webpage. Give me a bibtex entry: {url}\n\nHere are the contents of the page:\n\n{content}\n"
  response = openai.Completion.create(
    engine="text—davinci-003",
    prompt=prompt,
    max_tokens=200
  )
  return response.choices[0].text.strip()

if __name__ == "_ main_":
  if len(sys.argv) != 2:
    print("Usage: python url_to_bibtex.py <URL>")
    sys.exit(1)

  url = sys.argv[1]

  content = get_page_contents(url)
  bibtex_entry = generate_bibtex_entry(url, content)
  print(bibtex_entry)

@random_walker / GPT-4

Write a Python script that uses the GPT-3 API to convert a URL into a bibtex entry. Specifically, it should:

• accept a URL as a command line argument
• grab the first 1000 characters of the web page contents
  (not the source code); you can use python libraries for this
  or dump it using the lynx tool
• send the URL and the extracted page contents to the API
  and ask GPT-3 to generate a bibtex entry for it.
• output the result.