Skip to main content

An IPython extension for interacting with OpenAI's models

Project description

IPython + GPT-4

A quick implementation of a random idea: what if ipython had a magic function that sent the current session to gpt-4 along with a prompt, and automatically evaluated the result in the current session?

Installation

pip install ipython_openai

Then, you can either load the extension during an ipython session with %load_ext ipython_openai or add the following to your ipython config.

c.InteractiveShellApp.extensions = ['ipython_openai']

Configuration

Your OpenAI API key should be available in the OPENAI_API_KEY env var. You can configure the model used by adding an entry to your ipython config: c.GPT.model_name = 'gpt-4-turbo-preview'.

Usage

This extension adds two new "magic" functions documented below. Each one captures the current history of the ipython session and sends it to the OpenAI API along with the prompt given after the command. The result is streamed back to the session. Each invocation of the magic functions can also see the output from previous ones.

gen

In [1]: gen a 5 line greeter function with randomness
import random
def greet():
    names = ["Alice", "Bob", "Charlie", "Diana", "Edward"]
    greeting = ["Hello", "Hi", "Hey", "Greetings", "What's up"]
    print(f"{random.choice(greeting)}, {random.choice(names)}!")

Run it? (y/n) y

In [2]: greet()
Hello, Bob!

ask

In [3]: ask what happened
When the `greet()` function was called, the following occurred:
1. The function selected a random name from the list `names`, which contains five different names: "Alice", "Bob", "Charlie", "Diana", and "Edward".
2. It also selected a random greeting from the list `greeting`, which includes five different greetings: "Hello", "Hi", "Hey", "Greetings", and "What's up".
3. It combined the selected greeting and name with a formatted string to create a message.
4. Finally, the function printed this message to the console.
Since the selections are random, the exact output can vary with each call to `greet()`, displaying a greeting and a name randomly chosen from the lists provided.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ipython_openai-0.0.5.tar.gz (4.6 kB view details)

Uploaded Source

Built Distribution

ipython_openai-0.0.5-py3-none-any.whl (5.2 kB view details)

Uploaded Python 3

File details

Details for the file ipython_openai-0.0.5.tar.gz.

File metadata

  • Download URL: ipython_openai-0.0.5.tar.gz
  • Upload date:
  • Size: 4.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.0.0 CPython/3.9.14

File hashes

Hashes for ipython_openai-0.0.5.tar.gz
Algorithm Hash digest
SHA256 c083117f2954360da54a66e8d6e787aa84bb92d138e6f1b354c3f3f80a7e5514
MD5 a25f738c12053655882c99606da276fe
BLAKE2b-256 31ca12ba3ed55a6c9881f6936e97d8a8251d9f6c4f823b9d29892581a253c22c

See more details on using hashes here.

File details

Details for the file ipython_openai-0.0.5-py3-none-any.whl.

File metadata

File hashes

Hashes for ipython_openai-0.0.5-py3-none-any.whl
Algorithm Hash digest
SHA256 be521641ceb578449cc0c73b642c4e4be9fb29c54ae0862ae185c7ab285959b0
MD5 d16dd513efa5307dad50c52f106baa8b
BLAKE2b-256 253a22ae2cd1c169c8399c1668af7cb04d1a6138827561eeca9eebdf63e19006

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page