asyncPyGithub

A fully asynchronous Python library for the GitHub API.

Install

pip

pip install asyncPyGithub

Manual

git clone https://github.com/sudoDeVinci/asyncPyGithub.git
cd asyncPyGithub
pip install -r requirements.txt

Setup

Create a .env file:

GITHUB_TOKEN=your_token_here

Client Lifecycle

The library uses a shared httpx.AsyncClient under the hood. You have two options for managing it:

Option 1: Let it auto-start (simple)

The client starts automatically on first request. Just remember to close it when done:

import asyncio
from asyncPyGithub import GitHubPortal, GitHubUserPortal

async def main():
    # Client starts automatically on first request
    status, user = await GitHubPortal.authenticate("your_token")
    if status != 200:
        print(f"Auth failed: {user.message}")
        return
    
    print(f"Logged in as {user.login}")
    
    # Do stuff...
    status, other = await GitHubUserPortal.get_by_username("torvalds")
    if status == 200:
        print(f"Found: {other.login}")
    
    # Clean up when done
    await GitHubPortal.close()

asyncio.run(main())

Option 2: Use the context manager (scoped)

For scoped usage where you want automatic cleanup:

import asyncio
from asyncPyGithub import GitHubPortal, GitHubUserPortal

async def main():
    async with GitHubPortal.scoped_client():
        status, user = await GitHubPortal.authenticate("your_token")
        if status != 200:
            return
        
        status, other = await GitHubUserPortal.get_by_username("torvalds")
        # Client closes automatically when exiting the context

asyncio.run(main())

Concurrent Requests

The client uses connection pooling, so concurrent requests share connections efficiently:

import asyncio
from asyncPyGithub import GitHubPortal, GitHubUserPortal, GitHubRepositoryPortal

async def main():
    status, user = await GitHubPortal.authenticate("your_token")
    if status != 200:
        return
    
    # All requests run concurrently over the shared client
    results = await asyncio.gather(
        GitHubUserPortal.get_by_id(user.id),
        GitHubUserPortal.get_by_username(user.login),
        GitHubUserPortal.all(since=0, per_page=5),
        GitHubRepositoryPortal.get_organization_repos("LEGO", per_page=5),
    )
    
    for status, data in results:
        if status == 200:
            print(f"OK: {type(data).__name__}")
        else:
            print(f"Error: {data.message}")
    
    await GitHubPortal.close()

asyncio.run(main())

API

Every method returns tuple[int, Result | ErrorMessage]. Check the status code first.

GitHubPortal

Method	What it does
authenticate(token)	Auth and get your user info. Starts client if needed.
start()	Manually start the HTTP client
close()	Close the HTTP client
scoped_client()	Context manager that auto-closes on exit

GitHubUserPortal

Method	What it does
update(changes)	Update your profile
get_by_id(uid)	Get user by ID
get_by_username(username)	Get user by username
all(since, per_page)	List users
get_hovercard(username)	Get hovercard info

GitHubRepositoryPortal

Method	What it does
get_organization_repos(org, ...)	List org repos
create_organization_repo(org, name, ...)	Create repo in org
get_user_repo(owner, repo)	Get a specific repo
update_repository(owner, repo, ...)	Update repo settings
delete_repository(owner, repo)	Delete a repo
list_contributors(owner, repo)	List contributors
list_repository_languages(owner, repo)	Get language breakdown
list_repository_tags(owner, repo)	List tags
get_repository_topics(owner, repo)	Get topics

Error Handling

status, result = await GitHubUserPortal.get_by_username("doesnt-exist")

if status == 200:
    print(result.login)
else:
    # result is an ErrorMessage
    print(f"Error {result.code}: {result.message}")

Structure

GitHubPortal              # Base - auth, client management
├── GitHubUserPortal      # /user and /users endpoints
└── GitHubRepositoryPortal # /repos and /orgs/.../repos endpoints

All responses are Pydantic models (PrivateUser, SimpleUser, MinimalRepository, etc.).

Testing

pip install -r requirements-dev.txt
pytest

Limitations

• No rate limit handling
• No pagination helpers
• No webhooks