AltAPI 1.4.0

A simple and fast ASGI microframework for Python with WebSocket support.

AltAPI

A simple and fast ASGI microframework for Python with WebSocket support.

Changelog

v1.4.0

• Added Dependency Injection system with automatic cleanup:
  • Generator-based dependencies with yield for resource management
  • Nested dependencies (dependencies can depend on other dependencies)
  • Per-request caching of dependencies
  • Async dependency support
  • Automatic Request injection into dependencies
• Added Request State (request.state) for passing data between middleware and handlers
• Rate Limiting optimized — migrated to Shared Memory architecture:
  • Zero network overhead for multi-worker synchronization
  • Fixed race conditions in request counting
  • Memory leak fixes in storage backends
• Added examples:
  • 'examples/webapp.py' - full user managment system
• Fixed cleanup guarantees — finally blocks now always execute even on handler errors
• RedirectResponse returning 303 code, instead of 307

v1.3.0

• Added rate limiting with @rate_limit and @rate_limit_batch decorators
• Added shared manager for multi-worker rate limiting support
• Added support for all HTTP methods: GET, POST, PUT, DELETE, PATCH, HEAD, OPTIONS, TRACE, CONNECT

Installation

pip install altapi

Requirements:

• Python >= 3.10
• uvicorn >= 0.30.0
• anyio >= 4.0.0
• jinja2 >= 3.0.0
• ujson
• Cython >= 3.0.0

Quick Start

Minimal Example

from altapi import AltAPI
from altapi.http import JSONResponse

app = AltAPI()


@app.get("/")
async def home(request):
    return JSONResponse({"message": "Hello, World!"})


if __name__ == "__main__":
    app.run(host="0.0.0.0", port=8000)

Run:

python app.py

Or via uvicorn:

uvicorn app:app --reload

Features

• ✅ ASGI compliant
• ✅ JSON, HTML, and text responses
• ✅ Typed path parameters ({id:int}, {name:str}, {value:float}, {path:path})
• ✅ Sync and async handlers
• ✅ Full WebSocket support
• ✅ Built-in server (app.run()) with multi-worker support
• ✅ Jinja2 templates
• ✅ Response caching with per-worker InMemoryCache
• ✅ Rate limiting with Shared Memory optimization
• ✅ Dependency Injection with automatic cleanup
• ✅ Request State for middleware-to-handler data passing
• ✅ Static file mounting
• ✅ Optimized Cython router
• ✅ GC optimizations for better performance
• ✅ All HTTP methods: GET, POST, PUT, DELETE, PATCH, HEAD, OPTIONS, TRACE, CONNECT

Usage Examples

Path Parameters

from altapi.http import JSONResponse


@app.get("/users/{id:int}")
async def get_user(request):
    user_id = request.path_params["id"]  # automatically int
    return JSONResponse({"id": user_id, "name": f"User {user_id}"})


@app.get("/items/{name:str}")
async def get_item(request):
    name = request.path_params["name"]  # str
    return JSONResponse({"name": name})


@app.get("/files/{path:path}")
async def get_file(request):
    file_path = request.path_params["path"]  # captures full path with slashes
    return JSONResponse({"path": file_path})

Handling POST Requests

@app.post("/api/echo")
async def echo(request):
    data = await request.json()
    return JSONResponse({"echo": data})

WebSocket

from altapi.websocket import WebSocket


@app.websocket("/ws/echo")
async def websocket_echo(ws: WebSocket):
    await ws.accept()
    while True:
        text = await ws.receive_text()
        await ws.send_text(f"Echo: {text}")

Caching

from altapi.caching import cache

app = AltAPI()


@app.get("/api/data")
@cache(expires=3600)  # cache for 1 hour
async def get_data(request):
    return JSONResponse({"data": "cached"})

Rate Limiting

from altapi.ratelimit import rate_limit

app = AltAPI()


@app.get("/api/data")
@rate_limit(limit=10, period=60)  # 10 requests per minute
async def get_data(request):
    return JSONResponse({"data": "value"})


# Multiple limits on same endpoint
from altapi.ratelimit import rate_limit_batch


@app.get("/api/strict")
@rate_limit_batch([
    (10, 60),      # 10 per minute
    (100, 3600),   # 100 per hour
    (1000, 86400)  # 1000 per day
])
async def strict_endpoint(request):
    return JSONResponse({"data": "rate limited"})

Dependency Injection

from altapi import AltAPI
from altapi.http import JSONResponse
from altapi.depends import Depends
import sqlite3


def get_db():
    """Generator-based dependency with automatic cleanup."""
    conn = sqlite3.connect("app.db")
    try:
        yield conn
    finally:
        conn.close()


def get_user_repo(db=Depends(get_db)):
    """Nested dependency."""
    return UserRepository(db)


@app.get("/users")
async def list_users(repo=Depends(get_user_repo)):
    users = repo.get_all()
    return JSONResponse({"users": users})

See full example: examples/database_example.py

Jinja2 Templates

from altapi.templating import Jinja2Templates

app = AltAPI(templates_directory="templates")
templates = Jinja2Templates(directory="templates")


@app.get("/")
async def home(request):
    return templates.TemplateResponse(
        "index.html",
        {"request": request, "title": "Home Page"}
    )

Static Files

# Automatically serves files at /static
app = AltAPI(static_directory="static")

Multi-Worker Server

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=8000, workers=4, access_log=True)

Response Types

from altapi.http import (
    JSONResponse,      # JSON response
    HTMLResponse,      # HTML response
    PlainTextResponse, # Plain text response
    StreamingResponse, # Streaming response
    FileResponse,      # File download with range support
    RedirectResponse,  # Redirect
)

Documentation

Full documentation is available in DOCS.md.

License

AGPLv3, see LICENSE.txt for details.