higuma¶

Write in Python, serve with Rust.

python -m pip install -U higuma

from higuma import Higuma

app = Higuma(__name__)


@app.get("/")
def index():
    return {"message": "Hello from higuma"}


app.run()

Design¶

higuma combines a Flask-inspired decorator API with a built-in HTTP server powered by axum and Tokio. Application logic stays in Python; the Rust core handles sockets, route matching, file streaming, gzip, and WebSockets.

higuma is neither a drop-in Flask replacement nor an ASGI application passed to a separate ASGI server. Existing WSGI and ASGI applications can be mounted under a higuma route.

Features in 0.4.0¶

Annotated typed inputs, dependency injection, response models, and structured 422s
Typed converters, automatic HEAD/OPTIONS, sync/async handlers, and class-based views
Blueprints, middleware, request/response hooks, and lifespan
MiniJinja SSR, context processors, and a compiled template cache
Sync/async streaming, SSE, background tasks, range/ETag files, and automatic gzip
Text, binary, and JSON WebSockets with same-origin and auth preflight checks
Multipart uploads and a safe filename helper
OpenAPI 3.1 / Swagger UI schemas for dataclasses, TypedDicts, enums, and more
SQLite ORM with transactions and pagination
Signed sessions, login, roles/permissions, and OAuth 2.0 with PKCE
Password, token, CSRF, CORS, rate-limit, trusted-host, and proxy helpers
In-process testing, CLI commands, and a multi-process supervisor

OpenAPI and input validation

Markers such as Annotated[..., Body()] validate and convert runtime input and generate OpenAPI from the same declaration. response_model validates output. The legacy request_body= option and a return annotation by itself remain schema metadata.

Supported scope¶

CPython 3.10 or newer
ABI3 wheels for Windows, Linux, and macOS
The built-in ORM supports SQLite only
A reverse proxy is recommended for TLS, buffering, and network-level limits

Continue¶

Warning

higuma is under active development. Run workload-specific load, failure, and security tests before adopting it in production.