+
Skip to content

justmagici/capebase

Repository files navigation

CapeBase Logo

Enhance your Python backend with auto-generated APIs, real-time features, and granular permissions—all with minimal effort.


PyPI version License: MIT

🚧 Note: CapeBase is in early development - expect some API and interface tweaks along the way as we work towards a stable release!

Overview

CapeBase is a lightweight yet powerful Python library that transforms your backend development. It provides real-time capabilities, auto-generated REST APIs, and fine-grained data access controls—all with minimal setup. Designed for small to medium projects, CapeBase is database-agnostic, supporting SQLite, PostgreSQL, and MySQL while seamlessly integrating with FastAPI and SQLModel.

✨ Features

  • 🔌 FastAPI Integration: First-class support for FastAPI.
  • 🔄 Database Agnostic: Works seamlessly with SQLite, PostgreSQL, and MySQL
  • 🚀 Automatic API Generation: Instantly create RESTful APIs from your SQLModels.
  • Real-time Database: Subscribe to database changes in real-time
  • 🔐 Authentication: Easily plug in your own solution or use the built-in system (WIP)
  • 🛡️ Data Access Control: Role-based and resource-level permissions with dynamic filtering
  • 📊 Observability: API endpoint monitoring with Prometheus and comprehensive audit trail (WIP)

🚀 Quick Start

Installation

pip install capebase

Basic Usage

  1. Define your SQLModel
from datetime import datetime
from typing import Optional
from sqlmodel import SQLModel, Field
from capebase import FROM_AUTH_ID

class Todo(SQLModel, table=True):
    """Model for todo items"""
    id: Optional[int] = Field(default=None, primary_key=True)
    user_id: FROM_AUTH_ID = Field(index=True)  # Automatically populated from auth context
    task: str
    is_complete: bool = Field(default=False)
    inserted_at: datetime = Field(default_factory=datetime.utcnow)
  1. Set up your FastAPI app with CapeBase
from fastapi import FastAPI, Depends
from capebase import CapeBase

app = FastAPI(title="Real-time Todo App")
cape = CapeBase(
    app=app,
    db_path="sqlite+aiosqlite:///todos.db",
    auth_provider=auth_provider,  # Your auth provider
)

# Register model with Cape - automatically generates CRUD and subscription routes
cape.publish(Todo)  # Basic usage with default schemas

# Or with optional custom schemas for more control


cape.publish(
    Todo,
    create_schema=CreateTodo,  # Optional: Custom schema for creation
    update_schema=UpdateTodo   # Optional: Custom schema for updates
)

# Set up permissions
cape.permission_required(
    Todo, role="*", actions=["read"],
)  # Allow anyone to read todos (public read access)

cape.permission_required(
    Todo, owner_field="user_id", actions=["create", "update", "delete"]
)  # Users can only create/update/delete their own todos

# Example: Admin permissions (optional)
cape.permission_required(
    Todo, role="admin", actions=["read", "create", "update", "delete"]
)  # Admins have full access to all todos
  1. Define custom routes (optional)
from sqlalchemy import select
from typing import Sequence

@app.get("/todo/filter")
async def filter(
    completed: bool, 
    session=Depends(cape.get_db_dependency())
) -> Sequence[Todo]:
    """Filter todos by completion status"""
    result = await session.execute(
        select(Todo).where(Todo.is_complete == completed)
    )
    return result.scalars().all()
  1. Subscribe to real-time changes (optional)
@cape.subscribe(Todo)
async def on_todo_change(change):
    """Handle todo changes"""
    if change.event == "INSERT":
        print(f"New todo created: {change.payload.task}")
    elif change.event == "UPDATE":
        print(f"Todo updated: {change.payload.task}")
    elif change.event == "DELETE":
        print(f"Todo deleted: {change.payload.task}")

    # Use privileged session for unrestricted access
    async with cape.get_privileged_session() as session:
        result = await session.execute(
            select(Todo).where(Todo.user_id == change.payload.user_id)
        )
        print(f"User has created {len(result.scalars().all())} todos")

You can find the complete working code for this todo application in the examples/ directory.

Generated API Routes

The cape.publish(Todo) command automatically creates the following endpoints:

  • GET /todo - List todos
  • GET /todo/{id} - Get a specific todo
  • POST /todo - Create a new todo
  • PATCH /todo/{id} - Update a todo
  • DELETE /todo/{id} - Delete a todo
  • GET /todo/subscribe - Server-Sent Events (SSE) endpoint for real-time updates
  • (WIP) WS /todo/subscribe - WebSocket endpoint for real-time updates

You can also specify which routes to generate:

cape.publish(Todo, routes=["list", "get", "create"])

Custom Schemas (Optional)

You can customize which fields are allowed in create/update operations:

# Define custom schema for creation
class CreateTodo(SQLModel):
    task: str  # Required field
    is_complete: bool = Field(default=False)  # Optional with default

# Define custom schema for updates
class UpdateTodo(SQLModel):
    is_complete: bool = Field(default=False)  # Only allow updating completion status

# Register model with custom schemas
cape.publish(
    Todo,
    create_schema=CreateTodo,   # Control fields allowed during creation
    update_schema=UpdateTodo    # Control fields allowed during updates
)

This will generate:

  • POST /todo - Uses CreateTodo schema
  • PATCH /todo/{id} - Uses UpdateTodo schema
  • Other endpoints use the base Todo model

Authentication

Auth Context Provider

CapeBase uses an auth context provider to handle authentication and user context. You can implement your own provider based on your authentication needs. The provider should return an AuthContext object containing:

from capebase.auth import AuthContext

async def get_auth_context(request: Request) -> AuthContext:
    # Your authentication logic here
    return AuthContext(
        id="user_id",        # Required: Unique identifier for the user
        role="user_role",    # Optional: Role for permission checks
        context={}           # Optional: Additional context data
    )

cape = CapeBase(
    app=app,
    db_path="sqlite+aiosqlite:///app.db",
    get_auth_context=get_auth_context
)

Example: Integration with FastAPI Users

Here's an example of implementing an auth context provider using FastAPI Users:

from fastapi import Request, Depends
from fastapi_users import FastAPIUsers
from typing import Optional
from capebase.model import AuthContext

# Assuming you have FastAPI Users configured with your User model
fastapi_users = FastAPIUsers[User, int](
    get_user_manager,
    [auth_backend],
)

# Create auth context provider
async def get_auth_context(
    request: Request, 
    # You can use any FastAPI dependencies in your auth context provider
    user: Optional[User] = Depends(fastapi_users.current_user(active=True))
) -> AuthContext:
    if not user:
        return AuthContext()

    return AuthContext(
        id=str(user.id),
        role=user.role,
        context={}
    )

# Initialize CapeBase with the auth context
app = FastAPI()
cape = CapeBase(
    app=app,
    db_path="sqlite+aiosqlite:///app.db",
    get_auth_context=get_auth_context
)

This setup allows you to:

  • Use FastAPI Users' authentication system
  • Automatically populate user IDs in your models using FROM_AUTH_ID
  • Control access based on user roles
  • Access the authenticated user's context in your routes

Telemetry and Audit Trail

Metrics

CapeBase can be easily integrated with Prometheus for API endpoint monitoring using starlette-prometheus or other similar libraries:

from starlette_prometheus import PrometheusMiddleware, metrics

# Add Prometheus middleware to your FastAPI app
app.add_middleware(PrometheusMiddleware)
app.add_route("/metrics", metrics)

About

No description or website provided.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

点击 这是indexloc提供的php浏览器服务,不要输入任何密码和下载