Loading repository data…
iam-abbas / repository
A scalable and production ready boilerplate for FastAPI
A scalable and production ready boilerplate for FastAPI
This boilerplate follows a layered architecture that includes a model layer, a repository layer, a controller layer, and an API layer. Its directory structure is designed to isolate boilerplate code within the core directory, which requires minimal attention, thereby facilitating quick and easy feature development. The directory structure is also generally very predictable. The project's primary objective is to offer a production-ready boilerplate with a better developer experience and readily available features. It also has some widely used features like authentication, authorization, database migrations, type checking, etc which are discussed in detail in the Features section.
You need following to run this project:
I use asdf to manage my python versions. You can use it too. However, it is only supported on Linux and macOS. For Windows, you can use something like pyenv.
Once you have installed the above and have cloned the repository, you can follow the following steps to get the project up and running:
poetry shell
poetry install
docker-compose up -d
Copy the .env.example file to .env and update the values as per your needs.
Run the migrations:
make migrate
make run
The server should now be running on http://localhost:8000 and the API documentation should be available at http://localhost:8000/docs.
The project is designed to be modular and scalable. There are 3 main directories in the project:
core: This directory contains the central part of this project. It contains most of the boiler plate code like security dependencies, database connections, configuration, middlewares etc. It also contains the base classes for the models, repositories, and controllers. The core directory is designed to be as minimal as possible and usually requires minimal attention. Overall, the core directory is designed to be as generic as possible and can be used in any project. While building additional feature you may not need to modify this directory at all except for adding more controllers to the Factory class in core/factory.py.
app: This directory contains the actual application code. It contains the models, repositories, controllers, and schemas for the application. This is the directory you will be spending most of your time in while building features. The directory has following sub-directories:
models Here is where you add new tablesrepositories For each model, you need to create a repository. This is where you add the CRUD operations for the model.controllers For each logical unit of the application, you need to create a controller. This is where you add the business logic for the application.schemas This is where you add the schemas for the application. The schemas are used for validation and serialization/deserialization of the data.api: This directory contains the API layer of the application. It contains the API router, it is where you add the API endpoints.
The boilerplate contains a lot of features some of which are used in the application and some of which are not. The following sections describe the features in detail.
The migrations are handled by Alembic. The migrations are stored in the migrations directory. To create a new migration, you can run the following command:
make generate-migration
It will ask you for a message for the migration. Once you enter the message, it will create a new migration file in the migrations directory. You can then run the migrations using the following command:
make migrate
If you need to downgrade the database or reset it. You can use make rollback and make reset-database respectively.
The authentication used is basic implementation of JWT with bearer token. When the bearer token is supplied in the Authorization header, the token is verified and the user is automatically authenticated by setting request.user.id using middleware. To use the user model in any endpoint you can use the get_current_user dependency. If for any endpoint you want to enforce authentication, you can use the AuthenticationRequired dependency. It will raise a HTTPException if the user is not authenticated.
The boilerplate contains a custom row level permissions management module. It is inspired by fastapi-permissions. It is located in core/security/access_control.py. You can use this to enforce different permissions for different models. The module operates based on Principals and permissions. Every user has their own set of principals which need to be set using a function. Check core/fastapi/dependencies/permissions.py for an example. The principals are then used to check the permissions for the user. The permissions need to be defined at the model level. Check app/models/user.py for an example. Then you can use the dependency directly in the route to raise a HTTPException if the user does not have the required permissions. Below is an incomplete example:
from fastapi import APIRouter, Depends
from core.security.access_control import AccessControl, UserPrincipal, RolePrincipal, Allow
from core.database import Base
class User(Base):
__tablename__ = "users"
id = Column(Integer, primary_key=True)
name = Column(String)
email = Column(String, unique=True)
password = Column(String)
role = Column(String)
def __acl__(self):
return [
(Allow, UserPrincipal(self.id), "view"),
(Allow, RolePrincipal("admin"), "delete"),
]
def get_user_principals(user: User = Depends(get_current_user)):
return [UserPrincipal(user.id)]
Permission = AccessControl(get_user_principals)
router = APIRouter()
@router.get("/users/{user_id}")
def get_user(user_id: int, user: User = get_user(user_id), assert_access = Permission("view")):
assert_access(user)
return user
You can directly use the Cache.cached decorator from core.cache. Example
from core.cache import Cache
@Cache.cached(prefix="user", ttl=60)
def get_user(user_id: int):
...
The celery worker is already configured for the app. You can add your tasks in worker/ to run the celery worker, you can run the following command:
make celery-worker
The sessions are already handled by the middleware and get_session dependency which injected into the repositories through fastapi dependency injection inside the Factory class in core/factory.py. There is also Transactional decorator which can be used to wrap the functions which need to be executed in a transaction. Example:
@Transactional()
async def some_mutating_function():
...
Note: The decorator already handles the commit and rollback of the transaction. You do not need to do it manually.
If for any case you need an isolated sessions you can use standalone_session decorator from core.database. Example:
@standalone_session
async def do_something():
...
The boilerplate uses the repository pattern. Every model has a repository and all of them inherit base repository from core/repository. The repositories are located in app/repositories. The repositories are injected into the controllers inside the Factory class in core/factory/factory.py.py.
The base repository has the basic crud operations. All customer operations can be added to the specific repository. Example:
from core.repository import BaseRepository
from app.models.user import User
from sqlalchemy.sql.expression import select
class UserRepository(BaseRepository[User]):
async def get_by_email(self, email: str):
return await select(User).filter(User.email == email).gino.first()
To facilitate easier access to queries with complex joins, the BaseRepository class has a _query function (along with other handy functions like _all() and _one_or_none()) which can be used to write compplex queries very easily. Example:
async def get_user_by_email_join_tasks(email: str):
query = await self._query(join_)
query = query.filter(User.email == email)
return await self._one_or_none(query)
Note: For every join you want to make you need to create a function in the same repository with pattern _join_{name}. Example: _join_tasks for tasks. Example:
async def _join_tasks(self, query: Select) -> Select:
return query.options(joinedload(User.tasks))
Kind of to repositories, every logical unit of the application has a controller. The controller also has a primary repository which is injected into it. The controllers are located in app/controllers.
These controllers contain all the business logic of the application. Check app/controllers/auth.py for an example.
The schemas are located in app/schemas. The schemas are used to validate the request body and response body. The schemas are also used to generate the OpenAPI documentation. The schemas are inherited from BaseModel from pydantic. The schemas are primarily isolated into requests and responses which are pretty self explainatory.
You can use make format to format the code using black and isort.
You can use make lint to lint the code using pylint.
The project contains tests for all endpoints, some of the logical components like JWTHander and AccessControl and an example of testing complex inner components like BaseRepository. The tests are located in tests/. You can run the tests using make test.
Contributions are higly welcome. Please open an issue or a PR if you want to contribute.
This project is licensed under the terms of the MIT license. See the LICENSE file.
Selected from shared topics, language and repository description—not editorial ratings.
americadoodles /
Affordable NFT Marketplace API with Python and Django An optimized, full-featured boilerplate template designed for seamless Web3 integration, tailored specifically for NFT marketplace developers. This solution provides a comprehensive foundation for building scalable and efficient NFT marketplaces.
t4b4rz1n /
A production-ready Django REST Framework boilerplate with JWT auth, Celery, Docker, MinIO, and a custom ticketing system. Built for speed and scalability.
umairqadir97 /
Boilerplate for building Flask API, and scaling to production
piewared /
A production-ready FastAPI template for building scalable Python APIs with PostgreSQL, SQLModel, Redis, and Temporal—fully configured for local and cloud deployment.
mh-mukul /
🚀A FastAPI boilerplate for building scalable backends with user authentication, 🔑 role-based access control, and ⚡ SQLAlchemy integration. Ideal for REST APIs and microservices 🔥.
rishabh30 /
A boilerplate for a real-time user activity monitoring system built with Python, Kafka, and PostgreSQL. This project provides RESTful APIs for user management and tracks user activities through Kafka, processing and logging them into a database. Ideal for developers looking to implement scalable, real-time activity tracking in their applications.